"The Best Diversifiers For Your Equity Portfolio"

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
nedsaid
Posts: 13811
Joined: Fri Nov 23, 2012 12:33 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by nedsaid »

LilyFleur wrote: Fri Aug 23, 2019 11:40 am

So, bottom line: According to my AA, if I have invested in a bond fund with a .03% ER that approximates the Bloomberg Barclays Aggregate Fund, am I doing OK?
The answer is yes. Any Intermediate Term and Investment Grade bond fund with reasonable costs will work. A 0.03% expense ratio is very low. You are doing fine.
A fool and his money are good for business.
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

CULater wrote: Fri Aug 23, 2019 1:24 pm I don't know ...
I don't know either because I don't understand your last point.
CULater wrote: Fri Aug 23, 2019 1:24 pm The problem with using intermediate treasuries instead of long treasuries is that increasing the bond allocation reduces the expected portfolio return that you might want/need.
Why does using intermediate Treasuries instead of long Treasuries increase you bond allocation at all?
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

Doc wrote: Fri Aug 23, 2019 1:46 pm
CULater wrote: Fri Aug 23, 2019 1:24 pm I don't know ...
I don't know either because I don't understand your last point.
CULater wrote: Fri Aug 23, 2019 1:24 pm The problem with using intermediate treasuries instead of long treasuries is that increasing the bond allocation reduces the expected portfolio return that you might want/need.
Why does using intermediate Treasuries instead of long Treasuries increase you bond allocation at all?
I reckon what I was trying to say is that if you increase your bond allocation (e.g., from 40% to 50%), that is likely to lower risk at the expense of expected portfolio returns - that is the "de-risking" effect. Increasing the allocation to intermediate bonds will impact expected returns more than increasing the allocation to long bonds; due to fact the "diversification" effect is greater for long bonds than for intermediate bonds.
On the internet, nobody knows you're a dog.
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

CULater wrote: Fri Aug 23, 2019 2:43 pm I reckon what I was trying to say is that if you increase your bond allocation (e.g., from 40% to 50%), that is likely to lower risk at the expense of expected portfolio returns - that is the "de-risking" effect. Increasing the allocation to intermediate bonds will impact expected returns more than increasing the allocation to long bonds; due to fact the "diversification" effect is greater for long bonds than for intermediate bonds.
Back to the OP:

Benz said: "What surprised me in this latest data run, Susan, was that we didn't have to venture into long-term Treasuries to capture good diversification relative to equities. Short- and intermediate-term Treasuries did the job, too."

Is it "really different this time?"

I don't have a dog in this fight. My overall bond duration is only 2.96 years and almost all of the Treasury portion is less than one. And this has virtually nothing to do with "diversifying" my equites.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

Doc wrote: Fri Aug 23, 2019 3:10 pm
CULater wrote: Fri Aug 23, 2019 2:43 pm I reckon what I was trying to say is that if you increase your bond allocation (e.g., from 40% to 50%), that is likely to lower risk at the expense of expected portfolio returns - that is the "de-risking" effect. Increasing the allocation to intermediate bonds will impact expected returns more than increasing the allocation to long bonds; due to fact the "diversification" effect is greater for long bonds than for intermediate bonds.
Back to the OP:

Benz said: "What surprised me in this latest data run, Susan, was that we didn't have to venture into long-term Treasuries to capture good diversification relative to equities. Short- and intermediate-term Treasuries did the job, too."

Is it "really different this time?"

I don't have a dog in this fight. My overall bond duration is only 2.96 years and almost all of the Treasury portion is less than one. And this has virtually nothing to do with "diversifying" my equites.
I think there's two points of view on bonds. Most Bogleheads see bonds as safe "ballast" for their portfolios and look to equities for meaningful returns. Nothing wrong with that view. I'm a little more inclined to consider the diversifying value of bonds, because Treasuries do seem to be the single best diversifier for stocks; long term Treasuries especially. I just don't like the idea of depending on stocks-- they're too fickle!
On the internet, nobody knows you're a dog.
NMBob
Posts: 362
Joined: Thu Apr 23, 2015 8:13 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by NMBob »

Doc wrote: Fri Aug 23, 2019 6:58 am
NMBob wrote: Thu Aug 22, 2019 11:27 am 65percent us market with 35 short/mid/long treasuries 1978-july 2019


Portfolio Initial Balance Final Balance CAGR Stdev Best Year Worst Year Max. Drawdown Sharpe Ratio Sortino Ratio US Mkt Correlation
35%short $10,000 $505,130 9.89% 9.99% 27.50% -21.74% -32.60% 0.56 0.82 0.99
35%mid $10,000 $603,588 10.36% 10.21% 30.42% -19.41% -31.05% 0.59 0.87 0.98
35%long $10,000 $754,569 10.96% 10.77% 33.79% -16.20% -30.08% 0.62 0.92 0.93

https://www.portfoliovisualizer.com/bac ... total3=100
There are two considerations in the bond/equity decision process: long term and short term. If you have a set it and forget it portfolio and only rebalance on your mother in law's birthday the above analysis is appropriate. But if you are going to rebalance during a stock market crash those long term correlations aren't going to give you good answers.
I adjusted rebalancing to quarterly, monthly, and the first default band setting, and the relationships between the 3 stay relatively the same. Monthly makes them all lower returns.
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

CULater wrote: Fri Aug 23, 2019 4:15 pm I think there's two points of view on bonds. Most Bogleheads see bonds as safe "ballast" for their portfolios and look to equities for meaningful returns. Nothing wrong with that view. I'm a little more inclined to consider the diversifying value of bonds, because Treasuries do seem to be the single best diversifier for stocks; long term Treasuries especially.
Agreed. The only caveat is the Benz remark that recent data suggests that short and intermediate T's do as well.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Doc wrote: Sat Aug 24, 2019 7:53 am
CULater wrote: Fri Aug 23, 2019 4:15 pm I think there's two points of view on bonds. Most Bogleheads see bonds as safe "ballast" for their portfolios and look to equities for meaningful returns. Nothing wrong with that view. I'm a little more inclined to consider the diversifying value of bonds, because Treasuries do seem to be the single best diversifier for stocks; long term Treasuries especially.
Agreed. The only caveat is the Benz remark that recent data suggests that short and intermediate T's do as well.
The problem being that Benz was wrong, as discussed earlier.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

vineviz wrote: Sat Aug 24, 2019 7:57 am The problem being that Benz was wrong, as discussed earlier.
I missed any reference to the "the most recent data". How can anyone say Benz is wrong since she didn't supply any reference to the data she is using.

That said if I had a 80/20 AA that 20 would be long T's.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Doc wrote: Sat Aug 24, 2019 8:15 am
vineviz wrote: Sat Aug 24, 2019 7:57 am The problem being that Benz was wrong, as discussed earlier.
I missed any reference to the "the most recent data". How can anyone say Benz is wrong since she didn't supply any reference to the data she is using.

That said if I had a 80/20 AA that 20 would be long T's.
She was wrong because she was using an improper concept of "diversification": she looked only at the correlations of the various Treasury durations with stocks instead of looking at the covariances.

It's like saying that NBA star Kevin Durant and NFL star Jonathan Ogden are the same size because they are both 6'9" tall, ignoring the fact that Durant weighs 240 pounds and Ogden weighs 345 pounds.

Just like height alone is insufficient/inadequate as a measure of mass or volume, correlation alone is insufficient/inadequate as a measure of diversifcation.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

vineviz wrote: Sat Aug 24, 2019 8:52 am She was wrong because she was using an improper concept of "diversification": she looked only at the correlations of the various Treasury durations with stocks instead of looking at the covariances.
I don't know exactly what she looked at. And I don't care about correlations or covariances over long time periods. I want to know what happens over the short time periods when the stock market crashes. If you look at long time periods you are going to see little correlation in price between Treasury bonds and equities. But in the short term that is not true if the stock market is crashing. In that case the price of Treasuries goes up as stock prices go down - a "flight to quality" situation. It has long been accepted that long Treasuries make the best "insurance" in such a situation. I took Benz's comment as meaning that recently intermediate and even short term Treasuries did almost as well as long treasuries.

It is important to note that I am talking about price not total return. Most data I have seen like portfolio visualizer looks at total return and over extended periods not individual short term periods. And even if one looks at say thirty day periods making a short term judgement by averaging all the 30 day periods over ten or twenty years does not address the question of what happened in '08 or what may happen in '21.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

I took Benz's comment as meaning that recently intermediate and even short term Treasuries did almost as well as long treasuries.
Well, instead of taking her word for it let's look at Portfolio Visualizer. I compared 50/50 stock-treasury portfolios using intermediate treasuries vs. long term treasuries and what happened during the severe stock bear market in 2007-09. In a sense, Benz is correct. There was a about a 21% portfolio drawdown whether you were holding 50% intermediate or 50% long treasuries. From 1978-2019 the Sharpe ratios were the same for both portfolios as well. But here's the difference:

(1) Over that period, the portfolio with long treasuries had a cumulative return that was 35% greater. So, you experienced about the same overall portfolio losses (in terms of drawdown) but you ended up with a lot more $$$ at the end.

(2) Alternatively, you could have matched the cumulative return of the 50/50 intermediate portfolio with about 28% stocks / 72% long treasuries, so you would have been taking much lower equity risk. That allocation would have experienced a drawdown of 16% in 2007-2009 vs. 21% for the 50/50 portfolio with intermediate treasuries.

Greater diversification of the stock + long treasury portfolio is the explanation for both (1) and (2) above. Greater diversification = higher returns per unit of portfolio risk. You can collect the diversification bonus either in the form of higher returns or more downside protection or a little of both, depending on what is more important to you.
On the internet, nobody knows you're a dog.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Doc wrote: Sat Aug 24, 2019 10:55 amI don't know exactly what she looked at. And I don't care about correlations or covariances over long time periods.
I'm not say you have to care about diversification, just pointing out that any discussion that includes correlation but ignores variance is no longer a discussion about diversification.
Doc wrote: Sat Aug 24, 2019 10:55 am I took Benz's comment as meaning that recently intermediate and even short term Treasuries did almost as well as long treasuries.
That may very well have been what she meant but - again - that has nothing to do with diversification.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

CULater wrote: Sat Aug 24, 2019 11:29 am
I took Benz's comment as meaning that recently intermediate and even short term Treasuries did almost as well as long treasuries.
Well, instead of taking her word for it let's look at Portfolio Visualizer. I compared 50/50 stock-treasury portfolios using intermediate treasuries vs. long term treasuries and what happened during the severe stock bear market in 2007-09. In a sense, Benz is correct. There was a about a 21% portfolio drawdown whether you were holding 50% intermediate or 50% long treasuries. From 1978-2019 the Sharpe ratios were the same for both portfolios as well. But here's the difference:

(1) Over that period, the portfolio with long treasuries had a cumulative return that was 35% greater. So, you experienced about the same overall portfolio losses (in terms of drawdown) but you ended up with a lot more $$$ at the end.

(2) Alternatively, you could have matched the cumulative return of the 50/50 intermediate portfolio with about 25% stocks / 75% long treasuries, so you would have been taking much lower equity risk. That allocation would have experienced a drawdown of 16% in 2007-2009 vs. 21% for the 50/50 portfolio with intermediate treasuries.

Greater diversification of the stock + long treasury portfolio is the explanation for both (1) and (2) above. Greater diversification = higher returns per unit of portfolio risk. You can collect the diversification bonus either in the form of higher returns or more downside protection, depending on what is more important to you.
Maybe we are getting somewhere.
I compared 50/50 stock-treasury portfolios using intermediate treasuries vs. long term treasuries and what happened during the severe stock bear market in 2007-09.
Good. Maybe that's what Benz was addressing.
CULater wrote: Sat Aug 24, 2019 11:29 am Over that period, the portfolio with long treasuries had a cumulative return that was 35% greater. So, you experienced about the same overall portfolio losses (in terms of drawdown) but you ended up with a lot more $$$ at the end.
If that period is 1978-2019 I believe it to be completely irrelevant to the current period. You are including times before the Fed could control inflation. In the following chart the bars represent a money market account while the lines are TBM and the 500.

Imageupload free picture
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

vineviz wrote: Sat Aug 24, 2019 12:13 pm I'm not say you have to care about diversification, just pointing out that any discussion that includes correlation but ignores variance is no longer a discussion about diversification.
I'm not trying to make any kind of distinction between diversification and variance but just asking about the the price of long T's going up significantly more than Intermediate T's during a stock market crash.

Diversification and variance apply to what happens over 10 to 20 years not 10 or 20 weeks. Statistical analysis is usually based on normal distributions and stock market crashes are out of sample events.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

Yes, one can speculate that backtesting results won't hold up for this reason or that reason and your conclusions will have an impact on your allocation decisions. Nothing constant in the world of investing. Nothing wrong with that. However, it seems to me that the principal drivers of returns for stocks and long term treasuries haven't changed. I think long treasuries will continue to offer greater diversification to stocks than shorter duration treasuries and for that reason prefer them. I like the "risk parity" approach to asset allocation and it takes about 80% intermediates, 20% stocks to get to parity, and about 50/50 with long treasuries. I just don't want to hold 80% in intermediate treasuries right now and would rather amp my stock allocation higher than 20%, so prefer the long treasuries for that.
On the internet, nobody knows you're a dog.
User avatar
willthrill81
Posts: 20832
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by willthrill81 »

CULater wrote: Sat Aug 24, 2019 12:55 pm Yes, one can speculate that backtesting results won't hold up for this reason or that reason and your conclusions will have an impact on your allocation decisions. Nothing constant in the world of investing. Nothing wrong with that. However, it seems to me that the principal drivers of returns for stocks and long term treasuries haven't changed.
Many people love to point out that we don't have enough data to make 'statistically reliable' conclusions about this or that. Many others wholeheartedly agree but also point that we still have to make investment decisions, and the data and theory we have are all that we can logically use to make those decisions. I'm definitely in the latter camp.

Image
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Doc wrote: Sat Aug 24, 2019 12:50 pm I'm not trying to make any kind of distinction between diversification and variance but just asking about the the price of long T's going up significantly more than Intermediate T's during a stock market crash.
Okay. This thing you just said ("the price of long T's going up significantly more than Intermediate T's during a stock market crash") is precisely what covariance measures. Because to answer the question you ask you need to know both correlation (do Treasuries return more or less than average when stocks return less than average) and variance (how strongly do Treasuries go up and down). Covariance is the combination of correlation and variance, so that's what you seem to really care about whether you call it that or not.

Doc wrote: Sat Aug 24, 2019 12:50 pmDiversification and variance apply to what happens over 10 to 20 years not 10 or 20 weeks. Statistical analysis is usually based on normal distributions and stock market crashes are out of sample events.
Neither of these things is necessarily true.

Correlation and variance can be calculated for any number of periods of any length: you could calculate the correlation and variance for any pair of assets over the last 30 minutes of the trading day if you wanted to. That'd be a pretty terrible sample, leaving you with a pretty terrible estimate of the population parameter, but you could do it. So there's no reason to think that variance and correlation only applies over long time periods. These parameters happen to be very stationary, as opposed to mean returns which are not, so it's usually fine to use long historical samples as a basis for estimating future values. But this isn't strictly required.

As for the second point, many parts of modern portfolio theory do utilize assumptions of normality. This is not because financial economists think markets behave in a strictly normal fashion, but rather because the assumption of normality is required in order to estimate the parameters needed for portfolio construction.

If you're making dozens or hundreds of large trades a day this assumption is going to be problematic, but we are looking at monthly or even weekly returns - as you might if you are doing a monthly rebalancing, for instance - it turns out not to be such a big deal: the errors introduced by the normality assumption are greatly diminished to the point of being negligible. But if you didn't agree with that assessment you just need to step up your statistical game and start estimating things like coskewness and cokurtosis in addition to covariance. I don't think that any individual investor is going to be able to justify doing that.

Instead, either use MPT to form a diversified portfolio or just fall back on historical precedent. Since 2002 Vanguard Total Stock Market Index Fund (VTSMX) has had negative returns in 64 different months. The average return of iShares 1-3 Year Treasury Bond ETF (SHY) during those 64 "down" months was 0.33%. The average return of iShares 20+ Year Treasury Bond ETF (TLT) was 2.00% during this same months. This is the case even though SHY and TLT had virtually identical correlations with VTSMX: -0.32 and -0.31 respectively. You can see why Benz was in error by looking only at correlation instead of taking variance into account.

But wait, you say, what about those really bad months that are outside the normal distribution? There were 11 months when VTSMX had returns that were more than 2 standard deviations below average (i.e. -7.2% or less). The average return of SHY in these 11 months was 0.42% versus 2.5% for TLT.

Is this relationship perfect? Of course not. Will there be times when short-term Treasuries do better than long-term Treasuries? Probably so. Given what we know, in the face of an uncertain future, would it be smart to conclude that short-term Treasuries will be as good a diversifier as long-term Treasuries? Categorically not.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
catalina355
Posts: 283
Joined: Sun Jun 10, 2018 6:46 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by catalina355 »

vineviz wrote: Sat Aug 24, 2019 3:56 pm
Doc wrote: Sat Aug 24, 2019 12:50 pm I'm not trying to make any kind of distinction between diversification and variance but just asking about the the price of long T's going up significantly more than Intermediate T's during a stock market crash.
Okay. This thing you just said ("the price of long T's going up significantly more than Intermediate T's during a stock market crash") is precisely what covariance measures. Because to answer the question you ask you need to know both correlation (do Treasuries return more or less than average when stocks return less than average) and variance (how strongly do Treasuries go up and down). Covariance is the combination of correlation and variance, so that's what you seem to really care about whether you call it that or not.

Doc wrote: Sat Aug 24, 2019 12:50 pmDiversification and variance apply to what happens over 10 to 20 years not 10 or 20 weeks. Statistical analysis is usually based on normal distributions and stock market crashes are out of sample events.
Neither of these things is necessarily true.

Correlation and variance can be calculated for any number of periods of any length: you could calculate the correlation and variance for any pair of assets over the last 30 minutes of the trading day if you wanted to. That'd be a pretty terrible sample, leaving you with a pretty terrible estimate of the population parameter, but you could do it. So there's no reason to think that variance and correlation only applies over long time periods. These parameters happen to be very stationary, as opposed to mean returns which are not, so it's usually fine to use long historical samples as a basis for estimating future values. But this isn't strictly required.

As for the second point, many parts of modern portfolio theory do utilize assumptions of normality. This is not because financial economists think markets behave in a strictly normal fashion, but rather because the assumption of normality is required in order to estimate the parameters needed for portfolio construction.

If you're making dozens or hundreds of large trades a day this assumption is going to be problematic, but we are looking at monthly or even weekly returns - as you might if you are doing a monthly rebalancing, for instance - it turns out not to be such a big deal: the errors introduced by the normality assumption are greatly diminished to the point of being negligible. But if you didn't agree with that assessment you just need to step up your statistical game and start estimating things like coskewness and cokurtosis in addition to covariance. I don't think that any individual investor is going to be able to justify doing that.

Instead, either use MPT to form a diversified portfolio or just fall back on historical precedent. Since 2002 Vanguard Total Stock Market Index Fund (VTSMX) has had negative returns in 64 different months. The average return of iShares 1-3 Year Treasury Bond ETF (SHY) during those 64 "down" months was 0.33%. The average return of iShares 20+ Year Treasury Bond ETF (TLT) was 2.00% during this same months. This is the case even though SHY and TLT had virtually identical correlations with VTSMX: -0.32 and -0.31 respectively. You can see why Benz was in error by looking only at correlation instead of taking variance into account.

But wait, you say, what about those really bad months that are outside the normal distribution? There were 11 months when VTSMX had returns that were more than 2 standard deviations below average (i.e. -7.2% or less). The average return of SHY in these 11 months was 0.42% versus 2.5% for TLT.

Is this relationship perfect? Of course not. Will there be times when short-term Treasuries do better than long-term Treasuries? Probably so. Given what we know, in the face of an uncertain future, would it be smart to conclude that short-term Treasuries will be as good a diversifier as long-term Treasuries? Categorically not.
How would things look with IT Treasuries?
Call_Me_Op
Posts: 7980
Joined: Mon Sep 07, 2009 2:57 pm
Location: Milky Way

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Call_Me_Op »

Day9 wrote: Tue Aug 20, 2019 12:01 pm
nedsaid wrote: Tue Aug 20, 2019 10:37 am ...You can't collect 6%-7% anymore from a portfolio of quality bonds, you will get more like 2%...
This reminds me of how Jack Bogle used to say having an allocation to bonds is like having an "anchor to windward". I am no sailor but I took that to mean you have a part of your portfolio that will be going up, regardless of what the rest of your portfolio is doing. With today's low rates it's more like an anchor to zero real returns. I wish I had a large enough nautical vocabulary to come up with a witty analogy here.
How about just an anchor (drop the "to windward" part)?
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

@vineviz

I don't have time to go through your post in detail. But I think we are looking at different time scales. If you look at the covariance say on a three month basis you get the answer you advocate. But my point is that of that out of maybe 100 or more data points much more than the expected 5% are out of the expected 2 sigma expected norm then we do not have a normal distribution and we need to take cognisence of that abnormality. I think that the flight to quality (?) irrational responce resulting from that situation makes statistics based on a normal distribution less reliable (suspect?). I have no way or even any inclination of proving my idea. It is just a possible explantation of Benz's assertion.

If I was a 30 something with a 80/20 AA my FI portfolio would be long T's. But as a 70 something with a 50/50 AA having the Treasury portion of my FI in intermediate term makes sense to me.

Thank you Ms. Benz.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Doc wrote: Sat Aug 24, 2019 7:41 pm But my point is that of that out of maybe 100 or more data points much more than the expected 5% are out of the expected 2 sigma expected norm then we do not have a normal distribution and we need to take cognisence of that abnormality.
I addressed that in my post: unless you are day-trading, the impact of any divergence from normality is inconsequential. And to the extent that you think it matters, there are parameters beyond covariance (e.g. coskewness) that can account for it.
Doc wrote: Sat Aug 24, 2019 7:41 pmIf I was a 30 something with a 80/20 AA my FI portfolio would be long T's. But as a 70 something with a 50/50 AA having the Treasury portion of my FI in intermediate term makes sense to me.
Again, I'm not arguing that you should change your asset allocation based on anything I'm saying. If having your bond allocation in intermediate Treasuries is the best choice for you then I'll be the last one to suggest you change it.

But your personal asset allocation doesn't vindicate Morningstar on this issue: their claim that short- or intermediate-term Treasuries are equal to long-term Treasuries in terms of diversification benefit is simply incorrect. That doesn't mean that intermediate-term Treasuries are in any way a bad choice for a portfolio.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Random Musings
Posts: 5791
Joined: Thu Feb 22, 2007 4:24 pm
Location: Pennsylvania

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Random Musings »

Perhaps this thought is stale, but I recall that the volatility/expected returns tradeoff was not particularly appealing for LT bonds compared to shorter duration bonds.

Perhaps at current very low rates, that relationship has disapated since we now are at a permanently low plateau of interest rates. :D

RM
I figure the odds be fifty-fifty I just might have something to say. FZ
Day9
Posts: 988
Joined: Mon Jun 11, 2012 6:22 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Day9 »

Random Musings wrote: Sat Aug 24, 2019 9:05 pm Perhaps this thought is stale, but I recall that the volatility/expected returns tradeoff was not particularly appealing for LT bonds compared to shorter duration bonds.

Perhaps at current very low rates, that relationship has disapated since we now are at a permanently low plateau of interest rates. :D

RM
Yes long term bonds have a bad risk/return tradeoff when viewed in isolation. Their proponents say you have to look at them in the context of a whole portfolio. They also work for entities who have long term nominal liabilities they want to match, or they need consistent nominal income over a long period of time. They say insurance companies are an example of such an entity. But retirement investors have real liabilities.

I am a young (relative to the members on this board) accumulator with an 85/15 portfolio, bonds are split between long term treasuries and I Bonds. I Bonds are so good right now I believe everyone can improve their portfolio by buying them and extending the duration of the rest of your portfolio such that it is the same as it was before you bought the I Bonds. (And if you take credit risk, adjust your bond portfolio for that as well)
I'm just a fan of the person I got my user name from
andrew99999
Posts: 733
Joined: Fri Jul 13, 2018 8:14 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by andrew99999 »

vineviz wrote: Wed Aug 21, 2019 2:12 pm “De-risking” is most commonly defined as the process of reducing the variance of a portfolio.

“Diversification” is most commonly defined as the process of balancing the risk exposures within a portfolio.

It is possible to change a portfolio in ways that do both, do neither, or do one but not the other.

Starting with a portfolio that is 100% S&P 500, for instance and speaking in broad generalities:

Adding a small cap value fund would diversify but not de-risk.

Adding cash would de-risk but not diversify.

Adding a long-term Treasury bond fund would de-risk and diversify.

Adding a 3x leveraged S&P 500 fund would neither de-risk nor diversify.
What would you say gold does in terms of de-risking and diversifying?
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

andrew99999 wrote: Sun Aug 25, 2019 5:01 am
vineviz wrote: Wed Aug 21, 2019 2:12 pm “De-risking” is most commonly defined as the process of reducing the variance of a portfolio.

“Diversification” is most commonly defined as the process of balancing the risk exposures within a portfolio.

It is possible to change a portfolio in ways that do both, do neither, or do one but not the other.

Starting with a portfolio that is 100% S&P 500, for instance and speaking in broad generalities:

Adding a small cap value fund would diversify but not de-risk.

Adding cash would de-risk but not diversify.

Adding a long-term Treasury bond fund would de-risk and diversify.

Adding a 3x leveraged S&P 500 fund would neither de-risk nor diversify.
What would you say gold does in terms of de-risking and diversifying?
Following the original example, I'd expect that adding gold to a portfolio that is 100% S&P 500 would have negative de-risking effects but improve diversification. Depending on how much gold was added, the net effect of de-risking and diversification on expected volatility could be positive or negative with small amounts (e.g. 30% or less) of gold resulting in lower expected portfolio volatility.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

andrew99999 wrote: Sun Aug 25, 2019 5:01 am
vineviz wrote: Wed Aug 21, 2019 2:12 pm “De-risking” is most commonly defined as the process of reducing the variance of a portfolio.

“Diversification” is most commonly defined as the process of balancing the risk exposures within a portfolio.

It is possible to change a portfolio in ways that do both, do neither, or do one but not the other.

Starting with a portfolio that is 100% S&P 500, for instance and speaking in broad generalities:

Adding a small cap value fund would diversify but not de-risk.

Adding cash would de-risk but not diversify.

Adding a long-term Treasury bond fund would de-risk and diversify.

Adding a 3x leveraged S&P 500 fund would neither de-risk nor diversify.
What would you say gold does in terms of de-risking and diversifying?
Not vineviz and would like to hear his/her response. My take is that gold is a highly diversifying asset with respect to stocks. For example, a 50/50 combination of TSM with Gold has a combined portfolio volatility that is around 30% lower than the weighted average volatility of TSM and Gold. TSM and Gold are generally negatively correlated and have similar volatilities, so you would expect this.

As I understand "de-risking", it is the process of adding a lower volatility asset which has the effect of lowering (i.e, "dampening") overall portfolio volatility. For example, adding cash to stocks acts like this. It seems clear that gold is not a lower volatility asset compared to stocks, so the reduction in the volatility of stocks + gold is almost entirely due to the diversification effect of adding gold.

This raises an interesting paradox in my mind. Gold has such a low expected long-term return that it has decreased portfolio returns over the long term by quite a lot, although there have been shorter periods when this wasn't the case when gold was in a bull market. It would seem that "diversification" is good, in the sense that it has a very beneficial effect on portfolio volatility -- but it's not necessarily good if it significantly impacts portfolio returns. It would seem that the ideal diversifier should have three characteristics: a low correlation, a similar volatility, and a similar expected return. With respect to stocks, gold has the first two but not the last one.

Interestingly enough, during the period 1982-2019, a 50/50 allocation between stocks and cash had about the same return and maximum drawdown as a 50/50 allocation between stocks and cash. Both cash and gold had the same annualized return during this period. Cash had a low volatility, while gold had a high volatility. Apparently, the former allocation had virtually no diversification benefit, while the latter had virtually no de-risking benefit during this period and you ended up in the same place.
On the internet, nobody knows you're a dog.
User avatar
Doc
Posts: 9797
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Doc »

vineviz wrote: Sat Aug 24, 2019 8:25 pm Doc wrote: ↑Sat Aug 24, 2019 7:41 pm
But my point is that of that out of maybe 100 or more data points much more than the expected 5% are out of the expected 2 sigma expected norm then we do not have a normal distribution and we need to take cognisance of that abnormality.
vineviz wrote:I addressed that in my post: unless you are day-trading, the impact of any divergence from normality is inconsequential. And to the extent that you think it matters, there are parameters beyond covariance (e.g. coskewness) that can account for it.
I'm still trying to wrap my head around what you are saying. Up thead you wrote:
Morningstar found this “surprising” result because they misdefined “diversification” by treating de-risking and diversification as equivalent.

They are not, and Morningstar should know better because they just republished an article about de-risking.
In the OP Taylor referenced an article from Morningstar: "The Best Diversifiers for Your Equity Portfolio"
Benz wrote:Well, there are a number of ways that you could try to get your arms around this. One that we like at Morningstar is a statistic called correlation coefficient. And essentially, that means that we compare two assets' performance to one another. And what we're looking for is a negative correlation coefficient
I am not familiar with the formal definition of "diversification" in the current context but "negative correlation" between two assets seems to me to be a good thing in Treasury/equity discussion. Is it the term "Diverifers" that you disagree with or the "negative correlation coefficient"?

As regard to the negative correlation coefficient the point I was trying to make is that when you look at the correlation coefficient between equities and Treasuries over a long time period you get essentially a zero coefficient. However when there is a significant downturn in the stock market as in a crash that correlation coefficient goes highly negative which is what we desire IF we are going to rebalance into equities during that crisis. Is this co-skewness?

(Will you please give us a reference to the Morningstar article on de-risking that you mentioned. Thanks.)
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
willthrill81
Posts: 20832
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by willthrill81 »

CULater wrote: Sun Aug 25, 2019 8:42 amIt would seem that the ideal diversifier should have three characteristics: a low correlation, a similar volatility, and a similar expected return. With respect to stocks, gold has the first two but not the last one.
But now that many bonds are essentially offering a 0% real return or even lower for many international bonds, that may make gold more appealing, even if you believe that its real return will be basically zero over the long-term.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

Also important to point out the role of re-balancing per portfolio diversification. Here's some illustrative data for 1978-2019:

50% TSM + 50% LTT
- Annually rebalanced the combined portfolio was 26% less volatile than the weighted average volatilities
- Not rebalanced the combined portfolio was 16% less volatile than the weighted average volatilities

50% TSM + 50% GOLD
- Annually rebalanced the combined portfolio was 27% less volatile than the weighted average volatilities
- Not rebalanced the combined portfolio was 15% less volatile than the weighted average volatiles

So we see that re-balancing resulted in greater diversification for both long treasuries or gold combined with stocks. Rebalancing is necessary to maximize portfolio diversification at a given asset allocation. If you allow the AA to "drift" then the portfolio could become less diversified. The explanation is that over time it will tend to become more concentrated in stocks because stocks have the highest expected return. Less diversification = more risk so that's why we say that re-balancing is important for portfolio risk management. If you elect a particular AA because you trying to engineer a given diversification benefit, then you should probably be prepared to stick with that AA to have the best chance of meeting your objective.
On the internet, nobody knows you're a dog.
andrew99999
Posts: 733
Joined: Fri Jul 13, 2018 8:14 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by andrew99999 »

vineviz wrote: Wed Aug 21, 2019 2:12 pm “De-risking” is most commonly defined as the process of reducing the variance of a portfolio.

“Diversification” is most commonly defined as the process of balancing the risk exposures within a portfolio.

It is possible to change a portfolio in ways that do both, do neither, or do one but not the other.

Starting with a portfolio that is 100% S&P 500, for instance and speaking in broad generalities:

Adding a small cap value fund would diversify but not de-risk.

Adding cash would de-risk but not diversify.

Adding a long-term Treasury bond fund would de-risk and diversify.
One more question.
Would you say that cash would de-risk more than LTT and that as you move from the spectrum from cash to STT to ITT to LTT, you are de-risking less?
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

I'll take a stab and Vineviz can correct my errors.

DIVERSIFICATION
Asset diversification depends on the correlation between assets and their relative volatilities. The less positively correlated and the more similar the volatilities, the greater the resulting portfolio diversification. Generally, Cash has a zero correlation to stocks while bonds tend to be negatively correlated (but not always), so that would indicate that Cash is generally the least diversifying asset, while long treasuries is the most diversifying. This is especially true during stock market crashes, and less true when stocks are doing well.

With respect to volatility, Cash is the most dissimilar to stocks while long treasuries is the most similar. Based on that, cash is also less diversifying than bonds, especially long bonds.

DE-RISKING
Asset de-risking also depends on correlation and relative volatilities, but in a different way from diversification. A perfectly de-risking asset would have a high positive correlation to stocks but lower volatility. That means that this asset moves up and down precisely with stock movements, but has a lower "amplitude", so it dampens the overall amplitude of the portfolio. The correlation of Cash with Stocks is about zero, so it it de-risks but is not a "perfect" de-risker.

The correlation between stock movements and bond movements varies over time more than the correlation to Cash. But, if we assumed a long-term correlation between stocks and bonds was also zero (same as Cash) then Cash would de-risk more than bonds because it is less volatile than bonds.

BOTTOM LINE
During most periods, Cash de-risks Stocks more than Bonds because of it's lower relative volatility to Stocks and zero correlation; this is more true as we move from short term to long term bonds.

Cash, STT, ITT, and LTT all diversify Stocks to some extent with diversification increasing as we move from Cash to LTT. This would be true even if we assumed that the correlations with stocks were all zero. It would be especially true when the correlation between bonds and stocks goes negative during market crashes.

So, yes. In general, the ratio of the impact of de-risking to diversification on portfolio volatility moves from higher to lower as we move from Cash to Long Treasuries with Cash having the greatest de-risking impact and lowest diversification impact and Long Treasuries having the least de-risking impact and the greatest diversification impact.
On the internet, nobody knows you're a dog.
andrew99999
Posts: 733
Joined: Fri Jul 13, 2018 8:14 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by andrew99999 »

Thanks CULater.
CULater wrote: Mon Aug 26, 2019 9:17 am DE-RISKING
Asset de-risking also depends on correlation and relative volatilities, but in a different way from diversification. A perfectly de-risking asset would have a high positive correlation to stocks but lower volatility. That means that this asset moves up and down precisely with stock movements, but has a lower "amplitude", so it dampens the overall amplitude of the portfolio. The correlation of Cash with Stocks is about zero, so it it de-risks but is not a "perfect" de-risker.
Bold added by myself

The lower volatility I understand, but when talking about de-risking in isolation, are you sure it is necessary that there would be a high (or any) correlation with stocks? I would have assumed that it is purely a lower amplitude of volatility.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

andrew99999 wrote: Mon Aug 26, 2019 7:48 pm Thanks CULater.
CULater wrote: Mon Aug 26, 2019 9:17 am DE-RISKING
Asset de-risking also depends on correlation and relative volatilities, but in a different way from diversification. A perfectly de-risking asset would have a high positive correlation to stocks but lower volatility. That means that this asset moves up and down precisely with stock movements, but has a lower "amplitude", so it dampens the overall amplitude of the portfolio. The correlation of Cash with Stocks is about zero, so it it de-risks but is not a "perfect" de-risker.
Bold added by myself

The lower volatility I understand, but when talking about de-risking in isolation, are you sure it is necessary that there would be a high (or any) correlation with stocks? I would have assumed that it is purely a lower amplitude of volatility.
This is right. If the variance is low enough an asset can only really function as a de-risker no matter what the correlation coefficient is.

Technically, an asset with a correlation of +1.0 would provide no diversification benefit no matter the variance, but with a low enough variance even a correlation of -1.0 does very little diversifiying.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
hdas
Posts: 1327
Joined: Thu Jun 11, 2015 8:24 am

[Deleted]

Post by hdas »

[Deleted]
Last edited by hdas on Wed Jan 29, 2020 8:58 pm, edited 1 time in total.
....
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

andrew99999 wrote: Mon Aug 26, 2019 7:48 pm Thanks CULater.
CULater wrote: Mon Aug 26, 2019 9:17 am DE-RISKING
Asset de-risking also depends on correlation and relative volatilities, but in a different way from diversification. A perfectly de-risking asset would have a high positive correlation to stocks but lower volatility. That means that this asset moves up and down precisely with stock movements, but has a lower "amplitude", so it dampens the overall amplitude of the portfolio. The correlation of Cash with Stocks is about zero, so it it de-risks but is not a "perfect" de-risker.
Bold added by myself

The lower volatility I understand, but when talking about de-risking in isolation, are you sure it is necessary that there would be a high (or any) correlation with stocks? I would have assumed that it is purely a lower amplitude of volatility.
The more highly positively correlated an asset is with stocks, the more it can "de-risk" based on how low it's volatility is; positive correlation multiplies the impact of lower volatility on de-risking. A "pure" de-risking" asset would have a correlation +1.0 with stocks and have lower volatility-- all of it's potential benefit would be due to de-risking alone.

Conversely, a "pure" diversifying asset would have a correlation of -1.0 and the same volatility as stocks. It zigs when stocks zag and with the same amplitude. An equally-weighted combination of this asset with stocks would have a very small volatility compared with the weighted average volatilities. But it would also have a much lower return because the returns of these two assets cancel each other out. Being long and being short the same stock would act like this. Diversification is good for portfolio returns -- to a point.

Assets will have both a de-risking and diversification impact when combined with stocks; it's a question of the relative impact. Low volatility assets are skewed toward de-risking, while high volatility assets are skewed toward diversification; but it depends on the correlations. High correlating assets are skewed toward de-risking and low correlated assets toward diversification; but it depends on the volatilities. Cash, short term bonds are mostly de-risking assets (low volatility, low correlations to stocks). Gold, commodities, and long Treasuries are mostly diversifying assets (high volatility, negative correlations to stocks).

Different categories of stock, such as foreign stocks, have not been great de-riskers or diversifiers in the technical sense because they are both highly correlated to U.S. stocks and have similar volatility. They don't de-risk much because of the similar volatility and they don't diversify much because of the high positive correlation. Since the volatility of foreign stocks isn't likely to change that much, their diversification benefit depends on the correlation between U.S. stocks and foreign stocks becoming weaker than it has been.
On the internet, nobody knows you're a dog.
Dottie57
Posts: 9168
Joined: Thu May 19, 2016 5:43 pm
Location: Earth Northern Hemisphere

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Dottie57 »

Lovely teaching story about the meaning of “anchor to windward” from “Circe Institute” (educational entity).

https://www.circeinstitute.org/blog/anchor-windward
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

We know that over the long run, adding SCV has produced higher portfolio returns which has been shown by advocates of "tilting" toward SCV. I was interested if this seems to have anything to do with the diversifying benefit of SCV. The answer I found seems to be "no." I looked at this using the asset class data from Portfolio Visualizer.

I looked at the data for an equally weighted portfolio of TSM and SCV for the period 1978 - 2019. The weighted average volatility was 16.6%, while the portfolio had a volatility of 16.2% with annual rebalancing. This indicates that SCV had very little diversification impact when combined with TSM.

Over this period, the CAGR for TSM was 10.4% while it was 13.9% for SCV. The conclusion that I draw from this rough analysis is that SCV has improved portfolio returns primarily because it has had higher returns than TSM, but it hasn't actually improved portfolio returns because it added any diversification "free lunch." It has simply increased portfolio risk and taking that additional risk has paid off with higher overall portfolio returns. No "special sauce" has been provided. As I recall, Nisiprius has posted on this as well and my analysis seems to confirm.
On the internet, nobody knows you're a dog.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

CULater wrote: Tue Aug 27, 2019 9:06 pm We know that over the long run, adding SCV has produced higher portfolio returns which has been shown by advocates of "tilting" toward SCV. I was interested if this seems to have anything to do with the diversifying benefit of SCV. The answer I found seems to be "no." I looked at this using the asset class data from Portfolio Visualizer.

I looked at the data for an equally weighted portfolio of TSM and SCV for the period 1978 - 2019. The weighted average volatility was 16.6%, while the portfolio had a volatility of 16.2% with annual rebalancing. This indicates that SCV had very little diversification impact when combined with TSM.
I'm not sure what parameters you were using, but these results aren't consistent with what I get for the same portfolios over the same time period.

The portfolio had an annualized standard deviation of 15.83% whereas the weighted average volatility of the two assets is 16.26%. The difference of 0.43% can be quantified as the diversification benefit, and it's not inconsequential.

To illustrate, here are three hypothetical portfolios for a retiree starting in 1978 with $100k and withdrawing an inflation-adjusted $565/month. All three are 60/40 stock/bond portfolios with the bonds being intermediate US treasuries.

Portfolio 1 is 60% TSM. Portfolio 2 is 55% TSM + 5% SCV. Portfolio 3 is 50% TSM + 10% SCV. In other words, Portfolios 2 and 3 represent the tiniest of SCV tilts. Observe the difference that 0.43% diversification benefit produced for a hypothetical retiree. Note that Portfolios 1 and 2 had standard deviations of 9.57% and 9.59% respectively: the advantage of Portfolio 2 wasn't due to more volatility.

Image
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

It turns out that I was actually using the period 1972-2019. The weighted average vol for TSM and SCV was 16.62 compared to a vol of 16.15 for a 50/50 portfolio, which reduces the volatility by only 3% which seems inconsequential to me.
On the internet, nobody knows you're a dog.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

CULater wrote: Wed Aug 28, 2019 6:34 am It turns out that I was actually using the period 1972-2019. The weighted average vol for TSM and SCV was 16.62 compared to a vol of 16.15 for a 50/50 portfolio, which reduces the volatility by only 3% which seems inconsequential to me.
I’d just urge you not to assume that any particular number is inherently inconsequential or very consequential. In other words, intuition about “big impact” or “small impact” isn’t necessarily reliable.

Especially for retiree portfolios in withdrawal, even diversification improvements that seem minuscule can result in material impacts.

It’s true that in rank ordering, SCV is probably going to be less powerful than international diversification and certainly less powerful than long bonds.

That doesn’t mean it’s inconsequential, and given that such diversification can basically be had for free I’d argue it’s irresponsible not to do it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
willthrill81
Posts: 20832
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by willthrill81 »

vineviz wrote: Wed Aug 28, 2019 6:52 am Especially for retiree portfolios in withdrawal, even diversification improvements that seem minuscule can result in material impacts.
Indeed. A visual examination of the chart with the three portfolios you posted above reveals this. After the first decade, a casual observer could have easily concluded that while adding SCV helped, it didn't 'help much'. But 20+ years down the line, it made a very substantial difference.

Our brains are just not wired to intuitively understand exponential growth.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

vineviz wrote: Wed Aug 28, 2019 6:52 am
CULater wrote: Wed Aug 28, 2019 6:34 am It turns out that I was actually using the period 1972-2019. The weighted average vol for TSM and SCV was 16.62 compared to a vol of 16.15 for a 50/50 portfolio, which reduces the volatility by only 3% which seems inconsequential to me.
I’d just urge you not to assume that any particular number is inherently inconsequential or very consequential. In other words, intuition about “big impact” or “small impact” isn’t necessarily reliable.

Especially for retiree portfolios in withdrawal, even diversification improvements that seem minuscule can result in material impacts.

It’s true that in rank ordering, SCV is probably going to be less powerful than international diversification and certainly less powerful than long bonds.

That doesn’t mean it’s inconsequential, and given that such diversification can basically be had for free I’d argue it’s irresponsible not to do it.
I think that's correct. But there is more than one moving part involved. Since SCV is more volatile than TSM, adding it will increase overall portfolio volatility depending on the weight; increasing the bond allocation can control for that. I played around with this using Portfolio Visualizer for a 50% bond allocation and the stock allocation split evenly between TSM and SCV. I found that shifting about 4% - 5% from stocks to bonds was necessary to do that; but you still ended up with a higher compound return than not adding SCV, so it seemed worth it over the long run.
On the internet, nobody knows you're a dog.
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

CULater wrote: Wed Aug 28, 2019 10:34 am
vineviz wrote: Wed Aug 28, 2019 6:52 am
CULater wrote: Wed Aug 28, 2019 6:34 am It turns out that I was actually using the period 1972-2019. The weighted average vol for TSM and SCV was 16.62 compared to a vol of 16.15 for a 50/50 portfolio, which reduces the volatility by only 3% which seems inconsequential to me.
I’d just urge you not to assume that any particular number is inherently inconsequential or very consequential. In other words, intuition about “big impact” or “small impact” isn’t necessarily reliable.

Especially for retiree portfolios in withdrawal, even diversification improvements that seem minuscule can result in material impacts.

It’s true that in rank ordering, SCV is probably going to be less powerful than international diversification and certainly less powerful than long bonds.

That doesn’t mean it’s inconsequential, and given that such diversification can basically be had for free I’d argue it’s irresponsible not to do it.
I think that's correct. But there is more than one moving part involved. Since SCV is more volatile than TSM, adding it will increase overall portfolio volatility depending on the weight; increasing the bond allocation can control for that. I played around with this using Portfolio Visualizer for a 50% bond allocation and the stock allocation split evenly between TSM and SCV. I found that shifting about 4% - 5% from stocks to bonds was necessary to do that; but you still ended up with a higher compound return than not adding SCV, so it seemed worth it over the long run.
Yep.

Another approach would to reduce the volatility of the large cap holdings using using some sort of low beta or dividend yield fund. Something like a utilities sector fund or VYM would probably work just fine.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Ferdinand2014
Posts: 1688
Joined: Mon Dec 17, 2018 6:49 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Ferdinand2014 »

vineviz wrote: Sat Aug 24, 2019 3:56 pm
Doc wrote: Sat Aug 24, 2019 12:50 pm I'm not trying to make any kind of distinction between diversification and variance but just asking about the the price of long T's going up significantly more than Intermediate T's during a stock market crash.
Okay. This thing you just said ("the price of long T's going up significantly more than Intermediate T's during a stock market crash") is precisely what covariance measures. Because to answer the question you ask you need to know both correlation (do Treasuries return more or less than average when stocks return less than average) and variance (how strongly do Treasuries go up and down). Covariance is the combination of correlation and variance, so that's what you seem to really care about whether you call it that or not.

Doc wrote: Sat Aug 24, 2019 12:50 pmDiversification and variance apply to what happens over 10 to 20 years not 10 or 20 weeks. Statistical analysis is usually based on normal distributions and stock market crashes are out of sample events.
Neither of these things is necessarily true.

Correlation and variance can be calculated for any number of periods of any length: you could calculate the correlation and variance for any pair of assets over the last 30 minutes of the trading day if you wanted to. That'd be a pretty terrible sample, leaving you with a pretty terrible estimate of the population parameter, but you could do it. So there's no reason to think that variance and correlation only applies over long time periods. These parameters happen to be very stationary, as opposed to mean returns which are not, so it's usually fine to use long historical samples as a basis for estimating future values. But this isn't strictly required.

As for the second point, many parts of modern portfolio theory do utilize assumptions of normality. This is not because financial economists think markets behave in a strictly normal fashion, but rather because the assumption of normality is required in order to estimate the parameters needed for portfolio construction.

If you're making dozens or hundreds of large trades a day this assumption is going to be problematic, but we are looking at monthly or even weekly returns - as you might if you are doing a monthly rebalancing, for instance - it turns out not to be such a big deal: the errors introduced by the normality assumption are greatly diminished to the point of being negligible. But if you didn't agree with that assessment you just need to step up your statistical game and start estimating things like coskewness and cokurtosis in addition to covariance. I don't think that any individual investor is going to be able to justify doing that.

Instead, either use MPT to form a diversified portfolio or just fall back on historical precedent. Since 2002 Vanguard Total Stock Market Index Fund (VTSMX) has had negative returns in 64 different months. The average return of iShares 1-3 Year Treasury Bond ETF (SHY) during those 64 "down" months was 0.33%. The average return of iShares 20+ Year Treasury Bond ETF (TLT) was 2.00% during this same months. This is the case even though SHY and TLT had virtually identical correlations with VTSMX: -0.32 and -0.31 respectively. You can see why Benz was in error by looking only at correlation instead of taking variance into account.

But wait, you say, what about those really bad months that are outside the normal distribution? There were 11 months when VTSMX had returns that were more than 2 standard deviations below average (i.e. -7.2% or less). The average return of SHY in these 11 months was 0.42% versus 2.5% for TLT.

Is this relationship perfect? Of course not. Will there be times when short-term Treasuries do better than long-term Treasuries? Probably so. Given what we know, in the face of an uncertain future, would it be smart to conclude that short-term Treasuries will be as good a diversifier as long-term Treasuries? Categorically not.
What about the months when LTT were down significantly more then STT when VTSMX was up? There are 2 sides to the LTT coin. Overall, historically I do get LTT with equities has led to increased total returns over STT with equities, which is of course for most circumstances what we should look for unless liquidity and stability for near term expenses are a need or consideration. With interest rates at all time lows, the potential capital gains in short term situations seem less likely to be a large as in the past when we started at interest rates of say 5-6%. Although the difference from 2-1 represents 50% vs say 5-4 which represents 20% difference despite the same 1% interest rate change in terms of convexity. With that said, do you think that the diversification benefit of LTT going forward at an interest rate of 2% or less is much lower then the detraction if inflation were to rise unexpectedly compared to the past 30 years?
“You only find out who is swimming naked when the tide goes out.“ — Warren Buffett
User avatar
vineviz
Posts: 7807
Joined: Tue May 15, 2018 1:55 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by vineviz »

Ferdinand2014 wrote: Wed Aug 28, 2019 12:58 pm What about the months when LTT were down significantly more then STT when VTSMX was up?
I don't understand the question: what about it? When bonds are down, long duration bonds will be down more than short duration bonds pretty much every time. That's what you're buying, and what you want.
Ferdinand2014 wrote: Wed Aug 28, 2019 12:58 pm With interest rates at all time lows, the potential capital gains in short term situations seem less likely to be a large as in the past when we started at interest rates of say 5-6%. Although the difference from 2-1 represents 50% vs say 5-4 which represents 20% difference despite the same 1% interest rate change in terms of convexity. With that said, do you think that the diversification benefit of LTT going forward at an interest rate of 2% or less is much lower then the detraction if inflation were to rise unexpectedly compared to the past 30 years?
Interest rates have been making new lows pretty consistently for the last 40 years, which I think should serve as a cautionary tale for anyone who thinks timing bond markets is inherently easier than timing stock markets.

Either way, the diversification benefits of an asset don't depend on its return. So no, I don't expect the diversification benefit of LTTs be lower in the future.

As for future returns, I think they'll be lower in the future than in the past because the YTM is lower now than it was 15 or 20 years ago. But that's true for bonds at every point along the yield curve, and trying to pick a spot on the yield curve based on what bond markets are forecast to do is not something I'd do without a functional crystal ball.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

Another approach would to reduce the volatility of the large cap holdings using using some sort of low beta or dividend yield fund. Something like a utilities sector fund or VYM would probably work just fine.
I've been using Portfolio Visualizer to see if I could find a combination of asset weights that allows me to add SCV without increasing the overall risk of a portfolio with TSM and LTT. It turns out, it's a tricky business with lots of moving parts. The only successful solution I found was to reduce the duration of the bond holding by adding an allocation to short term bonds. I was unsuccessful trying to do it using a low beta fund (USMV), and I was unsuccessful by reducing the overall stock allocation as well, because long bonds tend to have the same volatility as stocks so shifting the allocation between stocks and bonds doesn't accomplish much -- you have to reduce the duration of the bonds, which reduces their volatility and expected return. The result was not favorable to portfolio returns, which were lower as a result of decreasing bond duration somewhat.

I have to conclude that it's probably wrong to assume that you can add SCV without increasing overall portfolio risk as well, unless you reduce the duration of your bond holdings. So, if you simply "tilt" your stock allocation by replacing a portion of TSM with SCV, I think you have to accept the fact that portfolio risk will thereby be increased. I suspect that it's mostly the added risk that accounts for higher returns with SCV over the long run, and the small incremental portfolio diversification has little to do with it. SCV is likely a risk story and not a diversification story in the real world.
On the internet, nobody knows you're a dog.
Ferdinand2014
Posts: 1688
Joined: Mon Dec 17, 2018 6:49 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Ferdinand2014 »

vineviz wrote: Wed Aug 28, 2019 3:53 pm
Ferdinand2014 wrote: Wed Aug 28, 2019 12:58 pm What about the months when LTT were down significantly more then STT when VTSMX was up?
I don't understand the question: what about it? When bonds are down, long duration bonds will be down more than short duration bonds pretty much every time. That's what you're buying, and what you want.
Ferdinand2014 wrote: Wed Aug 28, 2019 12:58 pm With interest rates at all time lows, the potential capital gains in short term situations seem less likely to be a large as in the past when we started at interest rates of say 5-6%. Although the difference from 2-1 represents 50% vs say 5-4 which represents 20% difference despite the same 1% interest rate change in terms of convexity. With that said, do you think that the diversification benefit of LTT going forward at an interest rate of 2% or less is much lower then the detraction if inflation were to rise unexpectedly compared to the past 30 years?
Interest rates have been making new lows pretty consistently for the last 40 years, which I think should serve as a cautionary tale for anyone who thinks timing bond markets is inherently easier than timing stock markets.

Agree that timing the bond market is a fools errand. However, the possibilities are higher or lower. If they go higher, starting at such current lows, wouldn’t that be much more detrimental to returns with LTT as they can go much higher most likely then if we started at say 5%. If they go much lower, we are in negative interest rate territory. How that effects volatility, correlation and diversification is difficult for me to understand.

Either way, the diversification benefits of an asset don't depend on its return. So no, I don't expect the diversification benefit of LTTs be lower in the future.

Do they not depend on volatility and correlation? If volatility goes down because we bounce along at low or even negative interest rates, wouldn’t that generally favor short term bond if that situation holds true and also avoid the downside if interest rates were to rise from an already low base? Just trying to understand your explanation.

As for future returns, I think they'll be lower in the future than in the past because the YTM is lower now than it was 15 or 20 years ago. But that's true for bonds at every point along the yield curve, and trying to pick a spot on the yield curve based on what bond markets are forecast to do is not something I'd do without a functional crystal ball.
“You only find out who is swimming naked when the tide goes out.“ — Warren Buffett
Day9
Posts: 988
Joined: Mon Jun 11, 2012 6:22 pm

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by Day9 »

CULater wrote: Wed Aug 28, 2019 4:41 pm
Another approach would to reduce the volatility of the large cap holdings using using some sort of low beta or dividend yield fund. Something like a utilities sector fund or VYM would probably work just fine.
I've been using Portfolio Visualizer to see if I could find a combination of asset weights that allows me to add SCV without increasing the overall risk of a portfolio with TSM and LTT. It turns out, it's a tricky business with lots of moving parts. The only successful solution I found was to reduce the duration of the bond holding by adding an allocation to short term bonds. I was unsuccessful trying to do it using a low beta fund (USMV), and I was unsuccessful by reducing the overall stock allocation as well, because long bonds tend to have the same volatility as stocks so shifting the allocation between stocks and bonds doesn't accomplish much -- you have to reduce the duration of the bonds, which reduces their volatility and expected return. The result was not favorable to portfolio returns, which were lower as a result of decreasing bond duration somewhat.

I have to conclude that it's probably wrong to assume that you can add SCV without increasing overall portfolio risk as well, unless you reduce the duration of your bond holdings. So, if you simply "tilt" your stock allocation by replacing a portion of TSM with SCV, I think you have to accept the fact that portfolio risk will thereby be increased. I suspect that it's mostly the added risk that accounts for higher returns with SCV over the long run, and the small incremental portfolio diversification has little to do with it. SCV is likely a risk story and not a diversification story in the real world.
This is the topic of Larry Swedroe's book Reducing the Risk of Black Swans. I have not read the book but I know he uses deep SCV tilts with funds like BOSVX and 5-year safe bonds (Treasuries, but he finds higher yielding FDIC insured CDs). The funds he uses are probably tilted much deeper than the funds or indices you were using in your portfolio visualizer tests. Check out the book or Larry's free articles on this topic for more info.
I'm just a fan of the person I got my user name from
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

Re: "The Best Diversifiers For Your Equity Portfolio"

Post by CULater »

It turns out that from 2003-2019, a portfolio of 47% SCV/53% LTT had the same annualized return as a portfolio of 50% TSM/50% LTT. I used VTI, IJS, and TLT with Portfolio Visualizer. So you got the same portfolio return with a somewhat lower SCV stock allocation compared to TSM. Is the explanation for this the greater risk of SCV or is it greater portfolio diversification? It turns out to be risk.

The SCV portfolio had an annualized SD of 9.26%, compared to 8.08% for the TSM portfolio, so it was clearly riskier. But was it also more diversified?

Comparing the ratios of the weighted average volatilities of the portfolio assets to the portfolio volatilities, they are virtually the same: a 39.8% ratio for the TSM/LTT portfolio vs. a 40.3% ratio for the SCV/LTT portfolio. The ratio indicates the impact of diversification.

So the volatility of the SCV portfolio was 14.6% greater, while the diversification ratio was just 1.3% greater. Based on these data, the main impact of SCV on portfolio returns is due primarily to the added risk (volatility) and has very little to do with any marginal diversification benefit.
On the internet, nobody knows you're a dog.
Post Reply