De-Risking and Diversification aren't the same thing.

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

De-Risking and Diversification aren't the same thing.

Post by vineviz » Tue Jul 09, 2019 5:09 pm

I've previously pointed out that investors sometimes confuse diversification with a different concept known as "de-risking". The explicit goal of de-risking a portfolio is to lower its volatility. And just as diversification doesn't always de-risk a portfolio, de-risking doesn't always improve the diversification of portfolio.

Diversification is the process of combining assets that have less-than-perfectly-positive correlation in order to spread the risk of the portfolio across multiple independent sources of risk and return. Diversification may reduce the risk of the portfolio, but need not do so.

De-risking is the process of adding low-volatility assets to a portfolio in order to reduce the volatility of the portfolio.1 De-risking may improve the diversification of the portfolio, but need not do so.

The distinction between "diversification" and "de-risking" is important to understand because diversification effects provide investors with the proverbial free lunch whereas de-risking does not. Imprecise or confused thinking can lead investors to make sub-optimal portfolio choices, especially when it comes to allocating between stocks and bonds.

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.

By exploiting the properties of covariance, investors can combine not-totally-correlated risky assets into a portfolio which has a volatility that is less than the weighted average volatility of the two assets but with a return that is equal to the weighted average return of the assets.

Same return + less volatility = free lunch.

You can see that here, as two hypothetical uncorrelated (i.e. zero correlation) assets with the same return and volatility are combined at various ratios the volatility decreases as the portfolio approaches the balance point but the return stays constant. This is the "free lunch" benefit of diversification.

Image

In the de-risking case, though, where asset B has lower risk and lower return you can see that lower volatility is associated with lower return. Less risk, but no (or much less) of the "free lunch effect".

Image

It is this second tradeoff that many investors make when they build so-called "balanced" portfolios using low-risk instruments like cash, CDs, money markets, short-term bonds, or total bond market funds. They are sacrificing the "free lunch effect" by accepting less return for a given amount of risk.

The amount of this loss can be quantified and decomposed into two components: a "diversification effect" and a "de-risking effect". Investors are compensated for the former but not for the latter.

To illustrate, I constructed three different 60/40 balanced portfolios using Vanguard Total Stock Market Index (VTSMX) and cash, Vanguard Total Bond Market Index (VBMFX), and Vanguard Long-Term Treasury (VUSTX) from 1992 to present.

Over this period Vanguard Total Stock Market Index had an annualized volatility of 14.5% whereas each of the 60/40 portfolios had similar volatilities of 8.7% to 8.8%.

Image

You can see that the 60/40 portfolio using cash resulted almost entirely from de-risking and enjoyed virtually no diversification benefit.

At the other end, the 60/40 portfolio using long-term Treasury bonds benefited greatly from diversification effects and much less from de-risking effects.

In all three cases the total portfolio risk reduction was incredibly similar, but in the high-diversification case the investor reduced idiosyncratic risk (i.e. uncompensated risk) by a much greater amount. As a consequence, their "free lunch effect" was much greater and manifests as higher expected returns for a portfolio with the same overall expected volatility. This is a much more efficient portfolio.

Edit: Salient Partners published a white paper entitled The Free Lunch Effect that inspired this post. Check it out at http://www.salientpartners.com/wp-conte ... effect.pdf (just don't think I'm endorsing their attempt to sell you commodity funds, MLPs or hedge funds).




1 For reasons of convenience and clarity, I'm using the conventional practice of using the words "risk" and "volatility" interchangeably in this particular thread.
Last edited by vineviz on Tue Jul 09, 2019 6:33 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

Dottie57
Posts: 6726
Joined: Thu May 19, 2016 5:43 pm

Re: De-Risking and Diversification aren't the same thing.

Post by Dottie57 » Tue Jul 09, 2019 5:36 pm

How did you determine the percent assigned to de-risking vs diversification?

Very interesting post.

User avatar
hdas
Posts: 1103
Joined: Thu Jun 11, 2015 8:24 am

Re: De-Risking and Diversification aren't the same thing.

Post by hdas » Tue Jul 09, 2019 6:09 pm

vineviz wrote:
Tue Jul 09, 2019 5:09 pm

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.
Nice post, however, the tenet you mention is not true or at least heavily challenged by the existence of the Low Volatility anomaly. Cheers :greedy
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Tue Jul 09, 2019 6:22 pm

Dottie57 wrote:
Tue Jul 09, 2019 5:36 pm
How did you determine the percent assigned to de-risking vs diversification?
Example: say that VTSMX has a volatility of 14.51% and VBMFX has a volatility of 3.55%. Further, say that a 60/40 combination of those two funds has a volatility of 8.82%.

The "de-risking effect" is simply the difference between the VTSMX volatility (14.51%) and the weighted average of the two funds:

Code: Select all

14.51% - ((.6 * 14.51%) + (.4 * 3.55%)) = 14.51% - 10.13% = 4.38%
The "diversification effect" is the difference between the weighted average volatility of the two funds and the volatility of the 60/40 portfolio:

Code: Select all

((.6 * 14.51%) + (.4 * 3.55%)) - 8.82% = 10.13% - 8.82% = 1.31%
If the second asset has a sufficiently low correlation, it can be MORE volatile than the first asset and still reduce the overall portfolio volatility. If I add an example 60/40 portfolio using Vanguard Extended Duration Treasury ETF (EDV), the de-risking effect is actually negative (because EDV is more volatile than VTSMX) but the diversificaiton effect more than offsets it.

Image

In such a case, you're effectively being paid to take less portfolio risk. This case also highlights the importance of treating diversification and de-risking as distinct processes: this portfolio with EDV is the most diversified one I've considered in this post but it is NOT the lowest risk portfolio.
Last edited by vineviz on Tue Jul 09, 2019 6:26 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
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Tue Jul 09, 2019 6:24 pm

hdas wrote:
Tue Jul 09, 2019 6:09 pm
vineviz wrote:
Tue Jul 09, 2019 5:09 pm

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.
Nice post, however, the tenet you mention is not true or at least heavily challenged by the existence of the Low Volatility anomaly. Cheers :greedy
The principles here could be extended to a world with multidimensional risk, a world in which low volatility is a explainable as a risk factor instead of as an "anomaly". Such an extension would be above my pay grade, however.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Dottie57
Posts: 6726
Joined: Thu May 19, 2016 5:43 pm

Re: De-Risking and Diversification aren't the same thing.

Post by Dottie57 » Tue Jul 09, 2019 6:26 pm

Thank you for explanation above.

columbia
Posts: 1830
Joined: Tue Aug 27, 2013 5:30 am

Re: De-Risking and Diversification aren't the same thing.

Post by columbia » Tue Jul 09, 2019 6:39 pm

I’ve switched to a variant of the Larry Portfolio: 70% treasuries + TIAA Traditional for the safe part. I sleep like a baby and let it ride with the other 30%

That 30% is probably not up to standard for the diversity police; oh well.

User avatar
9-5 Suited
Posts: 411
Joined: Thu Jun 23, 2016 12:14 pm

Re: De-Risking and Diversification aren't the same thing.

Post by 9-5 Suited » Tue Jul 09, 2019 6:56 pm

Thanks for the great original content. Appreciate you taking the time to make this post with clear explanation. Now if only there were another half dozen asset classes like equities out there with similar expected return and minimal correlation :)

Ferdinand2014
Posts: 593
Joined: Mon Dec 17, 2018 6:49 pm

Re: De-Risking and Diversification aren't the same thing.

Post by Ferdinand2014 » Tue Jul 09, 2019 7:57 pm

vineviz wrote:
Tue Jul 09, 2019 5:09 pm
I've previously pointed out that investors sometimes confuse diversification with a different concept known as "de-risking". The explicit goal of de-risking a portfolio is to lower its volatility. And just as diversification doesn't always de-risk a portfolio, de-risking doesn't always improve the diversification of portfolio.

Diversification is the process of combining assets that have less-than-perfectly-positive correlation in order to spread the risk of the portfolio across multiple independent sources of risk and return. Diversification may reduce the risk of the portfolio, but need not do so.

De-risking is the process of adding low-volatility assets to a portfolio in order to reduce the volatility of the portfolio.1 De-risking may improve the diversification of the portfolio, but need not do so.

The distinction between "diversification" and "de-risking" is important to understand because diversification effects provide investors with the proverbial free lunch whereas de-risking does not. Imprecise or confused thinking can lead investors to make sub-optimal portfolio choices, especially when it comes to allocating between stocks and bonds.

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.

By exploiting the properties of covariance, investors can combine not-totally-correlated risky assets into a portfolio which has a volatility that is less than the weighted average volatility of the two assets but with a return that is equal to the weighted average return of the assets.

Same return + less volatility = free lunch.

You can see that here, as two hypothetical uncorrelated (i.e. zero correlation) assets with the same return and volatility are combined at various ratios the volatility decreases as the portfolio approaches the balance point but the return stays constant. This is the "free lunch" benefit of diversification.

Image

In the de-risking case, though, where asset B has lower risk and lower return you can see that lower volatility is associated with lower return. Less risk, but no (or much less) of the "free lunch effect".

Image

It is this second tradeoff that many investors make when they build so-called "balanced" portfolios using low-risk instruments like cash, CDs, money markets, short-term bonds, or total bond market funds. They are sacrificing the "free lunch effect" by accepting less return for a given amount of risk.

The amount of this loss can be quantified and decomposed into two components: a "diversification effect" and a "de-risking effect". Investors are compensated for the former but not for the latter.

To illustrate, I constructed three different 60/40 balanced portfolios using Vanguard Total Stock Market Index (VTSMX) and cash, Vanguard Total Bond Market Index (VBMFX), and Vanguard Long-Term Treasury (VUSTX) from 1992 to present.

Over this period Vanguard Total Stock Market Index had an annualized volatility of 14.5% whereas each of the 60/40 portfolios had similar volatilities of 8.7% to 8.8%.

Image

You can see that the 60/40 portfolio using cash resulted almost entirely from de-risking and enjoyed virtually no diversification benefit.

At the other end, the 60/40 portfolio using long-term Treasury bonds benefited greatly from diversification effects and much less from de-risking effects.

In all three cases the total portfolio risk reduction was incredibly similar, but in the high-diversification case the investor reduced idiosyncratic risk (i.e. uncompensated risk) by a much greater amount. As a consequence, their "free lunch effect" was much greater and manifests as higher expected returns for a portfolio with the same overall expected volatility. This is a much more efficient portfolio.

Edit: Salient Partners published a white paper entitled The Free Lunch Effect that inspired this post. Check it out at http://www.salientpartners.com/wp-conte ... effect.pdf (just don't think I'm endorsing their attempt to sell you commodity funds, MLPs or hedge funds).




1 For reasons of convenience and clarity, I'm using the conventional practice of using the words "risk" and "volatility" interchangeably in this particular thread.
Thank you for this post. Will the results of' 'diversification effect' and 'de-risking' vary over time due to correlation and volatility and therefore the ideal portfolio with lowest volatility for given return be known only in hindsight?
“You only find out who is swimming naked when the tide goes out.“ — Warren Buffett

User avatar
whodidntante
Posts: 6154
Joined: Thu Jan 21, 2016 11:11 pm
Location: outside the echo chamber

Re: De-Risking and Diversification aren't the same thing.

Post by whodidntante » Tue Jul 09, 2019 8:19 pm

9-5 Suited wrote:
Tue Jul 09, 2019 6:56 pm
Thanks for the great original content. Appreciate you taking the time to make this post with clear explanation. Now if only there were another half dozen asset classes like equities out there with similar expected return and minimal correlation :)
Would you settle for similar risk adjusted return? That market factor is a wild ride, and bucks many people off the bull. :happy

Nice post, OP.

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Tue Jul 09, 2019 8:23 pm

Ferdinand2014 wrote:
Tue Jul 09, 2019 7:57 pm
Thank you for this post. Will the results of' 'diversification effect' and 'de-risking' vary over time due to correlation and volatility and therefore the ideal portfolio with lowest volatility for given return be known only in hindsight?
Good question. While the ideal portfolio allocation probably will vary a bit through time and/or only reveal itself with certainty in hindsight, I don't think it's necessary to strive for the ideal portfolio but rather to aim to simply make better portfolio choices when we can.

For instance, the absolute volatility of the bond funds I mentioned (e.g. VBMFX and VUSTX) might vary somewhat over time and so might the correlation with the equity fund (e.g. VTMSX). You can see that here in the rolling correlations:

Image

However, the fundamental properties of the assets means that you can count on VUSTX being less correlated with VTMSX than VBMFX is (you can see this in the graph) and also having higher volatility than VBMFX (it'd be VERY shocking if it were otherwise). As a result, you can reasonably expect that VUSTX will always be a better portfolio diversified than VBMFX is regardless of any subtle changes in the summary statistics.

I think the important take away is for investors to be thoughtful about the choices they make, knowing that otherwise they may be inadvertently wasting a "free" opportunity to get better returns and/or lower portfolio volatility.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

JBTX
Posts: 5332
Joined: Wed Jul 26, 2017 12:46 pm

Re: De-Risking and Diversification aren't the same thing.

Post by JBTX » Tue Jul 09, 2019 9:23 pm

vineviz wrote:
Tue Jul 09, 2019 5:09 pm
I've previously pointed out that investors sometimes confuse diversification with a different concept known as "de-risking". The explicit goal of de-risking a portfolio is to lower its volatility. And just as diversification doesn't always de-risk a portfolio, de-risking doesn't always improve the diversification of portfolio.

Diversification is the process of combining assets that have less-than-perfectly-positive correlation in order to spread the risk of the portfolio across multiple independent sources of risk and return. Diversification may reduce the risk of the portfolio, but need not do so.

De-risking is the process of adding low-volatility assets to a portfolio in order to reduce the volatility of the portfolio.1 De-risking may improve the diversification of the portfolio, but need not do so.

The distinction between "diversification" and "de-risking" is important to understand because diversification effects provide investors with the proverbial free lunch whereas de-risking does not. Imprecise or confused thinking can lead investors to make sub-optimal portfolio choices, especially when it comes to allocating between stocks and bonds.

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.

By exploiting the properties of covariance, investors can combine not-totally-correlated risky assets into a portfolio which has a volatility that is less than the weighted average volatility of the two assets but with a return that is equal to the weighted average return of the assets.

Same return + less volatility = free lunch.

You can see that here, as two hypothetical uncorrelated (i.e. zero correlation) assets with the same return and volatility are combined at various ratios the volatility decreases as the portfolio approaches the balance point but the return stays constant. This is the "free lunch" benefit of diversification.

Image

In the de-risking case, though, where asset B has lower risk and lower return you can see that lower volatility is associated with lower return. Less risk, but no (or much less) of the "free lunch effect".

Image

It is this second tradeoff that many investors make when they build so-called "balanced" portfolios using low-risk instruments like cash, CDs, money markets, short-term bonds, or total bond market funds. They are sacrificing the "free lunch effect" by accepting less return for a given amount of risk.

The amount of this loss can be quantified and decomposed into two components: a "diversification effect" and a "de-risking effect". Investors are compensated for the former but not for the latter.

To illustrate, I constructed three different 60/40 balanced portfolios using Vanguard Total Stock Market Index (VTSMX) and cash, Vanguard Total Bond Market Index (VBMFX), and Vanguard Long-Term Treasury (VUSTX) from 1992 to present.

Over this period Vanguard Total Stock Market Index had an annualized volatility of 14.5% whereas each of the 60/40 portfolios had similar volatilities of 8.7% to 8.8%.

Image

You can see that the 60/40 portfolio using cash resulted almost entirely from de-risking and enjoyed virtually no diversification benefit.

At the other end, the 60/40 portfolio using long-term Treasury bonds benefited greatly from diversification effects and much less from de-risking effects.

In all three cases the total portfolio risk reduction was incredibly similar, but in the high-diversification case the investor reduced idiosyncratic risk (i.e. uncompensated risk) by a much greater amount. As a consequence, their "free lunch effect" was much greater and manifests as higher expected returns for a portfolio with the same overall expected volatility. This is a much more efficient portfolio.

Edit: Salient Partners published a white paper entitled The Free Lunch Effect that inspired this post. Check it out at http://www.salientpartners.com/wp-conte ... effect.pdf (just don't think I'm endorsing their attempt to sell you commodity funds, MLPs or hedge funds).




1 For reasons of convenience and clarity, I'm using the conventional practice of using the words "risk" and "volatility" interchangeably in this particular thread.

Am I looking at the above graph correctly - does it say 40% cash, 40% total bond and 40% LT treasuries result in the same volatility and practically the same return?

Or perhaps return is not illustrated at all.

Metx
Posts: 19
Joined: Fri Apr 05, 2019 8:08 pm
Location: WA

Re: De-Risking and Diversification aren't the same thing.

Post by Metx » Tue Jul 09, 2019 9:44 pm

Volatility is the change in price squared, right, so a change upward is just as jarring as change downward. But it sure doesn't feel that way. How could you tell if the volatility you are minimizing is in the unfavorable (downward) direction, and not simply a friction like brake on the main driver of the changes?

Sometimes I feel my international holdings are lowering the volatility of my portfolio by simply not gaining as much as the US portion. :)

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 6:51 am

JBTX wrote:
Tue Jul 09, 2019 9:23 pm
Am I looking at the above graph correctly - does it say 40% cash, 40% total bond and 40% LT treasuries result in the same volatility and practically the same return?

Or perhaps return is not illustrated at all.
Return was not illustrated at all in the original post, but otherwise yes: 40% cash, 40% total bond and 40% LT treasuries result in virtually the same portfolio volatility.

For portfolios with similar overall volatility, we can - as a rule - expect the more diversified portfolio to have a higher return. This is what we see in the historical record, but special care would be needed to untangle the effect of declining interest rates from the more general benefits. Some people will look at the chart below and attribute ALL of the difference in return to falling yields, but that wouldn't be accurate.

Nonetheless, here are the three portfolios. In the summary statistics, you can see that the portfolios had similar volatility but the return of the long-term treasury portfolio (#3) was much greater than the cash portfolio (#1). About 7030% of that difference is what I'd call a "normal" term premium and 3070% I'd attribute to the decline in interest rates.

Image
Last edited by vineviz on Wed Jul 10, 2019 12:08 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
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 6:57 am

Metx wrote:
Tue Jul 09, 2019 9:44 pm
Volatility is the change in price squared, right, so a change upward is just as jarring as change downward. But it sure doesn't feel that way. How could you tell if the volatility you are minimizing is in the unfavorable (downward) direction, and not simply a friction like brake on the main driver of the changes?
Over all but the shortest periods of time, the distribution of returns pretty closely approximates a normal distribution so that shocks upward are about as common as shocks downward. This isn't 100% true, but it's true enough.

To me, I think the important question is whether the investor prefers lower returns over higher returns for any given level of volatility.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
SimpleGift
Posts: 3376
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: De-Risking and Diversification aren't the same thing.

Post by SimpleGift » Wed Jul 10, 2019 7:31 am

The OP would make a good addition to the Boglehead Wiki, I believe.

The distinction between de-risking and diversification is a good one, and adds clarification to the portfolio construction discussion. It would be nice if there was a permanent page in the Wiki where one could reference these concepts from Forum threads.

stan1
Posts: 7346
Joined: Mon Oct 08, 2007 4:35 pm

Re: De-Risking and Diversification aren't the same thing.

Post by stan1 » Wed Jul 10, 2019 7:37 am

How does rebalancing play into capturing the diversification effect? Is a target volatility strategy a necessary part of capturing the benefit?

I don't care for the term "free lunch effect" that comes from the website you linked to. The quote we all know is "there's no such thing as a free lunch (you don't get something for nothing)" so when a random person hears "free lunch" the first thought is "what's the catch". Or there's the "free lunch" offered by some high cost financial advisors around my home if I sit through their sales pitch for an hour. Same problem as "Backdoor Roth Contribution" which sounds like you are doing something sneaky or hidden.

I do appreciate this initiative and I think it will help some of the confusion on the factor/international threads if people read and understand. Frankly I'm going to have to think about it a bit. It's a lot simpler to operate in a single variable world (performance=rate of return) where volatility/risk is a qualitative (willingness, ability, need) factor. In some respects life was easier when pension fund managers and insurance companies took care of all this stuff for us (but I don't see any sign that will ever come back so we are alone managing large to very large balances for decades). Every man and woman for themself.
Last edited by stan1 on Wed Jul 10, 2019 7:40 am, edited 1 time in total.

User avatar
Sandtrap
Posts: 8305
Joined: Sat Nov 26, 2016 6:32 pm
Location: Hawaii No Ka Oi , N. Arizona

Re: De-Risking and Diversification aren't the same thing.

Post by Sandtrap » Wed Jul 10, 2019 7:39 am

SimpleGift wrote:
Wed Jul 10, 2019 7:31 am
The OP would make a good addition to the Boglehead Wiki, I believe.

The distinction between de-risking and diversification is a good one, and adds clarification to the portfolio construction discussion. It would be nice if there was a permanent page in the Wiki where one could reference these concepts from Forum threads.
+1
It would be a good start. Then expand and explore the concepts, etc.
j
Wiki Bogleheads Wiki: Everything You Need to Know

Demisaba
Posts: 38
Joined: Mon May 13, 2019 10:28 am

Re: De-Risking and Diversification aren't the same thing.

Post by Demisaba » Wed Jul 10, 2019 7:42 am

OP has a good point but s/he is restating two results in modern portfolio theory.

In modern portfolio theory, investors make two decisions: (1) the composition of risky assets, i.e. the weights of the risky asset portfolio, that will make up the optimal portfolio. and (2) the composition of this optimal portfolio and a risk free asset.

The first is all about diversification. Investors are trying to get the best combination of risky assets by taking advantage of risk reduction that takes place when you add a risky security that is less than perfectly correlated.

The second one is also a risk reduction, but it has nothing to do with diversification. Here, investors are gauging their risk appetit by balancing between the risk free asset and the optimal portfolio from (1). There is no diversification effect, as the risk of a combination of risky asset and a risk free asset is just a weighted average. The only problem to solve here is to gauge investors risk tolerance (risk aversion).

So, all in all, the problem is solved by first figuring out the optimal portfolio (which, by the way, happens to be the market cap weighted portfolio of all risky assets) and then dividing the investment between this optimal portfolio and the risk free rate. The first step reduces risk through diversification. The second step reduces risk by investing more and more in a risk free asset.

Side note. A result of this excercises is that the first result - the composition of the optimal portfolio - doesn't depend on investor's personal risk tolerance. Everyone will invest in the same optimal portfolio. The difference comes in the second step when dividing the investment between this optimal portfolio and the risk free asset. This part is determined by the investor's risk aversion.

This side note is a well known result in investment called Two Fund Separation Theorem, first coined by Stephen Ross in 1970s.

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 7:46 am

SimpleGift wrote:
Wed Jul 10, 2019 7:31 am
The OP would make a good addition to the Boglehead Wiki, I believe.

The distinction between de-risking and diversification is a good one, and adds clarification to the portfolio construction discussion. It would be nice if there was a permanent page in the Wiki where one could reference these concepts from Forum threads.
Thanks. I actually started this post with a goal of updating the Wiki page on diversification and decided that it was probably smart to float the concept in the forum first to make sure I could explain it in a way that people would find useful.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Nowizard
Posts: 2220
Joined: Tue Oct 23, 2007 5:33 pm

Re: De-Risking and Diversification aren't the same thing.

Post by Nowizard » Wed Jul 10, 2019 7:48 am

Interesting distinction. Our approach to investing, one that has been successful in terms of our goals, is basically one that relies on what we consider to be "common sense" or "pragmatism" more than focusing on beta, Sharpe ratios and other more cognitive approaches that are only vaguely familiar. We do diversify across various areas as recommended, but our "De-Risking" approach, now that the term has been presented, is very simple. Other than diversification, we focus on selecting funds that have low percentile rankings relative to their benchmarks. For example, our current, overall portfolio which includes the typically recommended approaches, plus some deviation from Boglehead suggestions, falls at the 35th percentile. For us, that is interpreted as meaning that we have a fairly conservative portfolio. If the market and, therefore, the portfolio totals are going down, that leads to the adage we use of "Sometimes you make the most by losing the least." It works for us in terms of avoiding panic changes. We give up some potential return, but have limited need for appreciation, though we are currently up over 12% for the year currently. There are certainly ways to critique the approach, as with any other, but we are content with having found a basic approach that has worked for us over many years.

Tim

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 8:29 am

Demisaba wrote:
Wed Jul 10, 2019 7:42 am
So, all in all, the problem is solved by first figuring out the optimal portfolio (which, by the way, happens to be the market cap weighted portfolio of all risky assets) and then dividing the investment between this optimal portfolio and the risk free rate. The first step reduces risk through diversification. The second step reduces risk by investing more and more in a risk free asset.
This is, indeed, a classic description of "modern portfolio theory". It depends on some untenable assumption in the original work, though, and is not consistent with the current state of knowledge.
Last edited by vineviz on Wed Jul 10, 2019 9:26 am, 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

dbr
Posts: 30137
Joined: Sun Mar 04, 2007 9:50 am

Re: De-Risking and Diversification aren't the same thing.

Post by dbr » Wed Jul 10, 2019 9:10 am

Thank you for posting this most important set of concepts. Comments have noted that this material has been basic to portfolio theory for a long time.

Comments have also noted that the results are hypothetical to certain assumed data for returns, volatility, and correlations, all of which still need to be estimated and are not constant over time.

It is still the case that each individual investor has to decide how well the hoped for return and the uncertainty in getting that return will meet his objectives.

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 9:28 am

dbr wrote:
Wed Jul 10, 2019 9:10 am
It is still the case that each individual investor has to decide how well the hoped for return and the uncertainty in getting that return will meet his objectives.
I totally agree. My view is that investors are much more likely to make better choices when they understand the tradeoffs they are making than when those tradeoffs are ignored or concealed.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
HomerJ
Posts: 13219
Joined: Fri Jun 06, 2008 12:50 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HomerJ » Wed Jul 10, 2019 9:39 am

vineviz wrote:
Tue Jul 09, 2019 8:23 pm
Ferdinand2014 wrote:
Tue Jul 09, 2019 7:57 pm
Thank you for this post. Will the results of' 'diversification effect' and 'de-risking' vary over time due to correlation and volatility and therefore the ideal portfolio with lowest volatility for given return be known only in hindsight?
Good question. While the ideal portfolio allocation probably will vary a bit through time and/or only reveal itself with certainty in hindsight, I don't think it's necessary to strive for the ideal portfolio but rather to aim to simply make better portfolio choices when we can.

For instance, the absolute volatility of the bond funds I mentioned (e.g. VBMFX and VUSTX) might vary somewhat over time and so might the correlation with the equity fund (e.g. VTMSX). You can see that here in the rolling correlations:

Image

However, the fundamental properties of the assets means that you can count on VUSTX being less correlated with VTMSX than VBMFX is (you can see this in the graph) and also having higher volatility than VBMFX (it'd be VERY shocking if it were otherwise). As a result, you can reasonably expect that VUSTX will always be a better portfolio diversified than VBMFX is regardless of any subtle changes in the summary statistics.
What's the chart look like in the late 70s and early 80s when interest rates and inflation were going through the roof?

Would using long treasury bonds been a "better portfolio" back then?
The J stands for Jay

User avatar
HomerJ
Posts: 13219
Joined: Fri Jun 06, 2008 12:50 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HomerJ » Wed Jul 10, 2019 9:45 am

vineviz wrote:
Wed Jul 10, 2019 6:51 am
Nonetheless, here are the three portfolios. In the summary statistics, you can see that the portfolios had similar volatility but the return of the long-term treasury portfolio (#3) was much greater than the cash portfolio (#1). About 70% of that difference is what I'd call a "normal" term premium and 30% I'd attribute to the decline in interest rates.

Image
How did you calculate that only 30% can be attributed to the decline in interest rates?

My first thought would be that MAIN reason the long-term treasury portfolio did so much better in the 90s and 2000s was because interest rates were dropping and bonds did very well during that time period.

I mean, after 2008-2009, interest rates dropped quickly, CDs and money-markets were paying near nothing, and existing long-term bonds did very well.

So, of course, the cash portfolio did poorly and the long-term portfolio pulled away.

Again, what happens if the inverse is true in the future and interest rates are rising?
The J stands for Jay

User avatar
hdas
Posts: 1103
Joined: Thu Jun 11, 2015 8:24 am

Re: De-Risking and Diversification aren't the same thing.

Post by hdas » Wed Jul 10, 2019 9:56 am

vineviz wrote:
Tue Jul 09, 2019 6:24 pm
hdas wrote:
Tue Jul 09, 2019 6:09 pm
vineviz wrote:
Tue Jul 09, 2019 5:09 pm

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.
Nice post, however, the tenet you mention is not true or at least heavily challenged by the existence of the Low Volatility anomaly. Cheers :greedy
The principles here could be extended to a world with multidimensional risk, a world in which low volatility is a explainable as a risk factor instead of as an "anomaly". Such an extension would be above my pay grade, however.
Vineviz,
Your example is a good case of the anomaly at work. People looking for more diversification/protection and perhaps return gravitate to something like EDV. But you and I know this is a sub optimal option mostly imposed by constrains on leverage. Using a combination of futures contracts you can efficiently achieve something that would resemble a leverage position on IEF (iShares 7-10 Year Treasury Bond ETF) that matches the volatility in EDV. This simple change will give you significant better returns (+2-3% additional CAGR) with equal diversification benefits. Cheers :greedy

PS. Nevermind the people using TMF in Hedgefundie's clan, however I understand the logistics and leverage constrains at play.
"whenever there is a randomized way of doing something, then there is a nonrandomized way that delivers better performance but requires more thought" ET Jaynes

robertmcd
Posts: 511
Joined: Tue Aug 09, 2016 9:06 am

Re: De-Risking and Diversification aren't the same thing.

Post by robertmcd » Wed Jul 10, 2019 10:16 am

hdas wrote:
Wed Jul 10, 2019 9:56 am
vineviz wrote:
Tue Jul 09, 2019 6:24 pm
hdas wrote:
Tue Jul 09, 2019 6:09 pm
vineviz wrote:
Tue Jul 09, 2019 5:09 pm

One of the tenets of modern portfolio theory is that investors demand compensation for shouldering undiversifiable (i.e. systematic) risks. The more systematic risk the investor bears, the higher the expected return must be.
Nice post, however, the tenet you mention is not true or at least heavily challenged by the existence of the Low Volatility anomaly. Cheers :greedy
The principles here could be extended to a world with multidimensional risk, a world in which low volatility is a explainable as a risk factor instead of as an "anomaly". Such an extension would be above my pay grade, however.
Vineviz,
Your example is a good case of the anomaly at work. People looking for more diversification/protection and perhaps return gravitate to something like EDV. But you and I know this is a sub optimal option mostly imposed by constrains on leverage. Using a combination of futures contracts you can efficiently achieve something that would resemble a leverage position on IEF (iShares 7-10 Year Treasury Bond ETF) that matches the volatility in EDV. This simple change will give you significant better returns (+2-3% additional CAGR) with equal diversification benefits. Cheers :greedy

PS. Nevermind the people using TMF in Hedgefundie's clan, however I understand the logistics and leverage constrains at play.
Even better than this is to use 2 yr treasury futures. 10% VTI/ 90% 2 yr treasuries, lever it up 10x.

User avatar
NYCPete
Posts: 734
Joined: Thu Apr 12, 2007 1:24 pm
Location: New York, NY

Re: De-Risking and Diversification aren't the same thing.

Post by NYCPete » Wed Jul 10, 2019 11:34 am

Wow, this is a really interesting post. First one on BHs in years that has me considering reevaluating my IPS and seeing if I need to change things.

A few things my brain is working through as I unpack this...

A) Is this way of looking at things actionable for me, or is it trying to over complicate things for benefits I'm not likely to notice on a year in year out basis?

B) For over 10 years I've practiced Larry Swedroe's approach of "take your risk on the equity side". His books led me to constructing a portfolio where I would reduce the risk I was taking with the safe assets by using (mainly) short term treasuries, and ramping up the risk I was taking on the stock side by weighting to small and value.

Larry's position, at least as I understood it in reading his books, is that small and value have higher returns not because they're a free lunch, but because they're riskier. I've subscribed to that theory in my portfolio, along with all the drawbacks that go with it (tracking error, etc.). I haven't been deterred from this path, despite small value's under-performance in the past decade.

As I currently understand this concept that Vineviz has posted, what I'm doing is more diversification just on the stock side, and the bond side only has de-risking and limited/no diversification benefit. Is this concept only applicable to a portfolio, or can it be applied on the level of just equities or just fixed income? Others please correct me if my understanding is not accurate. Trying to figure this out.

C) The whole point of holding safer assets, for me, is so I don't bail out during a market crash. From my understanding, I'd think I would want the safest types of assets, which means not taking interest rate risk. But this post got me looking at long term treasuries (LTT) during the 2008 market crash. When the stock market was at its worst in 2008, LTT did just fine. At the time I was holding a combo of total bond and STT, and I didn't bail out. If I'd held LTT instead, I don't think I would have wanted to bail any more or less (at least I think). That should mean LTT would pass the bar for why I hold safe assets, which is to not bail when things go south.

D) But....If I moved my safe assets to LTT, what about if interest rates rise? In fact, I've used Vanguard's LTT to explain to others whether rising interest rates really are all that bad. In the mid 90s, the Fed essentially doubled the fed funds rate from the beginning of 1994 to mid 1995 from 3% to 6%. The Vanguard LTT went down in 1994 by close to 8%, and it had gotten back to even in 18 months. Could I live with this? yes. Would it happen this way again? No, because nothing ever happens like it did in the past, and that's where I struggle.

E) Interest rate risk is a bigger deal over longer periods, right? If interest rates rise and continued to rise over a multi year period, LTT wouldn't do well, right? What about in the late 70s to early 80s?

I looked this up.

Long term treasury returns:
1977 -0.25
1978 -1.64
1979 -0.73
1980 -3.13
1981 0.17

(for comparison)
Short term treasury returns:
1977 3.22
1978 2.86
1979 7.38
1980 7.91
1981 11.72

(for more comparison)
Total Bond Market returns:
1977 2.88
1978 1.24
1979 1.78
1980 2.55
1981 6.09

(for comparison on what was going on the other side of the equation)
Total US stock market returns:
1977 -4.39
1978 7.34
1979 22.45
1980 32.62
1981 -3.78

Ugh, that sounds like it wasn't fun to get from your "safe" assets using LTT. Especially because inflation was at double digits at the time. STT seemed to be the way to go. Total bond was a nice split the difference between the two.

Anyway, just thinking out loud, and thought others might be wondering similar things...

Thanks, Vineviz, for an interesting post.

Best,
Peter
To the extent that a fool knows his foolishness, | He may be deemed wise | A fool who considers himself wise | Is indeed a fool. | | Buddha

User avatar
SimpleGift
Posts: 3376
Joined: Tue Feb 08, 2011 3:45 pm
Location: Central Oregon

Re: De-Risking and Diversification aren't the same thing.

Post by SimpleGift » Wed Jul 10, 2019 12:03 pm

NYCPete wrote:
Wed Jul 10, 2019 11:34 am
Ugh, that sounds like it wasn't fun to get from your "safe" assets using LTT. Especially because inflation was at double digits at the time. STT seemed to be the way to go. Total bond was a nice split the difference between the two.
Hello, Pete. The diversification effect of long-term Treasuries in the portfolio is predicated on a negative correlation between stocks and bonds — which has mainly occurred in history during periods of low inflation, like we've seen in recent decades. This is why LTT worked so well in the 2000 and 2008 market meltdowns.

But the stock-bond correlation has often been positive historically during periods of high inflation, such as the 1977-1981 period you examined (see this thread for more discussion). During these high-inflation periods, short-term Treasuries proved to be the best diversifier of stock risk.

Since we don't know what the predominant inflation regime will be in the future (high, low or medium), and thus what the correlation of stock and bonds will be (positive or negative), many investors settle on intermediate-term bonds as a prudent middle path. Just my two cents.
Last edited by SimpleGift on Wed Jul 10, 2019 12:09 pm, edited 2 times in total.

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 12:06 pm

HomerJ wrote:
Wed Jul 10, 2019 9:39 am
What's the chart look like in the late 70s and early 80s when interest rates and inflation were going through the roof?
VUSTX is the oldest extant long-term Treasury bond fund, and it's inception date is 1986. And of course total bond market didn't exist earlier than that either.

That said, all stock/bond correlations were higher in the 1970s and 1980s than they are now however the correlation between total bond and stocks has been almost universally lower than the correlation between long-term Treasuries and stocks. And the variance of total bond has always been lower, so the relative diversification effect has always been as I illustrated it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 12:14 pm

HomerJ wrote:
Wed Jul 10, 2019 9:45 am
How did you calculate that only 30% can be attributed to the decline in interest rates?
I'm sorry, I had a typo in my initial reply. It should have said 70% attributable to the decline in yields and 30% to the "normal" term premium.

It's an admittedly imperfect estimate: I simply took the difference in returns and backed out the initial yield spread (implicitly assuming that yields didn't change at all from 1992 to now). Because the historical returns are tangential to this topic, I think that estimate is have to suffice for now.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

HEDGEFUNDIE
Posts: 3293
Joined: Sun Oct 22, 2017 2:06 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HEDGEFUNDIE » Wed Jul 10, 2019 12:29 pm

Demisaba wrote:
Wed Jul 10, 2019 7:42 am
OP has a good point but s/he is restating two results in modern portfolio theory.

In modern portfolio theory, investors make two decisions: (1) the composition of risky assets, i.e. the weights of the risky asset portfolio, that will make up the optimal portfolio. and (2) the composition of this optimal portfolio and a risk free asset.

The first is all about diversification. Investors are trying to get the best combination of risky assets by taking advantage of risk reduction that takes place when you add a risky security that is less than perfectly correlated.

The second one is also a risk reduction, but it has nothing to do with diversification. Here, investors are gauging their risk appetit by balancing between the risk free asset and the optimal portfolio from (1). There is no diversification effect, as the risk of a combination of risky asset and a risk free asset is just a weighted average. The only problem to solve here is to gauge investors risk tolerance (risk aversion).

So, all in all, the problem is solved by first figuring out the optimal portfolio (which, by the way, happens to be the market cap weighted portfolio of all risky assets) and then dividing the investment between this optimal portfolio and the risk free rate. The first step reduces risk through diversification. The second step reduces risk by investing more and more in a risk free asset.

Side note. A result of this excercises is that the first result - the composition of the optimal portfolio - doesn't depend on investor's personal risk tolerance. Everyone will invest in the same optimal portfolio. The difference comes in the second step when dividing the investment between this optimal portfolio and the risk free asset. This part is determined by the investor's risk aversion.

This side note is a well known result in investment called Two Fund Separation Theorem, first coined by Stephen Ross in 1970s.
Intermediate, long term, and ultra long term Treasuries are "risky" assets in terms of variance and should be therefore included in the risky assets bucket with stocks. The mean-variance optimal portfolio that results would then look similar to what vineviz is proposing (i.e. it would be heavily weighted to long duration Treasuries).

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 12:36 pm

NYCPete wrote:
Wed Jul 10, 2019 11:34 am
A) Is this way of looking at things actionable for me, or is it trying to over complicate things for benefits I'm not likely to notice on a year in year out basis?
I wouldn't say this discussion, if taken to the actionable conclusion, should make the portfolio any more complicated. It's merely one way to visualize (and quantify) the diversification benefits you give up when you "take your risk on the equity side" instead of taking the more rational approach of controlling your expected risk and return at the portfolio level instead of the asset class or asset level.
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
Larry's position, at least as I understood it in reading his books, is that small and value have higher returns not because they're a free lunch, but because they're riskier.
I'd agree with that assessment for the most part. The similarity between what he advocates on the equity side and the expansion here to the whole portfolio is that the equity risk factors (e.g. market beta, size, and value) are individually more risky than is a diversified combination of those factors. Implicit in the example I presented is the extension of this diversification to include fixed income risk factors (primarily the term, or duration, factor) as well.

Short-term bonds and cash are effectively risk-free assets, so adding them to the portfolio is effective at de-risking the portfolio but NOT at diversifying it. Using longer-term bonds will, over much of the efficient frontier, do both. It's likely that the most risk-averse and/or short-term investors will also need to apply some de-risking in addition to diversification in order to get the portfolio volatility down to an acceptable level, but I don't think that'd normally be a factor until the portfolio was well over 50% in bonds.
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
D) But....If I moved my safe assets to LTT, what about if interest rates rise?
My advice is to focus on the LTT only as they relate to the entire portfolio. If interest rates take a sharp turn upwards, that might hurt long-term Treasuries in the short-run but that would likely correspond to a HUGE surge in stocks.
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
E) Interest rate risk is a bigger deal over longer periods, right?
Not so much. Interest rate risk is a function of the difference between your bond duration and your investment horizon. If your retirement cash flows don't start for another 10 or 15 years, short-term bonds have more interest rate risk to you than long-term bonds do.
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
Ugh, that sounds like it wasn't fun to get from your "safe" assets using LTT.
For the pre-Volcker era, the simulated data we have for long-term bond returns isn't directly comparable to the performance of long-term bonds after that for reasons that have nothing to do with inflation. That's a topic for another thread, just know that long-term bond returns before 1981 aren't relevant for modeling the behavior of long-term bonds after 1981.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 12:40 pm

SimpleGift wrote:
Wed Jul 10, 2019 12:03 pm
The diversification effect of long-term Treasuries in the portfolio is predicated on a negative correlation between stocks and bonds . . . .
It's more accurate to say that the diversification effect is predicated on the correlation being less than 1.0 and the volatility of long-term bonds being greater than 0.0%.

So long as both of those conditions are true (as they have been at every point in history), it just becomes a question of how strong the diversification effect is and not a question of whether the effect exists.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
HomerJ
Posts: 13219
Joined: Fri Jun 06, 2008 12:50 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HomerJ » Wed Jul 10, 2019 12:43 pm

vineviz wrote:
Wed Jul 10, 2019 12:36 pm
My advice is to focus on the LTT only as they relate to the entire portfolio. If interest rates take a sharp turn upwards, that might hurt long-term Treasuries in the short-run but that would likely correspond to a HUGE surge in stocks.
This is not something I've heard before... Didn't the market just go down a week ago because they were hoping for a interest rate CUT and they were worried it wasn't going to happen?

And it just jumped today because the interest rate cut might be back on.

Is that accepted wisdom? That high interest rates are good for the stock market?

Appears to be the opposite.
Last edited by HomerJ on Wed Jul 10, 2019 12:50 pm, edited 2 times in total.
The J stands for Jay

User avatar
HomerJ
Posts: 13219
Joined: Fri Jun 06, 2008 12:50 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HomerJ » Wed Jul 10, 2019 12:45 pm

vineviz wrote:
Wed Jul 10, 2019 12:36 pm
For the pre-Volcker era, the simulated data we have for long-term bond returns isn't directly comparable to the performance of long-term bonds after that for reasons that have nothing to do with inflation. That's a topic for another thread, just know that long-term bond returns before 1981 aren't relevant for modeling the behavior of long-term bonds after 1981.
We need that other thread then.

Are you stating that all our calculations regarding bonds in a portfolio should only look at data from 1981-present?

That's pretty good (sorry, absolutely excellent) period for bonds. Basing all portfolio calculations using data that ONLY contains the best-case scenario for bonds seems to be risky.
Last edited by HomerJ on Wed Jul 10, 2019 12:53 pm, edited 1 time in total.
The J stands for Jay

HEDGEFUNDIE
Posts: 3293
Joined: Sun Oct 22, 2017 2:06 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HEDGEFUNDIE » Wed Jul 10, 2019 12:51 pm

HomerJ wrote:
Wed Jul 10, 2019 12:45 pm
vineviz wrote:
Wed Jul 10, 2019 12:36 pm
For the pre-Volcker era, the simulated data we have for long-term bond returns isn't directly comparable to the performance of long-term bonds after that for reasons that have nothing to do with inflation. That's a topic for another thread, just know that long-term bond returns before 1981 aren't relevant for modeling the behavior of long-term bonds after 1981.
We need that other thread then.

Are you stating that all our calculations regarding bonds in a portfolio should only look at data from 1981-present?
It exists:

viewtopic.php?t=260386#p4150162

HEDGEFUNDIE
Posts: 3293
Joined: Sun Oct 22, 2017 2:06 pm

Re: De-Risking and Diversification aren't the same thing.

Post by HEDGEFUNDIE » Wed Jul 10, 2019 12:52 pm

HomerJ wrote:
Wed Jul 10, 2019 12:43 pm
vineviz wrote:
Wed Jul 10, 2019 12:36 pm
My advice is to focus on the LTT only as they relate to the entire portfolio. If interest rates take a sharp turn upwards, that might hurt long-term Treasuries in the short-run but that would likely correspond to a HUGE surge in stocks.
This is not something I've heard before... Didn't the market just go down a week ago because they were hoping for a interest rate CUT and they were worried it wasn't going to happen?

And it just jumped today because the interest rate cut might be back on.

Is that accepted wisdom? That high interest rates are good for the stock market?

Appears to be the opposite.
It may not be accepted wisdom but it is what the facts show:

https://advisors.vanguard.com/VGApp/iip ... PrStkRtrns
The historical research we've done, however, doesn’t show a pattern of falling stock prices during rate-hiking cycles. In fact, hiking regimes often take place when the economy is performing strongly and earnings growth is robust, and therefore stocks tend to perform respectably during those periods.

WS1
Posts: 68
Joined: Wed Jan 07, 2015 3:45 pm

Re: De-Risking and Diversification aren't the same thing.

Post by WS1 » Wed Jul 10, 2019 1:03 pm

I’ve been following all the recent (and not so recent) conversations on role long term treasuries can play.

How do zero coupons fit into this conversation?

The brokerage window of my workplace account has two commission free long term treasury choices, but I’m not even sure if they’re comparable in an accumulator’s portfolio.

SPTL
17.92 years duration
0.06% expense

ZROZ
27.42 years duration
0.15% expense

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 1:10 pm

WS1 wrote:
Wed Jul 10, 2019 1:03 pm
I’ve been following all the recent (and not so recent) conversations on role long term treasuries can play.

How do zero coupons fit into this conversation?

The brokerage window of my workplace account has two commission free long term treasury choices, but I’m not even sure if they’re comparable in an accumulator’s portfolio.

SPTL
17.92 years duration
0.06% expense

ZROZ
27.42 years duration
0.15% expense
I use SPTL as a commission-free choice in my brokerage account, but I'm not a fan of ZROZ because it looks like Pimco doesn't do a very good job of keeping the bid/ask spread reasonable. Vanguard Extended Duration Treasury ETF (EDV) would be a much better choice IMHO.

You might check out this thread ( viewtopic.php?f=10&t=255320 ) or search those tickers on the forum, so as not to get lost in this discussion.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

robertmcd
Posts: 511
Joined: Tue Aug 09, 2016 9:06 am

Re: De-Risking and Diversification aren't the same thing.

Post by robertmcd » Wed Jul 10, 2019 1:21 pm

WS1 wrote:
Wed Jul 10, 2019 1:03 pm
I’ve been following all the recent (and not so recent) conversations on role long term treasuries can play.

How do zero coupons fit into this conversation?

The brokerage window of my workplace account has two commission free long term treasury choices, but I’m not even sure if they’re comparable in an accumulator’s portfolio.

SPTL
17.92 years duration
0.06% expense

ZROZ
27.42 years duration
0.15% expense
Zero coupon treasuries have higher duration than long term treasuries and higher volatility, along with a slightly higher yield. EDV holds 20-30 zero coupons treasuries, ZROZ 25-30 yr zero coupon treasuries. EDV is even more volatile than stocks.

So for a risk parity portfolio of US stocks and treasury bonds, you roughly have the following:

60% VTI & 40% EDV (20-30 zero coupon treasury bonds)
50% VTI & 50% TLT (20-30 treasury bonds)
40% VTI & 60% SPTL (10-30 treasury bonds)
20% VTI & 80% VGIT (3-10 yr treasury bonds)
10% VTI & 90% VGSH (1-3 yr treasury bonds)

I don't like ZROZ compared to EDV due to its higher ER (.15 vs .07) and higher bid/ask spread when trading it (.3% vs .07% spread)

So in my opinion if you hold 60% or higher in US stocks, then your bond allocation should be in EDV. The exception to this is if you are willing to use leverage. You can lever up any part of the curve to build a portfolio with desired CAGR, std dev., sharpe, drawdown, etc. Since shorter term treasuries have a higher sharpe ratio and their yields tend to fall more than long term treasuries in times of distress, the strategies using higher leverage with shorter term bonds have done better than the strategies long term treasuries. This is the strategy Ray Dalio and Bridgewater used.

User avatar
Topic Author
vineviz
Posts: 5149
Joined: Tue May 15, 2018 1:55 pm

Re: De-Risking and Diversification aren't the same thing.

Post by vineviz » Wed Jul 10, 2019 1:35 pm

WS1 wrote:
Wed Jul 10, 2019 1:03 pm
The brokerage window of my workplace account has two commission free long term treasury choices, but I’m not even sure if they’re comparable in an accumulator’s portfolio.
I don't know if you can buy individual bonds through the brokerage window at Fidelity, but if you can and are within 30 years of retirement then routinely buying individual 30-year TIPS (with the intent of holding to maturity) with your bond allocation is reasonable strategy as well.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
alpine_boglehead
Posts: 351
Joined: Fri Feb 17, 2017 9:51 am
Location: Austria

Re: De-Risking and Diversification aren't the same thing.

Post by alpine_boglehead » Wed Jul 10, 2019 2:24 pm

Very interesting post, thanks very much.
9-5 Suited wrote:
Tue Jul 09, 2019 6:56 pm
Thanks for the great original content. Appreciate you taking the time to make this post with clear explanation. Now if only there were another half dozen asset classes like equities out there with similar expected return and minimal correlation :)
It finally made click for me what lies at the basis of all the factor discussions ... trying to find those other "asset classes" with somewhat equity-like returns but at least partially uncorrelated to stocks.

User avatar
NYCPete
Posts: 734
Joined: Thu Apr 12, 2007 1:24 pm
Location: New York, NY

Re: De-Risking and Diversification aren't the same thing.

Post by NYCPete » Wed Jul 10, 2019 2:59 pm

vineviz wrote:
Wed Jul 10, 2019 12:36 pm
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
A) Is this way of looking at things actionable for me, or is it trying to over complicate things for benefits I'm not likely to notice on a year in year out basis?
I wouldn't say this discussion, if taken to the actionable conclusion, should make the portfolio any more complicated. It's merely one way to visualize (and quantify) the diversification benefits you give up when you "take your risk on the equity side" instead of taking the more rational approach of controlling your expected risk and return at the portfolio level instead of the asset class or asset level.

[.....] etc.
Thank you for the detailed replies to my thoughts and questions, Vineviz. This is very helpful.
HEDGEFUNDIE wrote:
Wed Jul 10, 2019 12:51 pm
HomerJ wrote:
Wed Jul 10, 2019 12:45 pm
vineviz wrote:
Wed Jul 10, 2019 12:36 pm
For the pre-Volcker era, the simulated data we have for long-term bond returns isn't directly comparable to the performance of long-term bonds after that for reasons that have nothing to do with inflation. That's a topic for another thread, just know that long-term bond returns before 1981 aren't relevant for modeling the behavior of long-term bonds after 1981.
We need that other thread then.

Are you stating that all our calculations regarding bonds in a portfolio should only look at data from 1981-present?
It exists:

viewtopic.php?t=260386#p4150162
Wow, very interesting. Certainly puts the 1977-1981 historical returns I posted earlier in a different light.

Best,
Peter
To the extent that a fool knows his foolishness, | He may be deemed wise | A fool who considers himself wise | Is indeed a fool. | | Buddha

azanon
Posts: 2560
Joined: Mon Nov 07, 2011 10:34 am

Re: De-Risking and Diversification aren't the same thing.

Post by azanon » Wed Jul 10, 2019 3:40 pm

Great thread I definitely agree with, given that my unlevered risk parity portfolio on this forum makes a best attempt at hitting a home-run where true diversification is concerned. The cornerstone of Ray Dalio's criticism of typical portfolios is that they're not really that diversified. He doesn't specifically say it, but it can be implied that his observation is that most people just de-risk their portfolios to varying extents (and not that much given that most portfolios are dominated by stocks), but don't really focus that much on improving diversification.

User avatar
LadyGeek
Site Admin
Posts: 56460
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: De-Risking and Diversification aren't the same thing.

Post by LadyGeek » Thu Jul 25, 2019 10:15 pm

vineviz wrote:
Wed Jul 10, 2019 7:46 am
SimpleGift wrote:
Wed Jul 10, 2019 7:31 am
The OP would make a good addition to the Boglehead Wiki, I believe.

The distinction between de-risking and diversification is a good one, and adds clarification to the portfolio construction discussion. It would be nice if there was a permanent page in the Wiki where one could reference these concepts from Forum threads.
Thanks. I actually started this post with a goal of updating the Wiki page on diversification and decided that it was probably smart to float the concept in the forum first to make sure I could explain it in a way that people would find useful.
I assume you are referring to: Diversification

It looks like we might need to have a discussion to get consensus on the content. Would it be better to work on a draft page first, then update the "live" page when it's ready? If you want assistance, I can get this started for you.

(Both vineviz and SimpleGift are wiki editors. If anyone has lost their password, please PM me.)
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

bluquark
Posts: 720
Joined: Mon Oct 22, 2018 2:30 pm

Re: De-Risking and Diversification aren't the same thing.

Post by bluquark » Thu Jul 25, 2019 11:07 pm

vineviz wrote:
Wed Jul 10, 2019 12:36 pm
NYCPete wrote:
Wed Jul 10, 2019 11:34 am
D) But....If I moved my safe assets to LTT, what about if interest rates rise?
My advice is to focus on the LTT only as they relate to the entire portfolio. If interest rates take a sharp turn upwards, that might hurt long-term Treasuries in the short-run but that would likely correspond to a HUGE surge in stocks.
Higher interest rates increase companies' borrowing costs and probably even more importantly, drives stock market valuations down in order to maintain a similar risk premium over the risk-free rate (which just improved) -- i.e. the inverse behavior of the "yield chasing" we have been seeing in the current low rate environment.

Maybe your point is that interest rates are likely to be raised as a result of some process which also caused a huge surge in stocks. In that case, I would expect the interest rate rise to dampen the surge. The sum of both effects might result in stocks doing just OK and long-term bonds crashing, no?

User avatar
LadyGeek
Site Admin
Posts: 56460
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: De-Risking and Diversification aren't the same thing.

Post by LadyGeek » Fri Jul 26, 2019 8:46 am

longinvest has flagged the wiki article, as it does not represent the Bogleheads' definition of diversification. See the discussion page: Discussion

The article is now marked with a notice that the content is biased. See: Diversification

The page is in need of an update, as well as incorporating vineviz's comments. I have created a draft page for this purpose. Once we have a consensus, the "live" page can then be updated.

See: User:LadyGeek/Diversification

Comments / corrections / suggestions can be posted here. All wiki editors are welcome to edit the draft page.

(Although longinvest's comments are different from the intent of this thread, I think it's important to keep the discussion together.)
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

schooner
Posts: 268
Joined: Sun Jun 09, 2019 8:27 am

Re: De-Risking and Diversification aren't the same thing.

Post by schooner » Fri Jul 26, 2019 10:25 am

I think the discussion of this topic in the Wiki would lead to more confusion.

I don’t see the term “de-risk” used in any academic or authoritative literature on investing (at least as it’s discussed here).

Instead, it appears to be a term used to advocate various “alternative” investment strategies (commodities, liquid-alts etc.). There’s nothing wrong with this, but I’m not sure how this fits with Bogleheads’ philosophy.

Here’s one of many articles I found through a simple Google search. Most of the authors appear to operate funds within the space:

https://www.etf.com/sections/etf-strate ... -portfolio

And I agree with HomerJ re the statistical evidence.

Just my two cents.

Post Reply