First 20% of bonds in long-term Treasuries

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Sat Aug 10, 2019 4:23 pm

oldzey wrote:
Sat Aug 10, 2019 4:19 pm
Fortunately, I have access to TIAA Traditional, which I use for my entire fixed income allocation (20% of my portfolio).
The liquid version of TIAA Trad. is still very appealing, but the 'illiquid' version I have access to pays the same 3%, so there's very little reason to own it instead other than the possibility of higher rates in the future due to the 'vintage' issue.
“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
Doc
Posts: 9176
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: First 20% of bonds in long-term Treasuries

Post by Doc » Sat Aug 10, 2019 4:27 pm

willthrill81 wrote:
Sat Aug 10, 2019 2:43 pm
Doc wrote:
Sat Aug 10, 2019 2:38 pm
Long Treasuries are possibly going to be inversely correlated to equities during a stock market crash.
Fixed. :wink:
Long Treasuries are probably going to be inversely correlated to equities during a stock market crash.
:P
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: 9176
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: First 20% of bonds in long-term Treasuries

Post by Doc » Sat Aug 10, 2019 4:44 pm

welderwannabe wrote:
Sat Aug 10, 2019 3:42 pm
Doc wrote:
Sat Aug 10, 2019 2:38 pm
Personally our portfolio is 50:50 with maybe 35% of the total portfolio in Treasuries in actual bills/notes or as Treasury funds. But given our age I no longer use long Treasuries at all and given the current yield curve there is no longer even much intermediate term left.
What the other 15% of your fixed income portfolio in?
Short and intermediate investment grade corporates. (We also have some 12% of our FI portfolio in Pimco Total Return Fund PTTRX which I treat as intermediate term investment grade for want of any real data supplied by Pimco. :annoyed )

For clarification: It's 15% of my total portfolio or 30% of the FI part since I'm 50:50.
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
welderwannabe
Posts: 1090
Joined: Fri Jun 16, 2017 8:32 am

Re: First 20% of bonds in long-term Treasuries

Post by welderwannabe » Sat Aug 10, 2019 5:09 pm

Doc wrote:
Sat Aug 10, 2019 4:44 pm
Short and intermediate investment grade corporates. (We also have some 12% of our FI portfolio in Pimco Total Return Fund PTTRX which I treat as intermediate term investment grade for want of any real data supplied by Pimco. :annoyed )

For clarification: It's 15% of my total portfolio or 30% of the FI part since I'm 50:50.
Thanks for sharing. Im 50% intermediate treasuries, 50% short corporates in tax deferred.

My taxable is an entirely different story as I view that money as having a shorter time horizon for me. Yes, I know its mental accounting. Its how I like it.
I am not an investment professional, but I did stay at a Holiday Inn Express last night.

ThrustVectoring
Posts: 756
Joined: Wed Jul 12, 2017 2:51 pm

Re: First 20% of bonds in long-term Treasuries

Post by ThrustVectoring » Tue Aug 13, 2019 2:22 pm

grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.

Pretty much everything else I can think of ends up being inflation-indexed in some way. Taxes on planned Roth conversions (tax brackets are inflation-indexed), living expenses, property taxes, and so much more. People just don't get into too many personal contracts to pay out a fixed amount of dollars 20+ years in the future.
Current portfolio: 60% VTI / 40% VXUS

grok87
Posts: 8792
Joined: Tue Feb 27, 2007 9:00 pm

Re: First 20% of bonds in long-term Treasuries

Post by grok87 » Tue Aug 13, 2019 7:12 pm

ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.

Pretty much everything else I can think of ends up being inflation-indexed in some way. Taxes on planned Roth conversions (tax brackets are inflation-indexed), living expenses, property taxes, and so much more. People just don't get into too many personal contracts to pay out a fixed amount of dollars 20+ years in the future.
agree.
RIP Mr. Bogle.

User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Tue Aug 13, 2019 10:24 pm

ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.
Despite being very aggressive with our investments, we've been aggressively paying down our mortgage for several years and hope to pay it off next spring. Assuming 12% taxes (estimated marginal tax bracket in retirement), doing so has provided us with a higher return than even long-term Treasuries over the same period. But the real reasons we made the choice to do so were (1) to improve our cash flow once the mortgage is gone, which is even more important to us now than back then and (2) because it really got my wife on board.
“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

ThrustVectoring
Posts: 756
Joined: Wed Jul 12, 2017 2:51 pm

Re: First 20% of bonds in long-term Treasuries

Post by ThrustVectoring » Wed Aug 14, 2019 3:49 am

willthrill81 wrote:
Tue Aug 13, 2019 10:24 pm
ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.
Despite being very aggressive with our investments, we've been aggressively paying down our mortgage for several years and hope to pay it off next spring. Assuming 12% taxes (estimated marginal tax bracket in retirement), doing so has provided us with a higher return than even long-term Treasuries over the same period. But the real reasons we made the choice to do so were (1) to improve our cash flow once the mortgage is gone, which is even more important to us now than back then and (2) because it really got my wife on board.
Of course paying down the mortgage will have a higher return than buying treasuries - you have to pay for the embedded call option somehow, and the interest rate is the only variable at play.

The downside to paying off the mortgage early is that if rates fall while you have a mortgage plus long term treasuries, you get the gains in the treasuries and can refinance the mortgage to the new lower rate. But if you aren't trying anything fancy then that really doesn't matter much at all.
Current portfolio: 60% VTI / 40% VXUS

grok87
Posts: 8792
Joined: Tue Feb 27, 2007 9:00 pm

Re: First 20% of bonds in long-term Treasuries

Post by grok87 » Wed Aug 14, 2019 12:14 pm

ThrustVectoring wrote:
Wed Aug 14, 2019 3:49 am
willthrill81 wrote:
Tue Aug 13, 2019 10:24 pm
ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.
Despite being very aggressive with our investments, we've been aggressively paying down our mortgage for several years and hope to pay it off next spring. Assuming 12% taxes (estimated marginal tax bracket in retirement), doing so has provided us with a higher return than even long-term Treasuries over the same period. But the real reasons we made the choice to do so were (1) to improve our cash flow once the mortgage is gone, which is even more important to us now than back then and (2) because it really got my wife on board.
Of course paying down the mortgage will have a higher return than buying treasuries - you have to pay for the embedded call option somehow, and the interest rate is the only variable at play.

The downside to paying off the mortgage early is that if rates fall while you have a mortgage plus long term treasuries, you get the gains in the treasuries and can refinance the mortgage to the new lower rate. But if you aren't trying anything fancy then that really doesn't matter much at all.
Right but then you have to pay capital gains tax on the treasuries
RIP Mr. Bogle.

User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Wed Aug 14, 2019 12:20 pm

grok87 wrote:
Wed Aug 14, 2019 12:14 pm
ThrustVectoring wrote:
Wed Aug 14, 2019 3:49 am
willthrill81 wrote:
Tue Aug 13, 2019 10:24 pm
ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.
Despite being very aggressive with our investments, we've been aggressively paying down our mortgage for several years and hope to pay it off next spring. Assuming 12% taxes (estimated marginal tax bracket in retirement), doing so has provided us with a higher return than even long-term Treasuries over the same period. But the real reasons we made the choice to do so were (1) to improve our cash flow once the mortgage is gone, which is even more important to us now than back then and (2) because it really got my wife on board.
Of course paying down the mortgage will have a higher return than buying treasuries - you have to pay for the embedded call option somehow, and the interest rate is the only variable at play.

The downside to paying off the mortgage early is that if rates fall while you have a mortgage plus long term treasuries, you get the gains in the treasuries and can refinance the mortgage to the new lower rate. But if you aren't trying anything fancy then that really doesn't matter much at all.
Right but then you have to pay capital gains tax on the treasuries
Or potentially a significantly higher marginal tax rate if pulled from tax-deferred accounts. Even though our mortgage rate is only 3.375%, assuming 12% taxes on bond returns, paying down our mortgage has turned out to be significantly better than bonds of any maturity over the same period.
“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
abuss368
Posts: 15157
Joined: Mon Aug 03, 2009 2:33 pm
Location: Where the water is warm, the drinks are cold, and I don't know the names of the players!

Re: First 20% of bonds in long-term Treasuries

Post by abuss368 » Wed Aug 14, 2019 12:26 pm

grok87 wrote:
Wed Aug 14, 2019 12:14 pm
ThrustVectoring wrote:
Wed Aug 14, 2019 3:49 am
willthrill81 wrote:
Tue Aug 13, 2019 10:24 pm
ThrustVectoring wrote:
Tue Aug 13, 2019 2:22 pm
grok87 wrote:
Thu Aug 08, 2019 4:39 am
the only role i would see for long term nominal treasuries is liability matching. i.e. if one somehow had long term nominal liabilities (a fixed rate mortgage?) or spending needs (hard to think of what that might be?)
Fixed rate mortgages tend to be above long-term bond rates, so typically you're better off paying down the mortgage. It's complicated though, since there's an embedded call option in the mortgage to close out your short bond position at par, so if rates have gone up you may be better off leaving that option open and buying long-term treasuries instead. It's complicated though, so overall I think it's not at all necessary to do - probably you just want to pay down the mortgage.
Despite being very aggressive with our investments, we've been aggressively paying down our mortgage for several years and hope to pay it off next spring. Assuming 12% taxes (estimated marginal tax bracket in retirement), doing so has provided us with a higher return than even long-term Treasuries over the same period. But the real reasons we made the choice to do so were (1) to improve our cash flow once the mortgage is gone, which is even more important to us now than back then and (2) because it really got my wife on board.
Of course paying down the mortgage will have a higher return than buying treasuries - you have to pay for the embedded call option somehow, and the interest rate is the only variable at play.

The downside to paying off the mortgage early is that if rates fall while you have a mortgage plus long term treasuries, you get the gains in the treasuries and can refinance the mortgage to the new lower rate. But if you aren't trying anything fancy then that really doesn't matter much at all.
Right but then you have to pay capital gains tax on the treasuries
Good point.
John C. Bogle: "Simplicity is the master key to financial success."

Alex GR
Posts: 88
Joined: Mon Jul 31, 2017 9:17 am

Re: First 20% of bonds in long-term Treasuries

Post by Alex GR » Sat Aug 31, 2019 2:12 am

What about a slightly different approach:
For the first 20% of bond portion use BLV, for the rest of it use BIV.
So if your bond portion is 40%, it would be 20% BLV and 20% BIV.
This may alleviate some concerns of locking into low returns expressed by previous (younger) posters

User avatar
coingaroo
Posts: 50
Joined: Fri Apr 26, 2019 11:31 am

Re: First 20% of bonds in long-term Treasuries

Post by coingaroo » Sat Aug 31, 2019 3:38 am

vineviz wrote:
Fri Aug 09, 2019 12:39 pm
Day9 wrote:
Fri Aug 09, 2019 12:22 pm
In 2008 a large allocation to long term treasuries would have prevented large portfolio losses. In this thread and a few other popular ones people are suggesting a large allocation to long term treasuries. Is this "planning for the last war"?
No, this is more akin to general readiness.

The diversification benefits provided by long-term bonds are as close to universal as it gets in investing.
Is there something I'm missing, or does the increased diversification benefit simply come from increased volatility exposure (interest rate risk), and hence greater returns / volatility?

Can you achieve similar levels of diversification by leveraging up shorter term bonds?

Image

schismal
Posts: 122
Joined: Sat Apr 13, 2019 8:53 pm

Re: First 20% of bonds in long-term Treasuries

Post by schismal » Sat Aug 31, 2019 6:45 am

We haven't really touched on tax efficient placement yet for this rule of thumb.

The prevailing recommendation is against holding LTTs in taxable accounts -- let's use EDV as an extreme example. But if one were limited on tax advantaged space, would you alter this recommendation?

User avatar
Topic Author
vineviz
Posts: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sat Aug 31, 2019 6:56 am

coingaroo wrote:
Sat Aug 31, 2019 3:38 am
Is there something I'm missing, or does the increased diversification benefit simply come from increased volatility exposure (interest rate risk), and hence greater returns / volatility?

Can you achieve similar levels of diversification by leveraging up shorter term bonds?
The diversification benefits are indeed coming from a combination of correlation and variance, so STTs or ITTs leveraged up to match the variance and/or duration of LTTs should provide similar benefits diversification benefits. I can't speak to other differences, since I don't use leveraged bonds myself.
"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: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sat Aug 31, 2019 6:59 am

Alex GR wrote:
Sat Aug 31, 2019 2:12 am
What about a slightly different approach:
For the first 20% of bond portion use BLV, for the rest of it use BIV.
So if your bond portion is 40%, it would be 20% BLV and 20% BIV.
This may alleviate some concerns of locking into low returns expressed by previous (younger) posters
BLV is definitely a better diversifier than BIV or BND, though not as effective as long-term Treasuries, so your proposal would probably be an improvement for many investors.

Personally, I'd favor simply lowering the bond allocation and sticking with Treasuries for the bonds. But that's merely a preference.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
coingaroo
Posts: 50
Joined: Fri Apr 26, 2019 11:31 am

Re: First 20% of bonds in long-term Treasuries

Post by coingaroo » Sat Aug 31, 2019 7:42 am

schismal wrote:
Sat Aug 31, 2019 6:45 am


The prevailing recommendation is against holding LTTs in taxable accounts -- let's use EDV as an extreme example. But if one were limited on tax advantaged space, would you alter this recommendation?

You can improve your tax efficiency in taxable accounts, through futures which are 60% LT capital gains, and 40% ST capital gains.

schismal
Posts: 122
Joined: Sat Apr 13, 2019 8:53 pm

Re: First 20% of bonds in long-term Treasuries

Post by schismal » Sat Aug 31, 2019 7:59 am

coingaroo wrote:
Sat Aug 31, 2019 7:42 am
schismal wrote:
Sat Aug 31, 2019 6:45 am


The prevailing recommendation is against holding LTTs in taxable accounts -- let's use EDV as an extreme example. But if one were limited on tax advantaged space, would you alter this recommendation?

You can improve your tax efficiency in taxable accounts, through futures which are 60% LT capital gains, and 40% ST capital gains.
This is one reason why I'm about to pull the trigger on making NTSX a core taxable holding. Its treasury exposure is all in futures.

User avatar
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sat Aug 31, 2019 8:08 am

vineviz wrote:
Thu Aug 08, 2019 3:28 pm
garlandwhizzer wrote:
Thu Aug 08, 2019 2:30 pm
I remember those days quite well. [...] The thrust of this opinion comes from the fact that inflation has basically gone only one way, down, for longer than most investors and many young academics have any memory of.
Leaving all that aside, it's readily apparent that many investors think they remember aspects of that era that upon closer examination turn out not to be true. [...]
Although it is clear to me that garlandwhizzer has a point about recency bias, I ran a few tests with the Simba backtesting spreadsheet, comparing 60% TSM + 40% TBM with the same including 20% LTTs and 20% TBM. And... hm... vineviz may have a point too! :wink:

I looked at various charts, e.g. CAGR, drawdowns, Ulcer index, growth over 30yrs accumulation cycles, SWR over 30yrs withdrawal cycles, etc. I am not going to post a bunch of charts, but the bottomline is that, AT THE PORTFOLIO LEVEL, the second portfolio performed in basically the same way when starting in the 70s while the 2nd portfolio undoubtedly did better when starting more recently. The 2nd portfolio did a little worse when starting in the 50s or 60s, but not by a lot. I'll confess I was expecting a slightly different outcome.

I think this is discarding emotions a little too fast though. Seeing bonds slaughtered by inflation in the 70s was undoubtedly extremely hard to bear for investors during this troubled time, and I doubt that at looking the portfolio level metrics was much comfort. Personally, my (relatively small) position in IT Bonds has very little to do with cold numbers and portfolio-level considerations, it has everything to do with managing my emotions in a time of crisis and there is no way I'll switch that position for LT bonds and their uncertainties. That's just me though.

Still, granted, if one chooses to focus on cold numbers, vineviz's recommendation might be reasonable. Or at least, it was in the past, to a large extent.

User avatar
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sat Aug 31, 2019 10:13 am

PS. as a side note, a Total-Bonds fund may include more long-term bonds than some might perceive... It is truly diversified. See the chart below.

Image

EDIT: clarified chart and text, this is about stated maturities, not effective maturities.
Last edited by siamond on Sun Sep 01, 2019 1:45 pm, edited 1 time in total.

lazyday
Posts: 3397
Joined: Wed Mar 14, 2007 10:27 pm

Re: First 20% of bonds in long-term Treasuries

Post by lazyday » Sat Aug 31, 2019 11:27 am

siamond wrote:
Sat Aug 31, 2019 8:08 am
... hm... vineviz may have a point too! :wink:

I looked at various charts, e.g. CAGR, drawdowns, Ulcer index, growth over 30yrs accumulation cycles, SWR over 30yrs withdrawal cycles, etc. I am not going to post a bunch of charts, but the bottomline is that, AT THE PORTFOLIO LEVEL, the second portfolio performed in basically the same way when starting in the 70s while the 2nd portfolio undoubtedly did better when starting more recently. The 2nd portfolio did a little worse when starting in the 50s or 60s, but not by a lot. I'll confess I was expecting a slightly different outcome.
Thanks, I haven't followed this thread closely and missed vineviz's earlier post that you quoted. I haven't backtested a mid 60s to early 70s retirement myself, and would have expected long bonds to do much worse.

Here's a relevant part of the post you quoted:
vineviz wrote:
Thu Aug 08, 2019 3:28 pm
imagine that A and B were retirees in 1966 instead, starting with $100k in a 60/40 portfolio and withdrawing $330/month adjusted for inflation. Same bonds as before: A is all intermediate Treasuries while B is 20% in long Treasuries and 20% in intermediate Treasuries.

Despite living very near the limits of their SWRs, both investors enjoy their inflation-adjusted income for a full 30 years of retirement. Investor A dies with $194K and Investor B dies with $92k. Investor B came out behind, for sure, but that is despite putting half his bonds in long-term Treasuries during the WORST period in history when one could have done so.

User avatar
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sat Aug 31, 2019 1:04 pm

lazyday wrote:
Sat Aug 31, 2019 11:27 am
Thanks, I haven't followed this thread closely and missed vineviz's earlier post that you quoted. I haven't backtested a mid 60s to early 70s retirement myself, and would have expected long bonds to do much worse.
Well, there is only so much impact a change of 20% of the portfolio from TBM to LTT can have. It's only 20% after all.

Still, I might have been a tad hasty in my previous post discussing 60/40 portfolios. I was only looking at 1955+ while giving a bit more weight to recent years (which made the glass slightly more 'full' than 'empty'), but if we look at the entire known history of the US market, it looks a bit different. Accumulator perspective on the left side, withdrawal perspective on the right side. Click for a larger display of the graph.

Image

There are two disturbing features I hadn't acknowledged enough in the previous post. First, the glass does seem more half-empty than half-full when looking at such extended history. Next, the really bad scenarios were made worse by the LTT-tilted portfolio. Hm. The case weakens, I'm afraid. Now, of course, it all depends on much you weigh the importance of older than more recent data...

User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Sat Aug 31, 2019 1:11 pm

siamond wrote:
Sat Aug 31, 2019 1:04 pm
lazyday wrote:
Sat Aug 31, 2019 11:27 am
Thanks, I haven't followed this thread closely and missed vineviz's earlier post that you quoted. I haven't backtested a mid 60s to early 70s retirement myself, and would have expected long bonds to do much worse.
Well, there is only so much impact a change of 20% of the portfolio from TBM to LTT can have. It's only 20% after all.

Still, I might have been a tad hasty in my previous post discussing 60/40 portfolios. I was only looking at 1955+ while giving a bit more weight to recent years (which made the glass slightly more 'full' than 'empty'), but if we look at the entire known history of the US market, it looks a bit different. Accumulator perspective on the left side, withdrawal perspective on the right side. Click for a larger display of the graph.

Image

There are two disturbing features I hadn't acknowledged enough in the previous post. First, the glass does seem more half-empty than half-full when looking at such extended history. Next, the really bad scenarios were made worse by the LTT-tilted portfolio. Hm. The case weakens, I'm afraid.
Interesting analysis. I wonder how this would change if long-term TIPS were used in lieu of nominal bonds, although we obviously don't have such data.

Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
“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
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sat Aug 31, 2019 1:14 pm

willthrill81 wrote:
Sat Aug 31, 2019 1:11 pm
Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
LOL, I edited my post to add a mitigating sentence at the end "Now, of course, it all depends on much you weigh the importance of older than more recent data...", but you quoted me before I completed my edit! :wink:

User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Sat Aug 31, 2019 1:15 pm

siamond wrote:
Sat Aug 31, 2019 1:14 pm
willthrill81 wrote:
Sat Aug 31, 2019 1:11 pm
Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
LOL, I edited my post to add a mitigating sentence at the end "Now, of course, it all depends on much you weigh the importance of older than more recent data...", but you quoted me before I completed my edit! :wink:
I can be a little quick on the click! :D
“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
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sat Aug 31, 2019 1:19 pm

willthrill81 wrote:
Sat Aug 31, 2019 1:15 pm
I can be a little quick on the click! :D
And my grammar can be quite deficient when I rush to edit a post... "older vs. more recent" might have been better worded... Better stick with the graphs and let people interpret them, I guess. :shock:

Northern Flicker
Posts: 4593
Joined: Fri Apr 10, 2015 12:29 am

Re: First 20% of bonds in long-term Treasuries

Post by Northern Flicker » Sat Aug 31, 2019 1:26 pm

The idea is an easy sell when interest rates just plummeted.
Risk is not a guarantor of return.

User avatar
Horton
Posts: 320
Joined: Mon Jan 21, 2008 3:53 pm

Re: First 20% of bonds in long-term Treasuries

Post by Horton » Sat Aug 31, 2019 2:35 pm

vineviz wrote:
Wed Aug 07, 2019 7:24 pm
There are a number of low-cost ETFs and mutual funds that hold long-term Treasuries, including:
• SPDR Portfolio Long Term Treasury ETF (SPTL)
• Vanguard Long-Term Treasury ETF (VGLT)
• iShares 20+ Year Treasury Bond ETF (TLT)
• Vanguard Extended Duration Treasury ETF (EDV)
• Fidelity Long-Term Treasury Bond Index (FNBGX)
• Vanguard Long-Term Treasury Index (VLGSX)
• T. Rowe Price US Treasury Long-Term Index (PRUUX)
It sure would be nice if there were more LT TIPS funds available. The 0.20% expense ratio for LTPZ seems a bit too much in the current interest rate environment. Thus, if you want LT TIPS you only have a few options:

1. Purchase them at auction or on the secondary market
2. Suck it up and pay the fee for LTPZ
3. Buy a cheaper LTT fund and accept the inflation risk
"You must know that there is nothing higher and stronger and more wholesome and good for life in the future than some good memory, especially a memory of childhood, of home." - Dostoyevsky

User avatar
abuss368
Posts: 15157
Joined: Mon Aug 03, 2009 2:33 pm
Location: Where the water is warm, the drinks are cold, and I don't know the names of the players!

Re: First 20% of bonds in long-term Treasuries

Post by abuss368 » Sat Aug 31, 2019 9:39 pm

Treasuries are fine and David Swensen recommends Treasuries and TIPS.

Vanguard investment experts recommend a two fund approach of Total Bond and Total International Bond.

Any short or intermediate term investment grade bond fund will provide safety and income to a portfolio.
John C. Bogle: "Simplicity is the master key to financial success."

stlutz
Posts: 5349
Joined: Fri Jan 02, 2009 1:08 am

Re: First 20% of bonds in long-term Treasuries

Post by stlutz » Sat Aug 31, 2019 11:16 pm

siamond wrote:
Sat Aug 31, 2019 1:14 pm
willthrill81 wrote:
Sat Aug 31, 2019 1:11 pm
Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
LOL, I edited my post to add a mitigating sentence at the end "Now, of course, it all depends on much you weigh the importance of older than more recent data...", but you quoted me before I completed my edit! :wink:
Increasingly I wonder if we will look back and identify the GFC as another such breakpoint. What used to be "normal" doesn't seem to be coming back.

Lee_WSP
Posts: 818
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: First 20% of bonds in long-term Treasuries

Post by Lee_WSP » Sat Aug 31, 2019 11:42 pm

I've been looking at a lot of scenarios during the "bad bond years" and I'm starting to question if people wouldn't have been better off with zero exposure to bonds or only 20% exposure. Why were bonds such a relatively bad investment from 1955 to 1980? Or were they? Am I off in my thinking?

KyleAAA
Posts: 7372
Joined: Wed Jul 01, 2009 5:35 pm
Contact:

Re: First 20% of bonds in long-term Treasuries

Post by KyleAAA » Sun Sep 01, 2019 12:04 am

Dude2 wrote:
Wed Aug 07, 2019 10:28 pm
No dog in this fight. Do what you want with your money. Yes, I do understand the concept of take the entire portfolio into consideration, i.e. this zigs when that zags.

But, I favor the "take the risk on the equity side" argument because there is such a huge possibility for a tremendous outcome. On the other hand, even long term bonds or junk bonds, whatever the highest yielding animal is, have only a certain level of upside potential. Plus, they are not without risks -- interest rate risk, inflation risk.

Take short term TIPS for instance. That's a pretty low risk animal. It's short enough that interest rate risk isn't significant, and it tracks inflation and unexpected inflation to a very high degree. It is relatively risk-less and also relatively reward-less. I'm not looking for reward on the bond side, that's what the stock side is for. The bond side is to reduce my risk, i.e. the risk of losing everything or the risk of needing money when the money isn't there.

I use bonds to reduce overall portfolio risk, so I start with the most risk-less item and go from there. I do not mix something with certain risks with another thing with certain other risks because I believe the risks are mutually exclusive. Also I am not trying to necessarily reduce overall portfolio volatility, I am looking to reduce overall portfolio risks.
Long term treasuries DO reduce portfolio risk more than total bond. More than riskless tbills, too.

All Seasons
Posts: 124
Joined: Sun Dec 10, 2017 4:14 pm

Re: First 20% of bonds in long-term Treasuries

Post by All Seasons » Sun Sep 01, 2019 3:40 am

This along with all the gold threads in the recent past...

Every day the discussions on this board move closer and closer and closer to realizing how prudent the Permanent Portfolio is as a strategy... but we never quite seem to get there. :mrgreen: :mrgreen: :mrgreen:
The market portfolio is always a legitimate portfolio.

User avatar
Topic Author
vineviz
Posts: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sun Sep 01, 2019 8:12 am

siamond wrote:
Sat Aug 31, 2019 1:14 pm
willthrill81 wrote:
Sat Aug 31, 2019 1:11 pm
Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
LOL, I edited my post to add a mitigating sentence at the end "Now, of course, it all depends on much you weigh the importance of older than more recent data...", but you quoted me before I completed my edit! :wink:
Thanks for running the numbers.

If you’ll forgive a qualitative comment, the cardinal rule of back testing is to never trust a strategy that outperformed in every historical period.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
willthrill81
Posts: 12789
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: First 20% of bonds in long-term Treasuries

Post by willthrill81 » Sun Sep 01, 2019 9:41 am

All Seasons wrote:
Sun Sep 01, 2019 3:40 am
This along with all the gold threads in the recent past...

Every day the discussions on this board move closer and closer and closer to realizing how prudent the Permanent Portfolio is as a strategy... but we never quite seem to get there. :mrgreen: :mrgreen: :mrgreen:
I personally have no problem with the Permanent Portfolio or the Golden Butterfly, especially for those who have largely already 'built' their portfolio and are comfortable dialing back some growth in return for protection.
“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
nisiprius
Advisory Board
Posts: 38974
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: First 20% of bonds in long-term Treasuries

Post by nisiprius » Sun Sep 01, 2019 9:43 am

vineviz wrote:
Wed Aug 07, 2019 7:24 pm
...From my observations, many Bogleheads do a reasonably good job of constructing the equity portion of their portfolio... However, it seems to me that precious few of these same Bogleheads are allocating their fixed income allocations in a manner congruent with modern financial knowledge...
You don't give a reference for what you consider to be "modern financial knowledge." Do you mean, say, The Only Guide to a Winning Bond Strategy You'll Ever Need: The Way Smart Money Preserves Wealth Today by Larry Swedroe and Joseph H. Hempden, 2006? Which is on my to-read list, but as yet unread. If not, what would you suggest?

For starters, it should be noted that Total Bond, perhaps the most-commonly-mentioned bond fund in this forum, has this breakdown:

Image

Depending on definitions and without trying to analyze the details of the "10-20 year" slice, it is roughly 15% long-term bonds. That's much much less than your rule suggests, but it's not zero, either.

I would argue that adjustments of the composition of bond component are much less consequential than adjustments in the composition of the stock component... which, in turn, is much less consequential than adjustment of the stock/bond ratio in the portfolio as a whole.

I'll use the SBBI data sets for large-company stocks, long-term government bonds, and intermediate-term government bonds, as uploaded to PortfolioVisualizer--which then, for some reason, lets me calculate back to 1930 but not earlier.

In each case, portfolio 1 and the blue line will represent the use of all intermediate-term, while portfolio 2 and the red line are the result of following your rule. Source links will not work, but in every case my choices are with time period changed to "month-to-month" and starting date changed to "Jan 1930" and everything else left at default settings. In every case, the period covered is Jan 1930 - Dec 2018.

We start with 80/20 stocks/bonds. Portfolio 1 contains 20% intermediate-term government bonds and portfolio 2 contains 20% long-term government bonds.

Image.

The next is 60/40. Portfolio 1 contains 40% intermediate-term, portfolio 2 contains 20% long-term and 20% intermediate-term.

Image

The next is 40/60. Portfolio 1 contains 60% intermediate-term, portfolio 2 contains 20% long-term bonds and 40% intermediate-term.

Image

In no case did the addition of long-term do any harm, and in every case there is an argument to be made that the difference is in the direction of improvement, but it isn't much. It isn't enough to budge the Sharpe ratio (a measure of risk-adjusted return) after roundoff.

Whether it is worth doing depends on the strength of ones' convictions about the future behavior of these asset classes in relationship to the past.

Although you said at the start that you didn't want to consider it, I think it is clear that differences in stock/bond ratio, e.g. 80/20 versus 60/40 are hugely more consequential than fine-tuning differences in the bond component.
Last edited by nisiprius on Sun Sep 01, 2019 10:42 am, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
nisiprius
Advisory Board
Posts: 38974
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: First 20% of bonds in long-term Treasuries

Post by nisiprius » Sun Sep 01, 2019 10:20 am

willthrill81 wrote:
Sat Aug 31, 2019 1:11 pm
Some here have argued that pre-Volcker data are not representative of how bonds will perform going forward. I'm really not sure if I espouse that belief, but it should likely be mentioned.
It may be worth looking at a foundation paper for factor investing: Eugene F. Fama and Kenneth R. French (1993), Common risk factors in the returns on stocks and bonds, Journal of Financial Economics 33:3-56. They chose to use data from 1963 through 1990. In this paper, they don't really give any explanation for 1963 as the starting point, but refer back to "The Cross-Section of Expected Stock Returns," in which they note that "book values of common equity (COMPUSTAT item 60) is not generally available prior to 1962." So the time period was limited by the stock data available to them. The source for the bond data is "Ibbotson Associates" and presumably goes back to 1926.

In any case, in my skimming through this paper, I cannot find anything suggesting that Fama and French recognized any reason to separate bond behavior pre-Volcker from post-Volcker. Can you?

Another difficulty is that, to me, this plot--

Image

--does not fit a model of a unique Volcker moment, separating a stable, small positive correlation pre-Volcker from a stable negative correlation afterwards.

I think it fits much better with my own notion: "the correlation, like most correlations, has been quite unstable. The long-term average has been close to zero, but it is better described as 0 ± 0.5. There is a two-year dip well below -0.5 right after the global financial crisis--a couple of decades after Volcker, and I think caused by the global financial crisis, not by Volcker."
Last edited by nisiprius on Sun Sep 01, 2019 10:28 am, edited 1 time in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
Doc
Posts: 9176
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: First 20% of bonds in long-term Treasuries

Post by Doc » Sun Sep 01, 2019 10:28 am

nisiprius wrote:
Sun Sep 01, 2019 9:43 am
Do you mean, say, The Only Guide to a Winning Bond Strategy You'll Ever Need: The Way Smart Money Preserves Wealth Today by Larry Swedroe and Joseph H. Hempden, 2006? Which is on my to-read list, but as yet unread.
Since my copy is within hands reach of my computer and is well "dog eared" hear's my very brief synopsis.

Credit quality: AAA and AA only

Duration extension: Extend as long as you get additional 20 bps per year. (DFA recommendation) (Edit)

MBS: None (Negative convexity)

TIPS: Allocation in steps from 0% real yields <1.5% to 75-100% for real yields > 3%. (Table is for 10 yr TIPS vs Short term FI.) Sorry Grok.

Currently this gets me here:
Image

(The very high short T's is because any new cash is going into T-Bills given the current yield curve. But it would agree with Larry's Duration extension criteria.)
Last edited by Doc on Sun Sep 01, 2019 10:35 am, edited 1 time in total.
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
nisiprius
Advisory Board
Posts: 38974
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: First 20% of bonds in long-term Treasuries

Post by nisiprius » Sun Sep 01, 2019 10:30 am

Doc wrote:
Sun Sep 01, 2019 10:28 am
...Duration extension: Extend as long as you get 20 bps per year. (DFA recommendation)...
I.e. an additional 20 bps of yield per additional year of duration?
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
Doc
Posts: 9176
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: First 20% of bonds in long-term Treasuries

Post by Doc » Sun Sep 01, 2019 10:34 am

nisiprius wrote:
Sun Sep 01, 2019 10:30 am
Doc wrote:
Sun Sep 01, 2019 10:28 am
...Duration extension: Extend as long as you get 20 bps per year. (DFA recommendation)...
I.e. an additional 20 bps of yield per additional year of duration?
Yes. Thanks for the comment. I'll edit my post.
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
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sun Sep 01, 2019 11:34 am

nisiprius wrote:
Sun Sep 01, 2019 9:43 am
For starters, it should be noted that Total Bond, perhaps the most-commonly-mentioned bond fund in this forum, has this breakdown:

Image

Depending on definitions and without trying to analyze the details of the "10-20 year" slice, it is roughly 15% long-term bonds. That's much much less than your rule suggests, but it's not zero, either.
Hm. That's weird. The historical chart I posted here is not consistent with the latest breakdown Vanguard provided (which you quoted). I mean, Vanguard regroups categories (e.g. 10-15 and 15-20 grouped as 10-20), but this clearly doesn't match.

The data I used came from historical data series provided by Morningstar, I can't find the current breakdown with the new Morningstar Web site (sigh), but we can find a snapshot on YCharts at: https://ycharts.com/mutual_funds/M:VBMFX, which matches the data I used. Here is a screenshot as I don't think one can use the YCharts link more than a few times.

Image

At first, I thought Vanguard messed up their grouping rules, but on second thought, I noticed that the weighted average of the various buckets does seem to roughly add up to the average effective maturity (~8yrs) in the Vanguard case, while it's completely off with the Morningstar/YCharts data. Plus frankly, VBMFX (Total-Bonds) using 32% 20-30yrs bonds, this doesn't seem credible. WEIRD. :shock:

stlutz
Posts: 5349
Joined: Fri Jan 02, 2009 1:08 am

Re: First 20% of bonds in long-term Treasuries

Post by stlutz » Sun Sep 01, 2019 12:00 pm

siamond wrote:
Sun Sep 01, 2019 11:34 am

Hm. That's weird.
Morningstar is using the stated maturity on the face of the bond, which is problematic in three respects:

a) With mortgage bonds each payment contains principal and interest.
b) Almost all mortgages get paid back before maturity
c) Many corporate bonds have call provisions

The "average" maturity on the websites like Vanguards or iShares takes these into account--the calculations are based on when the bond is likely to get paid back.

You can look at the further by downloading the holdings of AGG off the iShares site: https://www.ishares.com/us/products/239 ... market-etf

User avatar
Topic Author
vineviz
Posts: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sun Sep 01, 2019 12:24 pm

nisiprius wrote:
Sun Sep 01, 2019 9:43 am
You don't give a reference for what you consider to be "modern financial knowledge." Do you mean, say, The Only Guide to a Winning Bond Strategy You'll Ever Need: The Way Smart Money Preserves Wealth Today by Larry Swedroe and Joseph H. Hempden, 2006? Which is on my to-read list, but as yet unread. If not, what would you suggest?
I probably wasn't referencing that books, since I've never read it. I was referring to something much more basic, which is the fundamental observation of modern portfolio theory as advanced by Markowitz et al. and every financial economist since:
Modern portfolio theory argues that an investment's risk and return characteristics should not be viewed alone, but should be evaluated by how the investment affects the overall portfolio's risk and return.
nisiprius wrote:
Sun Sep 01, 2019 9:43 am
I'll use the SBBI data sets for large-company stocks, long-term government bonds, and intermediate-term government bonds, as uploaded to PortfolioVisualizer--which then, for some reason, lets me calculate back to 1930 but not earlier.
"Growth of $10k" charts like this don't reveal anything much of value, one way or the other, about diversification strategies. This kind of CAGR chart is basically a plot of geometric mean, whereas the benefits of diversification are more accurately captured by the harmonic mean (to which SWR is related).
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

stlutz
Posts: 5349
Joined: Fri Jan 02, 2009 1:08 am

Re: First 20% of bonds in long-term Treasuries

Post by stlutz » Sun Sep 01, 2019 12:32 pm

vineviz wrote:
Sun Sep 01, 2019 12:24 pm
"Growth of $10k" charts like this don't reveal anything much of value, one way or the other, about diversification strategies. This kind of CAGR chart is basically a plot of geometric mean, whereas the benefits of diversification are more accurately captured by the harmonic mean (to which SWR is related).
But in the case of nisi's example, that would reduce the [tiny] advantage of the portfolio with LT bonds, not improve it, as the one with LT bonds had higher volatility.

User avatar
siamond
Posts: 4908
Joined: Mon May 28, 2012 5:50 am

Re: First 20% of bonds in long-term Treasuries

Post by siamond » Sun Sep 01, 2019 12:45 pm

stlutz wrote:
Sun Sep 01, 2019 12:00 pm
Morningstar is using the stated maturity on the face of the bond, which is problematic in three respects: [...]
Aaaahhh... Thank you, makes sense.

stlutz
Posts: 5349
Joined: Fri Jan 02, 2009 1:08 am

Re: First 20% of bonds in long-term Treasuries

Post by stlutz » Sun Sep 01, 2019 12:46 pm

A few years back there was a poster here, jerry_lee who pretty much took the exact opposite side of the coin from vineviz, pointing out that the post-1999 timeframe has been very unique and should be considered an anomaly.

e.g.

viewtopic.php?t=99591#p1440660
viewtopic.php?f=10&t=103247#p1498364

Same data; different adjectives.

As for "modern financial knowledge", a bias toward long-term bonds has decidedly not been conventional wisdom. I I generally think of Antti Ilmanen's book, "Expected Returns" as being a representation of "modern" conventional knowledge. That book has a strong bias in favor of short-term bonds because the additional term risk in long-term bonds is not adequately compensated.

The bias toward short/intermediate bonds might be wrong, but it was not something manufactured on Bogleheads.

User avatar
CULater
Posts: 2228
Joined: Sun Nov 13, 2016 10:59 am

Re: First 20% of bonds in long-term Treasuries

Post by CULater » Sun Sep 01, 2019 12:47 pm

There's no question that long bonds, especially LTT, are a lot more valuable as diversifiers when stocks take a hit. The only time I can think of when that wouldn't be true is if there's a serious bout of protracted inflation as in the 1970s, which hurts both stocks and bonds especially long term bonds. So, you probably want a 3rd asset that will hopefully mitigate that double whammy. That would be TIPS, but I'm not sure that it would be a TIPS fund with a duration of 7 years or longer because the term risk is going negatively impact the share value of the TIPS just as it would Nominals. We know that short-term TIPS track the inflation rate more closely, so how about holding a portion in a short-term TIPS instrument such as VTIP? This would be sort of like a Money Market, or T-Bills, component but with the added benefit that it should do a better job of handling inflation. But how much better than just MM or T-Bills (BIL), I'm not sure.
On the internet, nobody knows you're a dog.

User avatar
Topic Author
vineviz
Posts: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sun Sep 01, 2019 12:52 pm

nisiprius wrote:
Sun Sep 01, 2019 10:20 am
--does not fit a model of a unique Volcker moment, separating a stable, small positive correlation pre-Volcker from a stable negative correlation afterwards.
I encourage you to read a white paper that PIMCO published in 2013 if you haven't already.

http://innovawealth.ca/PDFs/2013.Pimco. ... alysis.pdf

They build a model to explain the stock-bond correlation and find that inflation volatility is a key driver of the correlation. For instance:
Correlations become more negative (for all horizons) if we reduce inflation volatility by 50% of its initial value, and generally increase if we increase inflation volatility by 50%, holding all other factors constant (we provide details on all simulation parameters in the appendix). This analysis reveals model-implied correlations are quite sensitive to inflation volatility. If inflation volatility decreases significantly, the correlation should remain negative, even at the two-year horizon. Importantly, it shows some “tail risks” in the hedging effect of bonds on equity risk. If inflation volatility increases significantly, the stock-bond correlation rises to +20% for two-year returns. At this correlation level, bonds would still provide diversification benefits to risk assets, but perhaps not as much as investors are currently assuming in their asset allocation decisions.
But note that this was still not sufficient: they had to further introduce a dummy variable to account for what they describe as a "structural downward shift in the correlation post 1997". A legitimate warning could be raised here about the dangers of overfitting, though there are some plausible explanations for such a shift (the creation of US TIPS in January 1997 and the launch of the Euro currency from 1995 to 1998).

However the problems with the pre-Volcker LTT data series lies not in the correlation of the bond returns with stocks but rather with much more fundamental disturbances in the nature of those securities.

For one thing, the Fed was actively trying so suppress the volatility of long-term rates during this period (versus the current strategy of suppressing inflation volatility). As a result, the term structure for Treasuries was seriously distorted. A second factor was the fact that long-term Treasuries were callable, so the ability of investors to profit from temporary drops in yields was limited.

As a consequence, from 1951 to 1979 the total return of long-term Treasuries was 100bps less per year than the return of T-bills. This is partly explained by the generally rising rates of this period, but that's far from the whole story.

Here are some summary return stats for the term premium (20-year bond minus 30-day bills) covering the three different Fed eras:

Image

I think it should be very apparent that the performance of LTTs relative to t-bills has varied considerably depending on the policy regime. The 1951-1979 era not only was period in which the returns of LTTs was lower: volatility, skewness, and excess kurtosis were also much lower. We rarely observe such large swings in the basic characteristics of other securities over time, which I why I think treasury data pre-1979 should only run under a caution flag.
"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: 5137
Joined: Tue May 15, 2018 1:55 pm

Re: First 20% of bonds in long-term Treasuries

Post by vineviz » Sun Sep 01, 2019 1:05 pm

stlutz wrote:
Sun Sep 01, 2019 12:32 pm
vineviz wrote:
Sun Sep 01, 2019 12:24 pm
"Growth of $10k" charts like this don't reveal anything much of value, one way or the other, about diversification strategies. This kind of CAGR chart is basically a plot of geometric mean, whereas the benefits of diversification are more accurately captured by the harmonic mean (to which SWR is related).
But in the case of nisi's example, that would reduce the [tiny] advantage of the portfolio with LT bonds, not improve it, as the one with LT bonds had higher volatility.
The variables don't combine in that fashion. Substituting LT nominal bonds for half of the TBM allocation in a 60/40 portfolio reduces the SWR for investors who retired before 1979 and improved it for those who retired afterwards. It improved diversification for every cohort I've examined, and obviously modern investors have the option of long-term TIPS in addition to long-term nominal Treasuries.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

User avatar
oldzey
Posts: 1335
Joined: Sun Apr 13, 2014 8:38 pm
Location: Land of Lincoln

Re: First 20% of bonds in long-term Treasuries

Post by oldzey » Sun Sep 01, 2019 1:12 pm

willthrill81 wrote:
Sat Aug 10, 2019 4:23 pm
oldzey wrote:
Sat Aug 10, 2019 4:19 pm
Fortunately, I have access to TIAA Traditional, which I use for my entire fixed income allocation (20% of my portfolio).
The liquid version of TIAA Trad. is still very appealing, but the 'illiquid' version I have access to pays the same 3%, so there's very little reason to own it instead other than the possibility of higher rates in the future due to the 'vintage' issue.
Yes, TIAA Trad is currently at the minimum (3%) accumulation interest rate for both the liquid and illiquid versions. However, since it represents the portion of my portfolio that I cannot afford to lose, I'll gladly take my 3%, with the option to eventually annuitize some or all of it. It also helps me contain my risk-taking to the equity side of my portfolio.
"The broker said the stock was 'poised to move.' Silly me, I thought he meant up." ― Randy Thurman

Post Reply