First 20% of bonds in long-term Treasuries

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Locked
User avatar
Scott S
Posts: 1937
Joined: Mon Nov 24, 2008 2:28 am
Location: building my position

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

Post by Scott S »

Corsair wrote: Sat Apr 03, 2021 8:12 pm“Worst Q1 return for 30-year Treasury since 1919, worst Q1 for IG bonds since 1980, worst Q1 for gold since 1982: BOA”

Ouch :x
Did you mean to post that in the Bonds in free fall thread? :mrgreen:
"Old value investors never die, they just get their fix from rebalancing." -- vineviz
User avatar
aj76er
Posts: 1179
Joined: Tue Dec 01, 2015 10:34 pm
Location: Austin, TX

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

Post by aj76er »

I'm total in agreement that bond duration should match time horizon in order to eliminate interest rate risk.

However, will 20% in LTT really make a meaningful difference to the overall portfolio?

Taking the time period 1981 - 2001, which is arguably the best time period in modern history to be holding LTT, and comparing two portfolios (with U.S. Large Cap Stocks as the equity component):

Code: Select all

                   CAGR        Stdev         Max DD
80/20 (with LTT):  11.29%      12.18%       -39.38%
80/20 (with ITT):  10.84%      12.05%       -39.99%
So 45bps of CAGR with a bit more volatility and only slightly better Max DD. But over that same period LTT in isolation had nearly twice the stddev (~10% vs ~5%), and double the Max DD (~14% vs ~7%).

Looking at the period from 2001-2021 has nearly identical results as above.

So, in summary, if you can tolerate the swings of LTT and rebalance faithfully, you might earn some premium if the future is like the past. Then again, with starting yields in the 2% range (vs 10% range like they were in 80s, or 6% range like they were in 00's), do you expect to be rewarded the same way (as the past) for taking on the extra volatility and inflation risk associated with LTT?
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle
User avatar
LTCM
Posts: 412
Joined: Wed Sep 09, 2020 3:58 am

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

Post by LTCM »

I'm considering adding 10% EDV to my current 100% stock portfolio.

ER: 0.07%
NAV Premium: $0.42 ($127.11 Mkt price)

Is that NAV premium significant? It's an extra 0.33%. I suppose its not significant if it stays that way if/when I need to rebalance. But if it goes to zero or -0.33% it seems significant? Should I just buy 30 year treasuries instead?
55% VUG - 20% VEA - 20% EDV - 5% BNDX
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

aj76er wrote: Sun Apr 04, 2021 2:07 pm I'm total in agreement that bond duration should match time horizon in order to eliminate interest rate risk.

However, will 20% in LTT really make a meaningful difference to the overall portfolio?

Taking the time period 1981 - 2001, which is arguably the best time period in modern history to be holding LTT, and comparing two portfolios (with U.S. Large Cap Stocks as the equity component):

Code: Select all

                   CAGR        Stdev         Max DD
80/20 (with LTT):  11.29%      12.18%       -39.38%
80/20 (with ITT):  10.84%      12.05%       -39.99%
So 45bps of CAGR with a bit more volatility and only slightly better Max DD. But over that same period LTT in isolation had nearly twice the stddev (~10% vs ~5%), and double the Max DD (~14% vs ~7%).

Looking at the period from 2001-2021 has nearly identical results as above.

So, in summary, if you can tolerate the swings of LTT and rebalance faithfully, you might earn some premium if the future is like the past. Then again, with starting yields in the 2% range (vs 10% range like they were in 80s, or 6% range like they were in 00's), do you expect to be rewarded the same way (as the past) for taking on the extra volatility and inflation risk associated with LTT?
If your thesis for buying LTT hinges on past performance, you may be disappointed.

My thesis doesn't.

Even though I benefitted in the past.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
Hector
Posts: 1879
Joined: Fri Dec 24, 2010 1:21 pm
Contact:

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

Post by Hector »

aj76er wrote: Sun Apr 04, 2021 2:07 pm I'm total in agreement that bond duration should match time horizon in order to eliminate interest rate risk.

However, will 20% in LTT really make a meaningful difference to the overall portfolio?

Taking the time period 1981 - 2001, which is arguably the best time period in modern history to be holding LTT, and comparing two portfolios (with U.S. Large Cap Stocks as the equity component):

Code: Select all

                   CAGR        Stdev         Max DD
80/20 (with LTT):  11.29%      12.18%       -39.38%
80/20 (with ITT):  10.84%      12.05%       -39.99%
So 45bps of CAGR with a bit more volatility and only slightly better Max DD. But over that same period LTT in isolation had nearly twice the stddev (~10% vs ~5%), and double the Max DD (~14% vs ~7%).

Looking at the period from 2001-2021 has nearly identical results as above.

So, in summary, if you can tolerate the swings of LTT and rebalance faithfully, you might earn some premium if the future is like the past. Then again, with starting yields in the 2% range (vs 10% range like they were in 80s, or 6% range like they were in 00's), do you expect to be rewarded the same way (as the past) for taking on the extra volatility and inflation risk associated with LTT?
If interest rate changes by 1%, 30 year bond at 2 % interest rate swings a lot more compare to 30 year bond at 10% interest rate.
User avatar
anon_investor
Posts: 15122
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

Hector wrote: Sun Apr 04, 2021 2:28 pm
aj76er wrote: Sun Apr 04, 2021 2:07 pm I'm total in agreement that bond duration should match time horizon in order to eliminate interest rate risk.

However, will 20% in LTT really make a meaningful difference to the overall portfolio?

Taking the time period 1981 - 2001, which is arguably the best time period in modern history to be holding LTT, and comparing two portfolios (with U.S. Large Cap Stocks as the equity component):

Code: Select all

                   CAGR        Stdev         Max DD
80/20 (with LTT):  11.29%      12.18%       -39.38%
80/20 (with ITT):  10.84%      12.05%       -39.99%
So 45bps of CAGR with a bit more volatility and only slightly better Max DD. But over that same period LTT in isolation had nearly twice the stddev (~10% vs ~5%), and double the Max DD (~14% vs ~7%).

Looking at the period from 2001-2021 has nearly identical results as above.

So, in summary, if you can tolerate the swings of LTT and rebalance faithfully, you might earn some premium if the future is like the past. Then again, with starting yields in the 2% range (vs 10% range like they were in 80s, or 6% range like they were in 00's), do you expect to be rewarded the same way (as the past) for taking on the extra volatility and inflation risk associated with LTT?
If interest rate changes by 1%, 30 year bond at 2 % interest rate swings a lot more compare to 30 year bond at 10% interest rate.
When was the last time the 30 year was 10%?
irwinmfletcher
Posts: 58
Joined: Wed Feb 22, 2012 10:55 am

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

Post by irwinmfletcher »

whereskyle wrote: Sun Apr 04, 2021 10:29 am
irwinmfletcher wrote: Sun Apr 04, 2021 9:57 am
whereskyle wrote: Sun Apr 04, 2021 9:41 am
irwinmfletcher wrote: Sun Apr 04, 2021 9:16 am
whereskyle wrote: Sun Apr 04, 2021 7:46 am

Short-term reserves, I consider, an emergency fund (I.e., funds that are NOT invested). First bonds for investment in a high equity portfolio should be long term treasuries because of the recent negative correlation with equities in times of stress. The principle is simply this: your 80%+ equity position is driving your risk and return; why bother with any bonds that don't provide a robust return when equities crash?

Only when you want to actively "derisk" your portfolio should you consider intermediate and short-term fixed-income.

Emergency funds are not part of this equation.
So in other words, don't really put the first 20% of what is not in equities in LTT? That is, some % in short term and then LTT.

I don't keep a separate emergency fund at all (or as I think of it, every $ I have is part of the emergency fund, but some parts of it are safer than others, and every single dollar is "invested"). Then the question is, of the 20% of the portfolio that is not equities, what % is shorter term and what % is LTT?

For example, if one's 20% of non-equities was 3 years of expenses, perhaps split it 50/50 ITT/LTT, or maybe 66.6/33.3.
I would not hold "short-term reserves" in ITTs. Again, short-term reserves, imo, are not "investments." They are "reserves" you have access to when needed, and the only vehicle I'd consider appropriate is a HYS account. Such funds should not be put at risk at all. If someone wants a separate bucket besides an emergency fund, I don't see why they'd consider that something other than part of their portfolio. I would not put money in my portfolio in short-term bonds unless I had a strategic reason to do so. If I'm taking my risk through a large equity allocation and my port has a long time horizon, my first bonds should be LTTs. I don't see why anyone who is not in or near the withdrawal phase would hold short-term bonds unless volatility makes them sick and they're pursuing something like the permanent portfolio or the Larry portfolio (large allocation to ITTs and the rest in high-risk equities like scv and em)
I understand your position, but I don't have much concern with ITT dropping 5 or 10% if interest rates rise. I might be concerned if LTT drops 36% if rates rise 2%. I'm using reserves as short hand for relatively low risk (that is, unlikely to drop 20% in a year). I have just 1 bucket (every $ I have), and trying to decide what % of that bucket not in equities should be LTT. Right now all of that 20% non-equity bucket is in TBM and ITB, so it's just a change of degree. I'm completely comfortable with that mix, but contemplating changing that 20% up among ITT/LTT.
What's your investment horizon?

A great contribution to the forum from Vineviz is distinguishing diversification from derisking. Bonds that don't match your time horizon in that they undershoot it can certainly be effective "deriskers", in that they lower the overall risk of your portfolio. They lower the stomach acid, so to speak, but they don't necessarily provide the most diversification. Diversification is most readily measured by low, zero, or even negative correlation. The lowest correlation asset to equities is Long-term treasuries. Intermediate-term treasuries are not far off, and they can certainly provide diversification while lowering the risk that LTTs provide. But consider whether you should be taking risk off the table by "derisking" rather than investing in different sources of risk by "diversifying." I think ITTs are a fine choice for most portfolios. With my long investing horizon though, I go long.
Time horizon is about 10 years to retirement and about 40 to live if I had to guess. Your point is well taken: I haven't decided which I am trying to do with the 20% (de-risk or diversify), but probably both so may take the cowards route and go 50/50.
whereskyle
Posts: 1911
Joined: Wed Jan 29, 2020 9:29 am

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

Post by whereskyle »

irwinmfletcher wrote: Sun Apr 04, 2021 2:44 pm
whereskyle wrote: Sun Apr 04, 2021 10:29 am
irwinmfletcher wrote: Sun Apr 04, 2021 9:57 am
whereskyle wrote: Sun Apr 04, 2021 9:41 am
irwinmfletcher wrote: Sun Apr 04, 2021 9:16 am

So in other words, don't really put the first 20% of what is not in equities in LTT? That is, some % in short term and then LTT.

I don't keep a separate emergency fund at all (or as I think of it, every $ I have is part of the emergency fund, but some parts of it are safer than others, and every single dollar is "invested"). Then the question is, of the 20% of the portfolio that is not equities, what % is shorter term and what % is LTT?

For example, if one's 20% of non-equities was 3 years of expenses, perhaps split it 50/50 ITT/LTT, or maybe 66.6/33.3.
I would not hold "short-term reserves" in ITTs. Again, short-term reserves, imo, are not "investments." They are "reserves" you have access to when needed, and the only vehicle I'd consider appropriate is a HYS account. Such funds should not be put at risk at all. If someone wants a separate bucket besides an emergency fund, I don't see why they'd consider that something other than part of their portfolio. I would not put money in my portfolio in short-term bonds unless I had a strategic reason to do so. If I'm taking my risk through a large equity allocation and my port has a long time horizon, my first bonds should be LTTs. I don't see why anyone who is not in or near the withdrawal phase would hold short-term bonds unless volatility makes them sick and they're pursuing something like the permanent portfolio or the Larry portfolio (large allocation to ITTs and the rest in high-risk equities like scv and em)
I understand your position, but I don't have much concern with ITT dropping 5 or 10% if interest rates rise. I might be concerned if LTT drops 36% if rates rise 2%. I'm using reserves as short hand for relatively low risk (that is, unlikely to drop 20% in a year). I have just 1 bucket (every $ I have), and trying to decide what % of that bucket not in equities should be LTT. Right now all of that 20% non-equity bucket is in TBM and ITB, so it's just a change of degree. I'm completely comfortable with that mix, but contemplating changing that 20% up among ITT/LTT.
What's your investment horizon?

A great contribution to the forum from Vineviz is distinguishing diversification from derisking. Bonds that don't match your time horizon in that they undershoot it can certainly be effective "deriskers", in that they lower the overall risk of your portfolio. They lower the stomach acid, so to speak, but they don't necessarily provide the most diversification. Diversification is most readily measured by low, zero, or even negative correlation. The lowest correlation asset to equities is Long-term treasuries. Intermediate-term treasuries are not far off, and they can certainly provide diversification while lowering the risk that LTTs provide. But consider whether you should be taking risk off the table by "derisking" rather than investing in different sources of risk by "diversifying." I think ITTs are a fine choice for most portfolios. With my long investing horizon though, I go long.
Time horizon is about 10 years to retirement and about 40 to live if I had to guess. Your point is well taken: I haven't decided which I am trying to do with the 20% (de-risk or diversify), but probably both so may take the cowards route and go 50/50.
10 years to retirement actually sounds like a good time to start thinking ITTs. Another possibility is the barbell approach, in which you might use roughly equal allocations of LTTs and Short-term treasuries, or even short-term TIPS. Watchnerd does this. You can see his portfolio in his signature. He's posted a lot in this thread. The short-term/long-term treasury allocation is a feature of the permanent portfolio and is something I'm considering for the withdrawal phase. Also considering something like the Larry portfolio, where I up the allocation to ITTs and then invest in only high-risk equities.
"I am better off than he is – for he knows nothing and thinks that he knows. I neither know nor think that I know." - Socrates. "Nobody knows nothing." - Jack Bogle
corp_sharecropper
Posts: 590
Joined: Thu Nov 07, 2013 1:36 pm

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

Post by corp_sharecropper »

I've come to the conclusion most bogleheads that can't stand LTT just fundamentally think of "bonds" differently than those of us that specifically focus long duration treasuries. TBM and LTT share some similarities but can serve two entirely different purposes. It's frustrating to read that people assume LTT can't still be hugely beneficial to a portfolio even in a rising rate environment (which it has done before, I'm on vacation and using a phone so you're on your own to research said rising rate regimes). There's also a fundamental disconnect between thinking of bonds held to maturity and keeping a constant maturity range far out (which of course benefits from a different roll down the yield curve).
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

corp_sharecropper wrote: Sun Apr 04, 2021 4:24 pm I've come to the conclusion most bogleheads that can't stand LTT just fundamentally think of "bonds" differently than those of us that specifically focus long duration treasuries. TBM and LTT share some similarities but can serve two entirely different purposes.
I agree.

But to its credit, TBM is a lot easier for people who don't really understand bonds very well.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

watchnerd wrote: Sun Apr 04, 2021 5:13 pm
corp_sharecropper wrote: Sun Apr 04, 2021 4:24 pm I've come to the conclusion most bogleheads that can't stand LTT just fundamentally think of "bonds" differently than those of us that specifically focus long duration treasuries. TBM and LTT share some similarities but can serve two entirely different purposes.
I agree.

But to its credit, TBM is a lot easier for people who don't really understand bonds very well.
True, though there are plenty of good options in between we can recommend. Funds like iShares Core 10+ Year USD Bond ETF (ILTB), Vanguard Long-Term Bond ETF (BLV), or iShares 10-20 Year Treasury Bond ETF (TLH) are much less volatile than EDV or even TLT but act as MUCH more excellent diversifiers than TBM does.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
aj76er
Posts: 1179
Joined: Tue Dec 01, 2015 10:34 pm
Location: Austin, TX

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

Post by aj76er »

watchnerd wrote: Sun Apr 04, 2021 2:22 pm
aj76er wrote: Sun Apr 04, 2021 2:07 pm I'm total in agreement that bond duration should match time horizon in order to eliminate interest rate risk.

However, will 20% in LTT really make a meaningful difference to the overall portfolio?

Taking the time period 1981 - 2001, which is arguably the best time period in modern history to be holding LTT, and comparing two portfolios (with U.S. Large Cap Stocks as the equity component):

Code: Select all

                   CAGR        Stdev         Max DD
80/20 (with LTT):  11.29%      12.18%       -39.38%
80/20 (with ITT):  10.84%      12.05%       -39.99%
So 45bps of CAGR with a bit more volatility and only slightly better Max DD. But over that same period LTT in isolation had nearly twice the stddev (~10% vs ~5%), and double the Max DD (~14% vs ~7%).

Looking at the period from 2001-2021 has nearly identical results as above.

So, in summary, if you can tolerate the swings of LTT and rebalance faithfully, you might earn some premium if the future is like the past. Then again, with starting yields in the 2% range (vs 10% range like they were in 80s, or 6% range like they were in 00's), do you expect to be rewarded the same way (as the past) for taking on the extra volatility and inflation risk associated with LTT?
If your thesis for buying LTT hinges on past performance, you may be disappointed.

My thesis doesn't.

Even though I benefitted in the past.
My thesis was more along the lines of ITT providing virtually the same downside protection as LTT with much less interest rate sensitivity. The extra complication of duration matching along with the stomach churning volatility of LTT may not justify (historically minimal) reward over ITT.
"Buy-and-hold, long-term, all-market-index strategies, implemented at rock-bottom cost, are the surest of all routes to the accumulation of wealth" - John C. Bogle
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

aj76er wrote: Sun Apr 04, 2021 9:38 pm My thesis was more along the lines of ITT providing virtually the same downside protection as LTT with much less interest rate sensitivity. The extra complication of duration matching along with the stomach churning volatility of LTT may not justify (historically minimal) reward over ITT.
Does interest rate sensitivity matter if your duration is matched to your personal timeframes?

As for volatility:

The higher beta gives LTT greater ability to serve as a hedge to high beta stocks.

ITT hasn't been as good of a diversifier vs stocks, in comparison.

It's fine for low equity portfolios, though, where you don't need as much equity risk parity.

The other bummer about ITT:

Intermediate Treasuries currently have a much higher negative real return.

7 YR = -1.08%

vs

20 YR = -0.16%
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
marky2kk
Posts: 51
Joined: Tue Feb 11, 2020 9:53 am

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

Post by marky2kk »

vineviz wrote: Sun Apr 26, 2020 8:19 am
You can see that very young investors should, when considering human capital effects, have shorter portfolio durations. In practice, they accomplish this by having the vast majority of their financial wealth invested in stocks (to which this paper, following Bernanke and Kuttner (2005), assigns a duration of 5).
I only read this thread thoroughly until page 12 or so, but the discussion on equity duration sparked my interest.

Equities are arguably the longest duration asset out there, it is insane to think of them as having short duration (0-2 years was mentioned as well). Most cash-flows of stocks are in the far future. The value of the first ten-years of cash-flows only make up of around 10% of the S&P 500! (you can calculate this number from derivatives called dividend futures, see here https://www.federalreserve.gov/econres/ ... 201014.htm). The other 90% come from cash-flows far, far in the future. Accordingly, stocks are very interest rate sensitive. You can think of stocks as a very, very long-term bond + cash-flow risk.

It is easy to see from the Gordon-Growth model in which P = D / (r-g), where D=dividend/cash-flow, r=interest rate, g=perpetual cash-flow growth rate. Take the derivative of P w.r.t. to r, that is dP/dr, and you will see that dP/dr = -P/D. That is, equity duration in this model is the negative of the price dividend ratio. The price-dividend ratio of the S&P 500 is around 50 right now? So equity duration is 50 years. For growth stocks it may be closer to a 100 years or so. If you don't believe in the Gordon-Growth model, the relationship between equity duration and the price-dividend ratio literally follows from the definition of a return using something called the Campbell-Shiller approximation (see footnote 11 here https://papers.ssrn.com/sol3/papers.cfm ... id=3385579).

As to the main topic of interest, I think everything has been said. My two cents, which repeat what I think Noobvestor was getting after: Duration matching eliminates interest rate risk, yes. LTTs are the safe-asset if I have a long-term nominal liability, yes. If I have a 80/20 portfolio with 80% equities, however, I don't care so much about interest rate risk. I care about hedging cash-flow risk from equities. For that purpose LTTs have been tremendous over the last two decades. Will they be in future? Who knows. Fed Vice-chair Clarida seems to believe so (https://www.federalreserve.gov/newseven ... 91112a.htm): "I, myself, believe that the change in the U.S. monetary policy regime that began in 1979 under Paul Volcker and that was extended by Alan Greenspan in the 1990s very likely contributed to the change in the sign of the correlation between inflation and the output gap" [which drives the correlation between equities and bonds in the Campbell, Viceira & Pflueger paper he refers to before].



PS: On the empirical paper of Bernanke & Kuttner. Without having read the paper, I can hardly believe they would conclude that equity duration is only 5 years (they are too smart for that). I would rather believe that surprises in the fed funds rate measure something else than pure monetary shocks (e.g., perhaps the Fed has superior information about growth prospects). It is then hard to say something about equity duration, that is, how do equity prices react to an increase in the interest rate (risk free or with risk premia, doesn't matter) ALL ELSE EQUAL, because it is not all else equal when the market updates its growth prospects.
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

marky2kk wrote: Mon Apr 05, 2021 5:33 pm
As to the main topic of interest, I think everything has been said. My two cents, which repeat what I think Noobvestor was getting after: Duration matching eliminates interest rate risk, yes. LTTs are the safe-asset if I have a long-term nominal liability, yes. If I have a 80/20 portfolio with 80% equities, however, I don't care so much about interest rate risk. I care about hedging cash-flow risk from equities. For that purpose LTTs have been tremendous over the last two decades. Will they be in future? Who knows. Fed Vice-chair Clarida seems to believe so (https://www.federalreserve.gov/newseven ... 91112a.htm): "I, myself, believe that the change in the U.S. monetary policy regime that began in 1979 under Paul Volcker and that was extended by Alan Greenspan in the 1990s very likely contributed to the change in the sign of the correlation between inflation and the output gap" [which drives the correlation between equities and bonds in the Campbell, Viceira & Pflueger paper he refers to before].
This is an excellent paper.

He summarizes a lot of findings that I had previously read scattered across at least 5+ other papers.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

marky2kk wrote: Mon Apr 05, 2021 5:33 pm
vineviz wrote: Sun Apr 26, 2020 8:19 am
You can see that very young investors should, when considering human capital effects, have shorter portfolio durations. In practice, they accomplish this by having the vast majority of their financial wealth invested in stocks (to which this paper, following Bernanke and Kuttner (2005), assigns a duration of 5).
I only read this thread thoroughly until page 12 or so, but the discussion on equity duration sparked my interest.

Equities are arguably the longest duration asset out there, it is insane to think of them as having short duration (0-2 years was mentioned as well).
That argument has been made, definitely, but the academic consensus has essentially arrived at the conclusion that equity duration is much shorter than the Gordon model once led people to believe.

Because equities do not have fixed cash flows, changes in interest rates affect both the numerator and the denominator in the equation: both the nominal expected flows and the discount rate.

As a result the difference between different duration measures (eg modified duration vs Macaulay duration) are usually substantial for equities but trivial for nominal bonds. This means that interest rate sensitivity must be empirically examined, and all the best studies have empirically estimated equity duration to be of intermediate length: sat, 5 to 10 years.

This is mostly an academic topic. In practical terms, virtually all investors can safely ignore the concept of equity duration whereas bond duration is hyper-relevant.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
marky2kk
Posts: 51
Joined: Tue Feb 11, 2020 9:53 am

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

Post by marky2kk »

That argument has been made, definitely, but the academic consensus has essentially arrived at the conclusion that equity duration is much shorter than the Gordon model once led people to believe.
References? I would like to see a serious academic who concludes that equity duration is rather short. In fact, quite the opposite is the case:

https://papers.ssrn.com/sol3/papers.cfm ... id=3611428 : no equity premium over the last thirty years once you match long-term bond duration (which is still lower than equity duration) to equity duration. All explained by declining interest rates.

As I said, you don't have to rely on Gordon-Growth, the relationship between the price-dividend ratio and equity duration follows from the Campbell-Shiller decomposition. Unless price-dividend ratios are less than 10, which they are not, equity duration is not less than 10 years.

It can be rather easily seen from the definition of duration as well: as you most certainly know, duration is defined as the average time it takes to receive all the cash flows of a bond, weighted by the present value of each of the cash flows. Applied to equities, if equity duration was, say 5 years, most of the value of equities would come from the next 5 years of profits/dividends. So let's take, say, Microsoft's past profits and model them over the next 5 years with some growth rate g. Then discount them back to the present using some discount rate. Whatever we assume for the growth rate g and the discount rate, I guarantee you that the resulting present value of the discounted profits over the next 5 years will be a teeny tiny fraction of Microsoft's total market cap. Again, because Microsoft's cash-flows are in the future, not in the present. So "weighted by the present value of each of the cash flows" will be teeny tine for first 5 years of profits and equity duration can impossibly be less than 5 years.

Anybody that has ever done a discounted cash-flow valuation will understand that most of the value comes from the terminal value part and not from the first couple of cash-flows that you actually rigorously model.
Because equities do not have fixed cash flows, changes in interest rates affect both the numerator and the denominator in the equation: both the nominal expected flows and the discount rate.
Again, not true. Intuition from Gordon-Growth: there is no r in the numerator. Perhaps you are thinking of a case in which high inflation leads both nominal cash-flows and the discount rate to increase. In that case, the interest rate doesn't affect anything---instead, it is affected by inflation.

As a result the difference between different duration measures (eg modified duration vs Macaulay duration) are usually substantial for equities but trivial for nominal bonds. This means that interest rate sensitivity must be empirically examined, and all the best studies have empirically estimated equity duration to be of intermediate length: sat, 5 to 10 years.
I refuse to define equity duration this way. Equity duration is what it is, dP/dr, a change in equity prices in response to a change in the discount rate (assuming a constant discount rate across maturities) all else equal---just as it is defined for bonds. You wouldn't define bond duration as the movement of yields in response to unanticipated shocks in the fed funds rate. If you did, you would often find that long-term bonds have less duration than short-term bonds because true monetary policy shocks fade across the yield curve (empirically, you may sometimes find something else because the Fed may affect term (risk) premia, but this is a different topic).
This is mostly an academic topic. In practical terms, virtually all investors can safely ignore the concept of equity duration whereas bond duration is hyper-relevant.
Well.. not really. I wouldn't oppose someone saying that tech stocks have rallied so heavily during 2020 because they benefit most from decreases in long-term yields and long-term risk premia (hi, fed put) as opposed to having rallied from better profit growth expectations. Tech stocks are extremely long duration assets and as such benefit most from rate declines. Now that rates increase, they are not doing so well---as expected.
cheezit
Posts: 493
Joined: Sat Jul 14, 2018 7:28 pm

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

Post by cheezit »

vineviz wrote: Mon Apr 05, 2021 6:48 pm
marky2kk wrote: Mon Apr 05, 2021 5:33 pm
vineviz wrote: Sun Apr 26, 2020 8:19 am
You can see that very young investors should, when considering human capital effects, have shorter portfolio durations. In practice, they accomplish this by having the vast majority of their financial wealth invested in stocks (to which this paper, following Bernanke and Kuttner (2005), assigns a duration of 5).
I only read this thread thoroughly until page 12 or so, but the discussion on equity duration sparked my interest.

Equities are arguably the longest duration asset out there, it is insane to think of them as having short duration (0-2 years was mentioned as well).
That argument has been made, definitely, but the academic consensus has essentially arrived at the conclusion that equity duration is much shorter than the Gordon model once led people to believe.

Because equities do not have fixed cash flows, changes in interest rates affect both the numerator and the denominator in the equation: both the nominal expected flows and the discount rate.

As a result the difference between different duration measures (eg modified duration vs Macaulay duration) are usually substantial for equities but trivial for nominal bonds. This means that interest rate sensitivity must be empirically examined, and all the best studies have empirically estimated equity duration to be of intermediate length: sat, 5 to 10 years.

This is mostly an academic topic. In practical terms, virtually all investors can safely ignore the concept of equity duration whereas bond duration is hyper-relevant.
If duration can meaningfully be applied to securities other than bonds, what is the difference between duration and rho?
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

marky2kk wrote: Tue Apr 06, 2021 3:03 am
That argument has been made, definitely, but the academic consensus has essentially arrived at the conclusion that equity duration is much shorter than the Gordon model once led people to believe.
References? I would like to see a serious academic who concludes that equity duration is rather short. In fact, quite the opposite is the case:

https://papers.ssrn.com/sol3/papers.cfm ... id=3611428 : no equity premium over the last thirty years once you match long-term bond duration (which is still lower than equity duration) to equity duration. All explained by declining interest rates.
Binsbergen is examining a counterfactual in this paper: if we assume that the equity risk premium is zero, what is the duration of a bond portfolio that would produce the same performance? It's an interesting study, intellectually speaking, but the paper is computing the duration that would be needed to specifically answer the question it asks. Note the author's own caveat:
When computing the duration of the stock market in this paper, I will focus on the expression in Equation 13. While this duration number is also approximately equal to the ceteris paribus change in the value of the stock market for a one percent change in its own discount rate (i.e. a partial derivative), it must also be recognized that dividends, expected growth rates, risk premia and interest rates are determined jointly in equilibrium, leading to offsetting effects on the value of the stock market in response to macro economic shocks. As such, while developing stock duration type measures for the purpose of interest rate risk management purposes may be interesting in their own right, they are not the focus of this paper.
marky2kk wrote: Tue Apr 06, 2021 3:03 amAs I said, you don't have to rely on Gordon-Growth, the relationship between the price-dividend ratio and equity duration follows from the Campbell-Shiller decomposition. Unless price-dividend ratios are less than 10, which they are not, equity duration is not less than 10 years.

It can be rather easily seen from the definition of duration as well: as you most certainly know, duration is defined as the average time it takes to receive all the cash flows of a bond, weighted by the present value of each of the cash flows.
That is ONE definition of duration, but it's not the only one. The distinction I drew earlier between modified duration and Macualay duration is very important here, much more so than when dealing with fixed income assets.

marky2kk wrote: Tue Apr 06, 2021 3:03 am
As a result the difference between different duration measures (eg modified duration vs Macaulay duration) are usually substantial for equities but trivial for nominal bonds. This means that interest rate sensitivity must be empirically examined, and all the best studies have empirically estimated equity duration to be of intermediate length: sat, 5 to 10 years.
I refuse to define equity duration this way. Equity duration is what it is, dP/dr, a change in equity prices in response to a change in the discount rate (assuming a constant discount rate across maturities) all else equal---just as it is defined for bonds.
Refuse if you want, but you'll be parting company with economists who study this topic. Defining equity duration as dP/dr is simultaneously tautological and untestable, because the denominator (dr) is unobservable.

And even if you like the definition of duration that Binsbergen uses ("weighted average time it takes for the asset to return the discounted cash flows to its owner") you have to pay attention to the caveat that gets attached to it: "it must also be recognized that dividends, expected growth rates, risk premia and interest rates are determined jointly in equilibrium, leading to offsetting effects on the value of the stock market in response to macro economic shocks."
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
marky2kk
Posts: 51
Joined: Tue Feb 11, 2020 9:53 am

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

Post by marky2kk »

Refuse if you want, but you'll be parting company with economists who study this topic. Defining equity duration as dP/dr is simultaneously tautological and untestable, because the denominator (dr) is unobservable.
Not sure which economists you mean though, can't be Binsbergen. Just because something is not observable doesn't mean its useless as a concept. The expected equity premium going forward is unobservable.
And even if you like the definition of duration that Binsbergen uses ("weighted average time it takes for the asset to return the discounted cash flows to its owner") you have to pay attention to the caveat that gets attached to it: "it must also be recognized that dividends, expected growth rates, risk premia and interest rates are determined jointly in equilibrium, leading to offsetting effects on the value of the stock market in response to macro economic shocks."
Yes, that definition makes most sense to me. It wouldn't come to my mind to define equity duration as stock price movements in response to changes in short-term yields. It's an interesting question as explored in the Bernanke and Kuttler paper, but it's not what I call duration and I guess here we disagree. Equity prices move in response to the appropriate yields---long-term yields. Stocks are very interest-rate sensitive assets, which is what the concept of duration is economically supposed to capture.

As I said, we also don't define bond duration in response to short-term yield movements. We don't ask: how much will 30-year bonds decline when the one-year yield increases by one percentage point? We ask: how much will 30-year bonds decline when the 30-year yield increases by one percentage point (close to 30%, of course). Similarly, we should ask how much will stocks decline when the 30-year yield increases by one percentage point? And the answer is, all else equal, a lot.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

cheezit wrote: Tue Apr 06, 2021 8:00 am If duration can meaningfully be applied to securities other than bonds, what is the difference between duration and rho?
I don't see the term rho used much outside of options markets, but essentially rho and interest rate duration are the same thing.

But I think it's worth repeating that the concept of equity duration has extremely limited (i.e. virtually no) practical usage.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
imak
Posts: 212
Joined: Fri Mar 22, 2019 5:18 pm
Location: Austin, TX

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

Post by imak »

marky2kk wrote: Mon Apr 05, 2021 5:33 pm As to the main topic of interest, I think everything has been said. My two cents, which repeat what I think Noobvestor was getting after: Duration matching eliminates interest rate risk, yes. LTTs are the safe-asset if I have a long-term nominal liability, yes. If I have a 80/20 portfolio with 80% equities, however, I don't care so much about interest rate risk. I care about hedging cash-flow risk from equities. For that purpose LTTs have been tremendous over the last two decades. Will they be in future? Who knows. Fed Vice-chair Clarida seems to believe so (https://www.federalreserve.gov/newseven ... 91112a.htm): "I, myself, believe that the change in the U.S. monetary policy regime that began in 1979 under Paul Volcker and that was extended by Alan Greenspan in the 1990s very likely contributed to the change in the sign of the correlation between inflation and the output gap" [which drives the correlation between equities and bonds in the Campbell, Viceira & Pflueger paper he refers to before].
Thanks for sharing this reference, quite insightful indeed.
marky2kk
Posts: 51
Joined: Tue Feb 11, 2020 9:53 am

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

Post by marky2kk »

imak wrote: Tue Apr 06, 2021 10:04 pm
marky2kk wrote: Mon Apr 05, 2021 5:33 pm As to the main topic of interest, I think everything has been said. My two cents, which repeat what I think Noobvestor was getting after: Duration matching eliminates interest rate risk, yes. LTTs are the safe-asset if I have a long-term nominal liability, yes. If I have a 80/20 portfolio with 80% equities, however, I don't care so much about interest rate risk. I care about hedging cash-flow risk from equities. For that purpose LTTs have been tremendous over the last two decades. Will they be in future? Who knows. Fed Vice-chair Clarida seems to believe so (https://www.federalreserve.gov/newseven ... 91112a.htm): "I, myself, believe that the change in the U.S. monetary policy regime that began in 1979 under Paul Volcker and that was extended by Alan Greenspan in the 1990s very likely contributed to the change in the sign of the correlation between inflation and the output gap" [which drives the correlation between equities and bonds in the Campbell, Viceira & Pflueger paper he refers to before].
Thanks for sharing this reference, quite insightful indeed.
Yes, a lot of it really depends on the correlation between inflation and the output gap. Interest rate immunization for long-term investors is nice, but the elephant in the room is the 80% equities in your portfolio. If recessions become inflationary again as in the 70s, LTT holders will be f***** (imagine a COVID crash in equities and ten look at your LTTs decline by the same amount :(). If recessions stay deflationary, all good. Intermediate ground: bet on both and diversify across duration, i.e., hold some LTTs and some short-duration assets. Perhaps tilt more towards LTTs if you want to believe the Fed itself and Clarida himself + you get interest rate immunization as long-term investor. I guess that's not the worst bet. As such, I'd probably agree with the rule of thumb in this thread. But there are risks.

Another risk is that In 30 years time the dollar may not be the world's safe haven anymore, who knows, which would also diminish the benefit of long duration dollar-denominated assets in recessions. The US has had SUCH a privilege that the dollar is the safe haven and inflations have been deflationary. We have all been so spoiled. History shows that such privileges don't last forever...
international001
Posts: 2748
Joined: Thu Feb 15, 2018 6:31 pm

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

Post by international001 »

Aren't we here comparing 20% in LTT vs 20% in ITT?

Assume 80% in stocks. The advantages of LTT:

1) Higher volatility
2) Lower correlation with stocks
3) Higher return (over long term)

Even if 3 doesn't come true, because some unexpected monetary policy, 1 and 2 could make for it.
stimulacra
Posts: 1006
Joined: Wed Dec 21, 2016 2:50 pm
Location: Houston

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

Post by stimulacra »

Looking to add EDV to my portfolio.

Currently for bonds I have:

VBTLX (Vanguard Total US Bonds): 9%
VIPSX (Vanguard Inflation Protected Securities): 9%
VICSX (Vanguard Intermediate-Term Corporate Bonds): 8%

As far as I can tell, there is some overlap with VBTLX and VICSX but not too bad.

Would it make sense to totally swap out VBTLX for EDV? That would eliminate any overlap across the three bond holdings.

Age 43, preferred allocation: 70% equities/25% bonds/5% gold
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

stimulacra wrote: Tue Apr 13, 2021 2:17 pm Looking to add EDV to my portfolio.

Currently for bonds I have:

VBTLX (Vanguard Total US Bonds): 9%
VIPSX (Vanguard Inflation Protected Securities): 9%
VICSX (Vanguard Intermediate-Term Corporate Bonds): 8%

As far as I can tell, there is some overlap with VBTLX and VICSX but not too bad.

Would it make sense to totally swap out VBTLX for EDV? That would eliminate any overlap across the three bond holdings.

Age 43, preferred allocation: 70% equities/25% bonds/5% gold
Long Term Treasuries are not drop in replacements for total bond funds.

They behave very differently, especially one with a duration as long as EDV.

Why do you have the desire to swap?

FWIW, I'm not against LTT and hold a fair chunk myself (see sig), but people should understand the trade offs.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
stimulacra
Posts: 1006
Joined: Wed Dec 21, 2016 2:50 pm
Location: Houston

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

Post by stimulacra »

watchnerd wrote: Tue Apr 13, 2021 3:13 pm
stimulacra wrote: Tue Apr 13, 2021 2:17 pm Looking to add EDV to my portfolio.

Currently for bonds I have:

VBTLX (Vanguard Total US Bonds): 9%
VIPSX (Vanguard Inflation Protected Securities): 9%
VICSX (Vanguard Intermediate-Term Corporate Bonds): 8%

As far as I can tell, there is some overlap with VBTLX and VICSX but not too bad.

Would it make sense to totally swap out VBTLX for EDV? That would eliminate any overlap across the three bond holdings.

Age 43, preferred allocation: 70% equities/25% bonds/5% gold
Long Term Treasuries are not drop in replacements for total bond funds.

They behave very differently, especially one with a duration as long as EDV.

Why do you have the desire to swap?

FWIW, I'm not against LTT and hold a fair chunk myself (see sig), but people should understand the trade offs.
1) I'm still 20 years off from retirement.
2) Want an asset non-correlated to equities for rebalancing opportunities between now and 2040.
3) I'm not worried about interest rate hikes in the next 10-15 years.
4) I think volatility on both sides will benefit me in the long run (Last few bear markets I simply treated as rebalancing opportunities).
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

stimulacra wrote: Tue Apr 13, 2021 4:37 pm
watchnerd wrote: Tue Apr 13, 2021 3:13 pm
stimulacra wrote: Tue Apr 13, 2021 2:17 pm Looking to add EDV to my portfolio.

Currently for bonds I have:

VBTLX (Vanguard Total US Bonds): 9%
VIPSX (Vanguard Inflation Protected Securities): 9%
VICSX (Vanguard Intermediate-Term Corporate Bonds): 8%

As far as I can tell, there is some overlap with VBTLX and VICSX but not too bad.

Would it make sense to totally swap out VBTLX for EDV? That would eliminate any overlap across the three bond holdings.

Age 43, preferred allocation: 70% equities/25% bonds/5% gold
Long Term Treasuries are not drop in replacements for total bond funds.

They behave very differently, especially one with a duration as long as EDV.

Why do you have the desire to swap?

FWIW, I'm not against LTT and hold a fair chunk myself (see sig), but people should understand the trade offs.
1) I'm still 20 years off from retirement.
2) Want an asset non-correlated to equities for rebalancing opportunities between now and 2040.
3) I'm not worried about interest rate hikes in the next 10-15 years.
4) I think volatility on both sides will benefit me in the long run (Last few bear markets I simply treated as rebalancing opportunities).
Ok.

So why keep the intermediate corporates at all?

#2 is better served by more Treasuries, less corporates.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
sfmurph
Posts: 178
Joined: Mon Aug 12, 2019 8:15 pm

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

Post by sfmurph »

Also, consider ZROZ as a long-term treasury STRIPS fund. Why? The ER is similar to EDV (0.15 instead of 0.07%), but EDV consistently has short-term capital gains every year, while ZROZ doesn't. ZROZ also has a slightly longer duration (27 years instead of 25) which will be more interesting as a rebalance pair.
stimulacra
Posts: 1006
Joined: Wed Dec 21, 2016 2:50 pm
Location: Houston

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

Post by stimulacra »

watchnerd wrote: Tue Apr 13, 2021 4:59 pm
stimulacra wrote: Tue Apr 13, 2021 4:37 pm
watchnerd wrote: Tue Apr 13, 2021 3:13 pm
stimulacra wrote: Tue Apr 13, 2021 2:17 pm Looking to add EDV to my portfolio.

Currently for bonds I have:

VBTLX (Vanguard Total US Bonds): 9%
VIPSX (Vanguard Inflation Protected Securities): 9%
VICSX (Vanguard Intermediate-Term Corporate Bonds): 8%

As far as I can tell, there is some overlap with VBTLX and VICSX but not too bad.

Would it make sense to totally swap out VBTLX for EDV? That would eliminate any overlap across the three bond holdings.

Age 43, preferred allocation: 70% equities/25% bonds/5% gold
Long Term Treasuries are not drop in replacements for total bond funds.

They behave very differently, especially one with a duration as long as EDV.

Why do you have the desire to swap?

FWIW, I'm not against LTT and hold a fair chunk myself (see sig), but people should understand the trade offs.
1) I'm still 20 years off from retirement.
2) Want an asset non-correlated to equities for rebalancing opportunities between now and 2040.
3) I'm not worried about interest rate hikes in the next 10-15 years.
4) I think volatility on both sides will benefit me in the long run (Last few bear markets I simply treated as rebalancing opportunities).
Ok.

So why keep the intermediate corporates at all?

#2 is better served by more Treasuries, less corporates.
I found them very attractive at the time when I was devising my asset allocation. I liked how corporate bondholders would get paid first before stockholders. Also corporates allowed me to hold a higher percentage of bonds in total without feeling FOMO, otherwise I would hold closer to 15% bonds only.
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

stimulacra wrote: Tue Apr 13, 2021 5:41 pm
I found them very attractive at the time when I was devising my asset allocation. I liked how corporate bondholders would get paid first before stockholders. Also corporates allowed me to hold a higher percentage of bonds in total without feeling FOMO, otherwise I would hold closer to 15% bonds only.
I ditched corporates entirely.

This allows me to concentrate business risk on the equity side and hold a proportionally higher allocation to LTT relative to stocks.

I use a 1/4 ratio, which matches my preference for balancing equity risk with duration risk and gives the whole port a Sharpe ratio I like.
Last edited by watchnerd on Tue Apr 13, 2021 6:47 pm, edited 1 time in total.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

sfmurph wrote: Tue Apr 13, 2021 5:40 pm Also, consider ZROZ as a long-term treasury STRIPS fund. Why? The ER is similar to EDV (0.15 instead of 0.07%), but EDV consistently has short-term capital gains every year, while ZROZ doesn't. ZROZ also has a slightly longer duration (27 years instead of 25) which will be more interesting as a rebalance pair.
GOVZ iShares 25+ Year Treasury STRIPS Bond ETF tracks the same index as ZROZ but with the lower expense of EDV.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
corp_sharecropper
Posts: 590
Joined: Thu Nov 07, 2013 1:36 pm

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

Post by corp_sharecropper »

marky2kk wrote: Thu Apr 08, 2021 4:14 am
imak wrote: Tue Apr 06, 2021 10:04 pm
marky2kk wrote: Mon Apr 05, 2021 5:33 pm As to the main topic of interest, I think everything has been said. My two cents, which repeat what I think Noobvestor was getting after: Duration matching eliminates interest rate risk, yes. LTTs are the safe-asset if I have a long-term nominal liability, yes. If I have a 80/20 portfolio with 80% equities, however, I don't care so much about interest rate risk. I care about hedging cash-flow risk from equities. For that purpose LTTs have been tremendous over the last two decades. Will they be in future? Who knows. Fed Vice-chair Clarida seems to believe so (https://www.federalreserve.gov/newseven ... 91112a.htm): "I, myself, believe that the change in the U.S. monetary policy regime that began in 1979 under Paul Volcker and that was extended by Alan Greenspan in the 1990s very likely contributed to the change in the sign of the correlation between inflation and the output gap" [which drives the correlation between equities and bonds in the Campbell, Viceira & Pflueger paper he refers to before].
Thanks for sharing this reference, quite insightful indeed.
Yes, a lot of it really depends on the correlation between inflation and the output gap. Interest rate immunization for long-term investors is nice, but the elephant in the room is the 80% equities in your portfolio. If recessions become inflationary again as in the 70s, LTT holders will be f***** (imagine a COVID crash in equities and ten look at your LTTs decline by the same amount :(). If recessions stay deflationary, all good. Intermediate ground: bet on both and diversify across duration, i.e., hold some LTTs and some short-duration assets. Perhaps tilt more towards LTTs if you want to believe the Fed itself and Clarida himself + you get interest rate immunization as long-term investor. I guess that's not the worst bet. As such, I'd probably agree with the rule of thumb in this thread. But there are risks.

Another risk is that In 30 years time the dollar may not be the world's safe haven anymore, who knows, which would also diminish the benefit of long duration dollar-denominated assets in recessions. The US has had SUCH a privilege that the dollar is the safe haven and inflations have been deflationary. We have all been so spoiled. History shows that such privileges don't last forever...
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.

Back to your hypothetical, basically something that's different than has been experienced in the recent past (40 or so years), this is why I simply don't understand how so many bogleheads will say "no one knows nuthin', anything can happen, diversify, diversify!" and in the same breath preach (many in a very "evangelical" tone) that diversifying assets like gold, commodities, crypto, etc are outrageous to hold. It's as if they've never even spent the time to look into whether these sorts of assets have worked in the past to create better, more robust, portfolios. How many bogleheads know that even an asset with no internal rate of return or one that shows no appreciable long term growth, can yet still improve a portfolio? My guess is it's the minority, but it's true if the asset has enough relative volatility to the other assets and is not highly correlated. In the sorts of portfolios I personally like, nominal bonds are one of many sleeves, and long duration treasuries are the absolute best asset to play that nominal fixed income role imo.
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

corp_sharecropper wrote: Tue Apr 13, 2021 7:19 pm
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.
99.34% of the risk was from equities, actually, 2008-present.

https://www.portfoliovisualizer.com/bac ... tion2_1=20
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
marky2kk
Posts: 51
Joined: Tue Feb 11, 2020 9:53 am

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

Post by marky2kk »

Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.

Back to your hypothetical, basically something that's different than has been experienced in the recent past (40 or so years), this is why I simply don't understand how so many bogleheads will say "no one knows nuthin', anything can happen, diversify, diversify!" and in the same breath preach (many in a very "evangelical" tone) that diversifying assets like gold, commodities, crypto, etc are outrageous to hold. It's as if they've never even spent the time to look into whether these sorts of assets have worked in the past to create better, more robust, portfolios. How many bogleheads know that even an asset with no internal rate of return or one that shows no appreciable long term growth, can yet still improve a portfolio? My guess is it's the minority, but it's true if the asset has enough relative volatility to the other assets and is not highly correlated. In the sorts of portfolios I personally like, nominal bonds are one of many sleeves, and long duration treasuries are the absolute best asset to play that nominal fixed income role imo.
Well, I agree I guess. I don't think it's outrageous to hold Gold. I never get the argument that "it doesn't pay interest/dividends, so don't hold it." From a CAPM perspective its beta=0 means that its expected return is naturally quite low.
absolute zero
Posts: 1244
Joined: Thu Dec 29, 2016 3:59 pm

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

Post by absolute zero »

watchnerd wrote: Tue Apr 13, 2021 7:25 pm
corp_sharecropper wrote: Tue Apr 13, 2021 7:19 pm
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.
99.34% of the risk was from equities, actually, 2008-present.

https://www.portfoliovisualizer.com/bac ... tion2_1=20
I'm not sure how much weight I'd put in that number from PV. If you swap 20% TBM for a stronger diversifier like 20% VGLT, you get that 105% of the risk came from equities.
:confused
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

absolute zero wrote: Wed Apr 14, 2021 8:09 am
watchnerd wrote: Tue Apr 13, 2021 7:25 pm
corp_sharecropper wrote: Tue Apr 13, 2021 7:19 pm
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.
99.34% of the risk was from equities, actually, 2008-present.

https://www.portfoliovisualizer.com/bac ... tion2_1=20
I'm not sure how much weight I'd put in that number from PV. If you swap 20% TBM for a stronger diversifier like 20% VGLT, you get that 105% of the risk came from equities.
:confused
Take it with as much salt as you like.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
LukeHeinz57
Posts: 166
Joined: Wed Sep 03, 2014 12:01 pm
Location: Ohio

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

Post by LukeHeinz57 »

absolute zero wrote: Wed Apr 14, 2021 8:09 am
watchnerd wrote: Tue Apr 13, 2021 7:25 pm
corp_sharecropper wrote: Tue Apr 13, 2021 7:19 pm
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.
99.34% of the risk was from equities, actually, 2008-present.

https://www.portfoliovisualizer.com/bac ... tion2_1=20
I'm not sure how much weight I'd put in that number from PV. If you swap 20% TBM for a stronger diversifier like 20% VGLT, you get that 105% of the risk came from equities.
:confused
I think what you are seeing is the negative correlation LTT have with equities which is kind of the whole point of this thread...Right?
"Contentment", the only thing you ever truly need more of!
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

LukeHeinz57 wrote: Wed Apr 14, 2021 9:06 am I think what you are seeing is the negative correlation LTT have with equities which is kind of the whole point of this thread...Right?
Mathematically, a negatively correlated asset like LTT can cause an increase equity contribution to >100%.

However, one can see how some might find this to be counter-intuitive and perceived as impossible (absent leverage).

Understanding that 'risk' and 'volatility' are inter-related in this model is important.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
jeffyscott
Posts: 13487
Joined: Tue Feb 27, 2007 8:12 am

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

Post by jeffyscott »

LukeHeinz57 wrote: Wed Apr 14, 2021 9:06 am
absolute zero wrote: Wed Apr 14, 2021 8:09 am
watchnerd wrote: Tue Apr 13, 2021 7:25 pm
corp_sharecropper wrote: Tue Apr 13, 2021 7:19 pm
Ok, well in this situation your typical 80/20 TSM/TBM boglehead isn't fairing much better with his/her portfolio driven almost entirely by equity risk (really, 20% might sound like a lot but a 80/20 portfolio with TBM is basically run by about 90%-95% equity risk). I'd also wager that the LTT holder already has a decent headstart before this hypothetical inflationary recession.
99.34% of the risk was from equities, actually, 2008-present.
I'm not sure how much weight I'd put in that number from PV. If you swap 20% TBM for a stronger diversifier like 20% VGLT, you get that 105% of the risk came from equities.
:confused
I think what you are seeing is the negative correlation LTT have with equities which is kind of the whole point of this thread...Right?
You can extend the analysis back to 1993 by using VTSMX and VBMFX, instead of the ETFs. Doesn't change the result, though, 99.61% of risk is was from equities with VBMFX. Using VUSTX for long term treasuries gives basically the same result as total bond with 100.81% of risk from equities. Unsurprisingly, if you had 80% of your assets in stocks, that has been responsible for essentially 100% of your risk, regardless of what bond fund you used for the other 20%.

https://www.portfoliovisualizer.com/bac ... tion3_2=20
User avatar
imak
Posts: 212
Joined: Fri Mar 22, 2019 5:18 pm
Location: Austin, TX

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

Post by imak »

watchnerd wrote: Wed Apr 14, 2021 9:11 am
LukeHeinz57 wrote: Wed Apr 14, 2021 9:06 am I think what you are seeing is the negative correlation LTT have with equities which is kind of the whole point of this thread...Right?
Mathematically, a negatively correlated asset like LTT can cause an increase equity contribution to >100%.

However, one can see how some might find this to be counter-intuitive and perceived as impossible (absent leverage).

Understanding that 'risk' and 'volatility' are inter-related in this model is important.
Another way to look at it is - having 20% of negatively correlated volatile asset like LTT indirectly induces a disciplined "value averaging" approach for 80% of equities. It lowers the cost basis by buying more shares at cheaper valuations during bear markets. From this perspective, LTT are useful even if their expected return is 0% due to their volatility.
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

imak wrote: Wed Apr 14, 2021 10:01 am
watchnerd wrote: Wed Apr 14, 2021 9:11 am
LukeHeinz57 wrote: Wed Apr 14, 2021 9:06 am I think what you are seeing is the negative correlation LTT have with equities which is kind of the whole point of this thread...Right?
Mathematically, a negatively correlated asset like LTT can cause an increase equity contribution to >100%.

However, one can see how some might find this to be counter-intuitive and perceived as impossible (absent leverage).

Understanding that 'risk' and 'volatility' are inter-related in this model is important.
Another way to look at it is - having 20% of negatively correlated volatile asset like LTT indirectly induces a disciplined "value averaging" approach for 80% of equities. It lowers the cost basis by buying more shares at cheaper valuations during bear markets. From this perspective, LTT are useful even if their expected return is 0% due to their volatility.
While this makes sense, you can still get PV to show >100% equity risk contribution even if you entirely turn off rebalancing.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
dorster
Posts: 278
Joined: Tue Aug 15, 2017 4:20 pm

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

Post by dorster »

vineviz wrote: Tue Apr 13, 2021 6:43 pm
Vineviz, I've heard you recommend series I savings bonds (from Treasury Direct) in other threads, and I think described them as in some ways having a longish duration (though I realize they're not really comparable to normal bonds so maybe duration isn't the right word).

What do you see as tradeoffs between edv/vglt and series I or EE savings bonds for the first 20% of bonds in a portfolio?

I'm more or less 100% stocks and following a version of lifecycle investing (100/S to be specific) and am in the process of adding bonds. If I-Bond yields are similar to EDV would they make sense to consider? My investment horizon would be at least 25-35 years.

Thanks for your contributions to the forum (and I hope it's not rude to flag you with a specific question).
Baconator
Posts: 5
Joined: Sat Apr 03, 2021 12:30 am

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

Post by Baconator »

I’m interested in potentially adding long term treasuries but I have the following mental problems I can’t seem to get over:

1). I’m not convinced the negative correlations vs. equities will hold because it seems like it’s different this time with rates so low.

2). It doesn’t seem like a smart investment to me on its face when you think about lending the government money for 25+ years for 2.25%

Has anyone been able to get over these and how do you rationalize these away?
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

Baconator wrote: Sun Apr 18, 2021 4:51 pm I’m interested in potentially adding long term treasuries but I have the following mental problems I can’t seem to get over:

1). I’m not convinced the negative correlations vs. equities will hold because it seems like it’s different this time with rates so low.

2). It doesn’t seem like a smart investment to me on its face when you think about lending the government money for 25+ years for 2.25%

Has anyone been able to get over these and how do you rationalize these away?
1) Doesn't bother me.

2) I don't buy any Treasury of any duration to make money. If I can get 0 real return, I'm fine with that. Long Treasuries are pretty close to a 0 real return (20 year is a little negative, 30 year is a little positive).
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

dorster wrote: Sun Apr 18, 2021 4:37 pm
vineviz wrote: Tue Apr 13, 2021 6:43 pm
Vineviz, I've heard you recommend series I savings bonds (from Treasury Direct) in other threads, and I think described them as in some ways having a longish duration (though I realize they're not really comparable to normal bonds so maybe duration isn't the right word).

What do you see as tradeoffs between edv/vglt and series I or EE savings bonds for the first 20% of bonds in a portfolio?

I'm more or less 100% stocks and following a version of lifecycle investing (100/S to be specific) and am in the process of adding bonds. If I-Bond yields are similar to EDV would they make sense to consider? My investment horizon would be at least 25-35 years.

Thanks for your contributions to the forum (and I hope it's not rude to flag you with a specific question).
I think savings bonds can be an excellent alternative to EDV for someone <30 years from expected retirement

They can’t be used for rebalancing very easily, but that’s not really a drawback
"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: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

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

Post by vineviz »

Baconator wrote: Sun Apr 18, 2021 4:51 pm I’m interested in potentially adding long term treasuries but I have the following mental problems I can’t seem to get over:

1). I’m not convinced the negative correlations vs. equities will hold because it seems like it’s different this time with rates so low.

2). It doesn’t seem like a smart investment to me on its face when you think about lending the government money for 25+ years for 2.25%

Has anyone been able to get over these and how do you rationalize these away?
I’m with watchnerd on this .

I think concern about correlations is unfounded, and not only because it’s not likely they’ll turn positive. More importantly, it doesn’t matter if they do. They’re still the best way to diversify an equity portfolio eleven with positive correlations.

Treasury bond yields are always lower than expected stock returns. It’s no different now than it ever had been. You buy bonds because you want more certainty about your portfolio returns, not because you want higher returns.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
jeffyscott
Posts: 13487
Joined: Tue Feb 27, 2007 8:12 am

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

Post by jeffyscott »

Baconator wrote: Sun Apr 18, 2021 4:51 pm I’m interested in potentially adding long term treasuries but I have the following mental problems I can’t seem to get over:

1). I’m not convinced the negative correlations vs. equities will hold because it seems like it’s different this time with rates so low.

2). It doesn’t seem like a smart investment to me on its face when you think about lending the government money for 25+ years for 2.25%

Has anyone been able to get over these and how do you rationalize these away?
I'm not interested in long term treasuries, in any case, but correlations change and have changed over time. I would not rely on them to make up for the low yields. Other than whatever may be in our bond and balanced funds, I am not lending money to the government for 25+ years, but if I were, it'd be done by buying EE bonds to cover at least the first 20. I'd prefer the certainty of a higher yield, rather than hoping that there will be a rebalancing benefit that exceeds the benefit of the higher yields.
User avatar
anon_investor
Posts: 15122
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

jeffyscott wrote: Tue Apr 20, 2021 8:32 am
Baconator wrote: Sun Apr 18, 2021 4:51 pm I’m interested in potentially adding long term treasuries but I have the following mental problems I can’t seem to get over:

1). I’m not convinced the negative correlations vs. equities will hold because it seems like it’s different this time with rates so low.

2). It doesn’t seem like a smart investment to me on its face when you think about lending the government money for 25+ years for 2.25%

Has anyone been able to get over these and how do you rationalize these away?
I'm not interested in long term treasuries, in any case, but correlations change and have changed over time. I would not rely on them to make up for the low yields. Other than whatever may be in our bond and balanced funds, I am not lending money to the government for 25+ years, but if I were, it'd be done by buying EE bonds to cover at least the first 20. I'd prefer the certainty of a higher yield, rather than hoping that there will be a rebalancing benefit that exceeds the benefit of the higher yields.
Would your mind change if LTT offered a higher interest rate (e.g. 4%)?
User avatar
watchnerd
Posts: 13614
Joined: Sat Mar 03, 2007 10:18 am
Location: Gig Harbor, WA, USA

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

Post by watchnerd »

jeffyscott wrote: Tue Apr 20, 2021 8:32 am I'd prefer the certainty of a higher yield, rather than hoping that there will be a rebalancing benefit that exceeds the benefit of the higher yields.
I think trading off some equity portion for junk bonds is a rational choice if one has concerns about low Treasury yields and/or correlation persistence.

Where people get into hot water is keeping a high equity allocation and holding junk.
Global stocks, IG/HY bonds, gold & digital assets at market weights 75% / 19% / 6% || LMP: TIPS ladder
Locked