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
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
On the internet, nobody knows you're a dog.
HEDGEFUNDIE
Posts: 4801
Joined: Sun Oct 22, 2017 2:06 pm

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

Post by HEDGEFUNDIE »

CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
Long term TIPS are a thing. The fly in that ointment is that inflation insurance does not come for free.

https://www.portfoliovisualizer.com/fun ... chmark=TLT
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
Unexpected inflation is primarily a threat to consumption, so the impact on portfolios under decumulation is the real concern.

I expect that few Bogleheads are maintaining an 80/20 allocation in retirement, but I hope that any who ARE doing so at least consider using inflation-indexed Treasuries instead of nominal Treasuries.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
HawkeyePierce
Posts: 1486
Joined: Tue Mar 05, 2019 10:29 pm
Location: Colorado

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

Post by HawkeyePierce »

BigJohn wrote: Thu Apr 23, 2020 8:07 am
HEDGEFUNDIE wrote: Wed Apr 22, 2020 2:35 pm
garlandwhizzer wrote: Wed Apr 22, 2020 2:09 pm I'm missing the same thing. There is a high level of uncertainty for anyone to predict the exact time frame over which they'll need money and exactly how much money they'll need. All the nice arithmetic models and graphs given to support LTT alone assume that the investor knows this information with full reliability up front for periods of 30 years or more and that those needs are not interrupted by life. There is not a single individual in the world who knows that with full reliability up front regardless of what the model or graph show. If you're living on this planet in the real world that basic underlying assumption is flawed. Unexpected things happen all time that totally change our financial needs. Having a secure source of non-volatile assets has value IMO regardless of backtesting results, models, or graphs. It all depends on the individual's circumstances and preference IMO, not a one size fits all situation.

Garland Whizzer
I am 34.

Putting 100% of my bonds into LTT would be a mistake, since my spending needs are not all in the distant future.

Putting 100% of my bonds into STT would also be a mistake, since my spending needs are not all within the next few years.

So what mix of bonds should I hold?

Doing literally no math, it seems obvious to me that small minority of my spending needs are 1-10 years away. For these I will hold short and intermediate term bonds.

But a large majority of my spending needs are over 10 years away. For these I will hold long term bonds.

So what is the ideal weighted average duration of all my bonds?

Probably >20 years.

There, that wasn't so hard, was it?
I think that you and vinevz are both missing the point about how unpredictable life can be. Maybe it’s just lack of experience or maybe you understand but can ignore. However, in my mind there is a non-trivial chance of some life event increasing your annual expense by several fold and doing so well before your average duration or the duration of your LTT’s. If that occurs at a time when rates are up and prices are down, I might not have the choice to just ride out the volatility to maturity. I need a lot of cash now and could be forced to sell “early” and take a significant loss. Just a few examples to make my point...

A small business owner when COVID 19 hits that needs cash to prevent bankruptcy. A spouse who gets early onset Alzheimer’s and needs LTC for many years starting at age 45-50. A child in an accident or with an illness that is racking up significant medical bills. The list is endless and not all may apply to you but they apply to many. Like all risks, it’s about balance so for many the answer is intermediate term to mitigate this risk despite the potentially superior performance of long term.

So, I’m not saying your technical analysis is wrong but I don’t think it’s appropriate to trivialize these kinds of concerns as unimportant.
I just barbell my EDV holdings with enough cash to see me through a bad year. Problem solved.

Well, problem solved for me but I'm 29 with no dependents and private disability insurance to cover lost income if I'm unable to work for medical reasons.

Some of the unexpected life events you mention can be covered with insurance rather than investments. If we're aiming for liability-matching, that seems like the optimal course to me.
User avatar
willthrill81
Posts: 20915
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

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

Post by willthrill81 »

HEDGEFUNDIE wrote: Thu Apr 23, 2020 9:06 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
Long term TIPS are a thing. The fly in that ointment is that inflation insurance does not come for free.
With the 30 year TIPS break-even inflation rate currently being 1.29%, it could be argued to be free right now. If you believe that the Fed will hit its inflation target, then you're getting paid to take on the inflation insurance.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

HEDGEFUNDIE wrote: Thu Apr 23, 2020 9:06 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
Long term TIPS are a thing. The fly in that ointment is that inflation insurance does not come for free.

https://www.portfoliovisualizer.com/fun ... chmark=TLT
The thing about TIPS is that they can be a way of hedging the capital invested in them from inflation, but don't do much to protect other investment assets from inflation, as explained below. And the thing about a long term TIPS fund is that the correlation to inflation is quite low, so it doesn't even do a good job of hedging the capital invested in the long TIPS fund from inflation. About the best use of TIPS is to invest in a ladder of TIPS and hold them to maturity; that's the only way TIPS really can offset inflation, and that's for the actual capital invested in the ladder. In other words - TIPS are much weaker asset than expected if what you're trying to do is offset the impact of inflation on overall portfolio returns. They are a great asset for liability-matching.
According to Vanguard, a good inflation hedge combines a strong correlation to inflation and a high "inflation beta" (volatility benchmarked to inflation). Here's why. Money invested in an asset with just a high correlation to inflation provides inflation protection for the money invested in that asset; but it does not provide protection for the rest of the portfolio. Investing in an asset with both high correlation and high inflation beta means that it's protection goes beyond the invested position and can help to protect the rest of the portfolio as well. This is a basic principal of portfolio diversification: the diversification benefit of an asset depends on both it's correlation and it's beta, or volatility.

The problem with TIPS is that there's no way to invest in them that provides both a high correlation to unexpected inflation and high beta.

A TIPs bond ladder with bonds held to maturity has a very high correlation of returns to inflation -- that's what they're designed to do. However, the returns have very low inflation beta. This means that a TIPS bond ladder does a good job of protecting the capital allocated from inflation, but does nothing to protect other investment assets from unexpected inflation.

The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation.

Finally, we have TIPS funds with longer durations. This option has a higher inflation beta; however longer duration TIPS funds have a very low correlation of realized returns to inflation. Vanguard estimates that the correlation of the broad TIPs index with inflation is about 0.20 or less. This means that money invested in an intermediate to long term TIPS fund offers little protection from unexpected inflation to neither the invested position nor to the rest of the portfolio.
viewtopic.php?f=10&t=290538
On the internet, nobody knows you're a dog.
User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

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

Post by watchnerd »

CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP
User avatar
Doc
Posts: 9807
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 »

vineviz wrote: Wed Apr 22, 2020 7:00 pm Of all the traditional portfolio assets, long-term Treasuries are regarded as the best diversifier of stocks for a reason. Regardless of the "feels", an 80/20 investor using long-term Treasuries can expect to have less overall portfolio volatility and lower drawdowns than an 80/20 investor using total bond market. That means more available wealth when "life happens". When "life happens" you get your cash from the portfolio, and the portfolio most likely to give you the most cash with the most liquidity will contain long-term Treasuries.
Can you expand the 80/20 idea to a more "mature" portfolio. An 80/20 with the 20% in LT T's could be interpreted as 25% of you equites in LT T's. So if you had a 60/40 AA would the long T's recommendation be only 15% of your total. That would mean you long T's would be about 33% to you total FI portfolio not 100%.
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
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

Doc wrote: Thu Apr 23, 2020 10:46 am
vineviz wrote: Wed Apr 22, 2020 7:00 pm Of all the traditional portfolio assets, long-term Treasuries are regarded as the best diversifier of stocks for a reason. Regardless of the "feels", an 80/20 investor using long-term Treasuries can expect to have less overall portfolio volatility and lower drawdowns than an 80/20 investor using total bond market. That means more available wealth when "life happens". When "life happens" you get your cash from the portfolio, and the portfolio most likely to give you the most cash with the most liquidity will contain long-term Treasuries.
Can you expand the 80/20 idea to a more "mature" portfolio. An 80/20 with the 20% in LT T's could be interpreted as 25% of you equites in LT T's. So if you had a 60/40 AA would the long T's recommendation be only 15% of your total. That would mean you long T's would be about 33% to you total FI portfolio not 100%.
The expansion would be 60% in stocks, 20% in long-term Treasuries, and 20% in "something else".

That's what I intended in the post title.

Now, a hardcore investor could theoretically make the "something else" ALSO be long-term Treasuries but it's difficult to imagine many people for whom that would be prudent. For me, the "something else" will probably be mostly long-term TIPS.
Last edited by vineviz on Thu Apr 23, 2020 10:58 am, edited 1 time in total.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

watchnerd wrote: Thu Apr 23, 2020 10:38 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
As explained above, this is a gun with no ammo:

"The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation."
On the internet, nobody knows you're a dog.
User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

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

Post by watchnerd »

CULater wrote: Thu Apr 23, 2020 10:55 am
watchnerd wrote: Thu Apr 23, 2020 10:38 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
As explained above, this is a gun with no ammo:

"The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation."
What's the better alternative?
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

CULater wrote: Thu Apr 23, 2020 10:55 am
watchnerd wrote: Thu Apr 23, 2020 10:38 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
As explained above, this is a gun with no ammo:

"The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation."
Both short-term TIPS and the stocks offer pretty good long-run protection from inflation (both expected and unexpected) while the LTTs offer good protection from expected inflation (and unexpected inflation too if the LTTs are inflation-indexed).

So really the investor is, in principle, mainly exposed to unexpected inflation for a small portion of their portfolio. Unless the portfolio is expected to cover a large percentage of non-discretionary expenses (i.e. the investor has a very low Social Security benefit) the overall inflation risk probably isn't extreme.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

vineviz wrote: Thu Apr 23, 2020 11:05 am
CULater wrote: Thu Apr 23, 2020 10:55 am
watchnerd wrote: Thu Apr 23, 2020 10:38 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
As explained above, this is a gun with no ammo:

"The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation."
Both short-term TIPS and the stocks offer pretty good long-run protection from inflation (both expected and unexpected) while the LTTs offer good protection from expected inflation (and unexpected inflation too if the LTTs are inflation-indexed).

So really the investor is, in principle, mainly exposed to unexpected inflation for a small portion of their portfolio. Unless the portfolio is expected to cover a large percentage of non-discretionary expenses (i.e. the investor has a very low Social Security benefit) the overall inflation risk probably isn't extreme.
You can't make that true by just saying it. Where's the beef? It's just not true that short TIPS, intermediate TIPS, or long TIPS offer much inflation protection for the total portfolio. The thing that TIPS are good for is protecting the amount invested in TIPS from inflation, and TIPS funds are inferior to TIPS held to maturity for that.

If you want to protect your portfolio from inflation, about the only way to do that is to allocate more assets to stuff that doesn't lose a lot of real value to inflation; cash, short bonds. Long bonds are vulnerable to inflation; so if that's a black swan then don't own them -- this would be especially true for retirees. Especially alongside stocks.
On the internet, nobody knows you're a dog.
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

nps wrote: Thu Apr 23, 2020 6:09 am
SnowBog wrote: Thu Apr 23, 2020 4:59 am
Wow, that is a long post. Could I try to distill down what you are saying? I think it is that price changes very close to your need for the money will bother you, even if your end-state value over the entire 20 years is more favorable than alternative investing approaches. Does that sum it up?
Put more simply, if we enter a sustained period of rising interest rates (or just a really short spike at the wrong time), long term treasury funds will be penalized driving down their NAV (meaning I may have less money then I invested in it).

I'm not predicting we'll be in such a period. We may remain in a flat/failing rate environment - like we've been in the past 30 years. But my point is, I don't know, and I see that as a legitimate risk to my plan (which is heavily impacted by SOR).

A shorter term fund has less sensitivity to rate changes... Maybe it isn't as great a diversify, maybe it won't maximize my earnings, but I'm not seeing anyone say it's a bad choice (just not their preferred choice).

Maybe my continued disconnect is that my goal is safety, as I'm nearing were I have more to lose than gain. For me, I'll take the option with the lower downside risk.

But to the OP, I'm likely not their target audience. So to each their own...

I appreciate the attempts to explain how this isn't the case, how I'm looking at things all wrong, etc. I really do...
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

CULater wrote: Thu Apr 23, 2020 11:14 am It's just not true that short TIPS, intermediate TIPS, or long TIPS offer much inflation protection for the total portfolio.
I agree, which is why I never made such a claim.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

Back to the OP, advocating putting 20% into long term bonds; for example, owning an 80% stocks + 20% bonds for a younger investor, with the 20% in long treasuries. Seems reasonable, but I happened to take a look at this using Portfolio Visualizer and was a bit surprised. Turns out that an 80/20 portfolio rebalanced annually had about the same return for the 25-year period 1978-2003 whether the 20% was in long treasuries or intermediate treasuries (cagr = 12.23% vs. 12.04%) with the same Sharpe. There's a bigger difference after that, primarily due to the crash in 2008-09 but it's not much. And the drawdown was just about identical whether you had long or intermediate treasuries (39.3% vs. 39.9%). Why? Well, it's because with an 80% allocation to stocks the primary driver of portfolio returns is stock returns and this swamps the remaining 20% whatever it is.

Long story short, I'm not sure that you'll notice the difference between 20% in LTT vs. 20% in ITT if you have a portfolio highly skewed toward stocks. It makes a bigger difference if you are considering portfolios with a lower stock allocation. From a risk parity perspective that's about 50% in stocks and 50% in long bonds.
On the internet, nobody knows you're a dog.
User avatar
CULater
Posts: 2832
Joined: Sun Nov 13, 2016 10:59 am
Location: Hic sunt dracones

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

Post by CULater »

vineviz wrote: Thu Apr 23, 2020 11:29 am
CULater wrote: Thu Apr 23, 2020 11:14 am It's just not true that short TIPS, intermediate TIPS, or long TIPS offer much inflation protection for the total portfolio.
I agree, which is why I never made such a claim.
But you said this:
Both short-term TIPS and the stocks offer pretty good long-run protection from inflation (both expected and unexpected) while the LTTs offer good protection from expected inflation (and unexpected inflation too if the LTTs are inflation-indexed).
On the internet, nobody knows you're a dog.
User avatar
Doc
Posts: 9807
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 »

vineviz wrote: Thu Apr 23, 2020 10:54 am The expansion would be 60% in stocks, 20% in long-term Treasuries, and 20% in "something else".

That's what I intended in the post title.
That is not the answer I would have expected. I agree that long term nominal Treasuries are probably the best if one is going to rebalnce in a stock market crash. (Many people here did not rebalance in the recent few months.) If one is trying to protect one's future cash flow only I would suggest that long TIPS might be appropriate.

I agree that having a pool of nominal Treasuries to rebalance in a stock market crash is a good idea. Whether that pool is 20% of one's total portfolio or 25% of the equity portion is a nuance.

With today's yield curve being as flat and as low as it is I would use intermediate Treasuries instead of long term. Thus trading off some protection on the possible continued equity sell off against the very likely probability that interest rates will become more normal in the next few years. Thus making the big future loss in the FI portion of one's portfolio not the equity part.

One also needs to factor in one's personal duration along with the FI duration. :)
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.
BigJohn
Posts: 1895
Joined: Wed Apr 02, 2014 11:27 pm

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

Post by BigJohn »

HawkeyePierce wrote: Thu Apr 23, 2020 9:23 am I just barbell my EDV holdings with enough cash to see me through a bad year. Problem solved.

Well, problem solved for me but I'm 29 with no dependents and private disability insurance to cover lost income if I'm unable to work for medical reasons.

Some of the unexpected life events you mention can be covered with insurance rather than investments. If we're aiming for liability-matching, that seems like the optimal course to me.
Yes, that's just my point, it's covered for you and maybe hedgefundie and vinevz but not everyone is in your circumstances. Scenarios I laid out are not covered by just lost income insurance, you might need 2 or 3 times more than your normal expenditures. And these expense can last far longer than one year. For example, some Alzheimer's patients are in LTC for 5-10+ years at costs that can exceed $100k/yr depending on the area of the country.

Yes, insurance might be available to cover some of those expenses but that's very iffy. From everything I've read, most small business insurance specifically excludes pandemics, at least to date. LTC insurance is available and is a good deal if you bought it years ago. Today's policies offer much less protection at much higher cost and are often unaffordable (several threads on this forum if you're interested). Not everyone has access to low deductible, low out-of-pocket health insurance, you don't have to live under a rock to read stories about these kinds of huge expenses happening to real people. In the absence of solid insurance, you have to self-insure with the ability to get at your assets much sooner than planned.

Again, I'm not say any of what is proposed in the thread is wrong. It might be the best plan for some people with low/no chance of these kinds of scenarios. But a small dose of "maybe it's not the right answer for everyone" from the LTT's are the only answer crowd is in order in my opinion.
HEDGEFUNDIE
Posts: 4801
Joined: Sun Oct 22, 2017 2:06 pm

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

Post by HEDGEFUNDIE »

SnowBog wrote: Thu Apr 23, 2020 11:15 am
nps wrote: Thu Apr 23, 2020 6:09 am
SnowBog wrote: Thu Apr 23, 2020 4:59 am
Wow, that is a long post. Could I try to distill down what you are saying? I think it is that price changes very close to your need for the money will bother you, even if your end-state value over the entire 20 years is more favorable than alternative investing approaches. Does that sum it up?
Put more simply, if we enter a sustained period of rising interest rates (or just a really short spike at the wrong time), long term treasury funds will be penalized driving down their NAV (meaning I may have less money then I invested in it).

I'm not predicting we'll be in such a period. We may remain in a flat/failing rate environment - like we've been in the past 30 years. But my point is, I don't know, and I see that as a legitimate risk to my plan (which is heavily impacted by SOR).

A shorter term fund has less sensitivity to rate changes... Maybe it isn't as great a diversify, maybe it won't maximize my earnings, but I'm not seeing anyone say it's a bad choice (just not their preferred choice).

Maybe my continued disconnect is that my goal is safety, as I'm nearing were I have more to lose than gain. For me, I'll take the option with the lower downside risk.

But to the OP, I'm likely not their target audience. So to each their own...

I appreciate the attempts to explain how this isn't the case, how I'm looking at things all wrong, etc. I really do...
For near term needs we have no disagreement.

For longer term needs you are absolutely looking at it wrong.

Let's imagine the Treasury changes the EE bond rules tomorrow that reduce the doubling feature by half, so only 50% gain over 20 years instead of 100%.

My expected return in the long term (which I was counting on to retire) just got cut in half. That is not safety to me. And you can dismiss it as simply not "maximizing earnings" but the pernicious thing is that it is happening for the period which you can least afford it, when your human capital is exhausted.

I like safety both in near term and long term. Not just the former.
User avatar
anon_investor
Posts: 3501
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

SnowBog wrote: Thu Apr 23, 2020 11:15 am
nps wrote: Thu Apr 23, 2020 6:09 am
SnowBog wrote: Thu Apr 23, 2020 4:59 am
Wow, that is a long post. Could I try to distill down what you are saying? I think it is that price changes very close to your need for the money will bother you, even if your end-state value over the entire 20 years is more favorable than alternative investing approaches. Does that sum it up?
Put more simply, if we enter a sustained period of rising interest rates (or just a really short spike at the wrong time), long term treasury funds will be penalized driving down their NAV (meaning I may have less money then I invested in it).

I'm not predicting we'll be in such a period. We may remain in a flat/failing rate environment - like we've been in the past 30 years. But my point is, I don't know, and I see that as a legitimate risk to my plan (which is heavily impacted by SOR).

A shorter term fund has less sensitivity to rate changes... Maybe it isn't as great a diversify, maybe it won't maximize my earnings, but I'm not seeing anyone say it's a bad choice (just not their preferred choice).

Maybe my continued disconnect is that my goal is safety, as I'm nearing were I have more to lose than gain. For me, I'll take the option with the lower downside risk.

But to the OP, I'm likely not their target audience. So to each their own...

I appreciate the attempts to explain how this isn't the case, how I'm looking at things all wrong, etc. I really do...
This may be another way to explain the OP, at least this is how I understand it. The recommendation for LTT is not for income. Buying LLT at today's rates is accepting the low 1+% yields (not expecting growth from the yields). But the benefit of the LTT is that is counter balances equities (you will get your growth from the equities). So when equities fall, LTT will rise to reduce the volitility/standard deviation of the overall portfolio. The general idea is that at any point in time, while your LTT or your equities may be down, the overal portfolio will not be down as much (compared to having TBM instead of LLT). At least this makes a lot of sense to me. I also understand the point that you may not want more than 20% of your portfolio in LTT, and having the rest of the fixed income in something else would make the overall portfolio (in my mind) "all weather", as the LTT can counter balance the equities and the other fixed income (TIPS/I Bonds for inflation and intermediate/short term nominal bonds - maybe to serve as an emergency fund); this would help the portfolio to survive most environments/situations.

Personally as I am an investor with a long investment time horizon still in the accumlation phase, it may make sense for me to have LLT up to 20% and I Bonds/cash serve as an emergency fund, and the rest in equities.
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

CULater wrote: Thu Apr 23, 2020 11:35 am
vineviz wrote: Thu Apr 23, 2020 11:29 am
CULater wrote: Thu Apr 23, 2020 11:14 am It's just not true that short TIPS, intermediate TIPS, or long TIPS offer much inflation protection for the total portfolio.
I agree, which is why I never made such a claim.
But you said this:
Both short-term TIPS and the stocks offer pretty good long-run protection from inflation (both expected and unexpected) while the LTTs offer good protection from expected inflation (and unexpected inflation too if the LTTs are inflation-indexed).
I don't see the conflict, unless you are inferring something that I didn't write.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

HEDGEFUNDIE wrote: Thu Apr 23, 2020 7:39 am But absent an emergency, if you are holding EDV and your time horizon is still long, and EDV’s price drops because interest rates rise, that literally doesn’t affect you and that does not count as SORR. As vineviz has shown in his charts, SORR is actually minimized when you duration match to your spending need. You just have to close your eyes and bear the meaningless price drops along the way.

But I will point out again that AQR article which talks about people paying a premium not to be exposed to volatility, even meaningless volatility. So you’re definitely not alone, if that gives you any comfort.
You appear to be optimizing the overall portfolio, and the argument is LTT has historically offered the best diversification. I'll accept that...

But I disagree that because I'm not convinced that LTT is my best option, I'm paying a premium to not be exposed to volatility. It's the opposite...

By your/vineviz data - I'm exposing myself to increased volatility if I don't have 20% LTT. But you are looking at the whole portfolio - which is fine...

I'm looking at my first 10+ years of downdraw - starting in 10 years. My goal to minimize SOR is that I'll be pulling from "safe" assets during those years.

If I have enough "safe assets" to help weather those 10+ years, then I can "close my eyes" to the volatility of my stocks. If that volatility is higher because I don't have LTT, I don't care.

But I need to survive those first 10+ years...

If a good chunk of my "safe assets" are in LTT, and we hit that high interest rate period (however unlikely you may think it is), my LTT holdings would take a large hit likely while stocks are taking a large hit (no more easy money for businesses). If that happens, my plan could fall apart. I don't care if the odds of that are < 5% - it's the biggest risk I see in my plan.

So I'm not (as) concerned with the volatility of my overall portfolio (which you seem to be), I'm concerned with the volatility of my "safe" assets to bridge these years.
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

Doc wrote: Thu Apr 23, 2020 11:40 am
vineviz wrote: Thu Apr 23, 2020 10:54 am The expansion would be 60% in stocks, 20% in long-term Treasuries, and 20% in "something else".

That's what I intended in the post title.
That is not the answer I would have expected. I agree that long term nominal Treasuries are probably the best if one is going to rebalnce in a stock market crash. (Many people here did not rebalance in the recent few months.) If one is trying to protect one's future cash flow only I would suggest that long TIPS might be appropriate.
I know it is common to infer that I mean "nominal" Treasuries when I write "long-term Treasuries" but that's not always the case.

The "T" in "TIPS" stands for Treasury, after all.

And again, I think it's reasonable to assume that most investors with an 80% stock allocation are decidedly pre-retirement. Would you agree? And the traditional Boglehead choice for the bond fund, total bond market, is ALSO entirely comprised of nominal bonds.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
anon_investor
Posts: 3501
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

vineviz wrote: Thu Apr 23, 2020 12:08 pm
Doc wrote: Thu Apr 23, 2020 11:40 am
vineviz wrote: Thu Apr 23, 2020 10:54 am The expansion would be 60% in stocks, 20% in long-term Treasuries, and 20% in "something else".

That's what I intended in the post title.
That is not the answer I would have expected. I agree that long term nominal Treasuries are probably the best if one is going to rebalnce in a stock market crash. (Many people here did not rebalance in the recent few months.) If one is trying to protect one's future cash flow only I would suggest that long TIPS might be appropriate.
I know it is common to infer that I mean "nominal" Treasuries when I write "long-term Treasuries" but that's not always the case.

The "T" in "TIPS" stands for Treasury, after all.

And again, I think it's reasonable to assume that most investors with an 80% stock allocation are decidedly pre-retirement. Would you agree? And the traditional Boglehead choice for the bond fund, total bond market, is ALSO entirely comprised of nominal bonds.
Do any long term TIPS mutual funds/ETFs exist?
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

SnowBog wrote: Thu Apr 23, 2020 11:59 am
I'm looking at my first 10+ years of downdraw - starting in 10 years. My goal to minimize SOR is that I'll be pulling from "safe" assets during those years.
I think I understand your goal. For you, the years you are most heavily focused on are 2030 through 2040 or so.

You want to have the utmost confidence that your portfolio will safely generate enough income during those 10+ years.

Do I have this right?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

HEDGEFUNDIE wrote: Thu Apr 23, 2020 11:47 am For near term needs we have no disagreement.
I think we finally agree (mostly)...
HEDGEFUNDIE wrote: Thu Apr 23, 2020 11:47 am For longer term needs you are absolutely looking at it wrong.
...
I like safety both in near term and long term. Not just the former.
Clearly our long terms outlooks are different...

I'm lucky in that FireCalc says I don't need to invest any more $ to have a 95%+ chance of having our plan work - provided my savings can continue to grow over the next 10 years. (But I'm still saving anyway...)

Hence why I have more to lose than gain...

I'm lucky enough that my plan allows me to delay SS and pensions until 70 (spouses until 67), at which point I'm projecting they'll cover 75% of expenses (and majority have a COLA). If I can survive my first 10+ years, my projected withdrawal rate for a comfortable retirement is around 2% from 70+ (less if we need to cut back).

So my "risk" in the long term is minimal compared with my short term risk.

As I need to survive my first 10+ years... And my preferred option is the safest, least volatile assets as the backbone of those years. Which we now apparently agree isn't LTT [edited] funds (at least for my time horizon)...

(I'm also lucky that obviously baring bad health or job loss, we are choosing to have the option to retire early. If "life happens" we can choose to work longer, pickup side jobs, etc. That buys us more flexibility.)

I feel very blessed to be in the circumstances that we are in now (years of hard work, sacrifice, and lots of luck). And my "concerns" pale to the circumstances of others - especially those impacted by things like COVID-19.

Again, don't claim to be the target audience of OP or this thread... Thanks for the dialog!
Last edited by SnowBog on Thu Apr 23, 2020 12:45 pm, edited 1 time in total.
ruud
Posts: 320
Joined: Sat Mar 03, 2007 1:28 pm
Location: san francisco bay area

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

Post by ruud »

anon_investor wrote: Thu Apr 23, 2020 12:21 pm Do any long term TIPS mutual funds/ETFs exist?
LTPZ - PIMCO 15+ Year U.S. TIPS Index Exchange-Traded Fund
.
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

anon_investor wrote: Thu Apr 23, 2020 12:21 pm
Do any long term TIPS mutual funds/ETFs exist?
The choices are not very inspiring.

PIMCO 15+ Year US TIPS ETF (LTPZ) would be the most widely available, but the ER is higher than it should be (0.2%) for a fund that only owns 11 bonds. Plus the bid-ask spreads are typically wide.

DFA LTIP Institutional (DRXIX) is cheaper (0.15%) but not very widely available.

Honestly, I'd suggest using Series I savings bonds (at least up to your annual household purchase limit) for taxable and purchasing individual TIPS for the rest (or in a tax-advantaged account).
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

vineviz wrote: Thu Apr 23, 2020 12:24 pm
SnowBog wrote: Thu Apr 23, 2020 11:59 am
I'm looking at my first 10+ years of downdraw - starting in 10 years. My goal to minimize SOR is that I'll be pulling from "safe" assets during those years.
I think I understand your goal. For you, the years you are most heavily focused on are 2030 through 2040 or so.

You want to have the utmost confidence that your portfolio will safely generate enough income during those 10+ years.

Do I have this right?
Mostly... Years are a little off, but close enough...

Ideally I can survive [mostly] off a return of principal with small interest gains (to maintain purchasing power) for these years.

I can get that from buying individual bonds.

I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

SnowBog wrote: Thu Apr 23, 2020 12:38 pm I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
Okay, well you can definitely get it with TWO bond funds. I'd be happy to give you some detail if you want it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

vineviz wrote: Thu Apr 23, 2020 12:40 pm
SnowBog wrote: Thu Apr 23, 2020 12:38 pm I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
Okay, well you can definitely get it with TWO bond funds. I'd be happy to give you some detail if you want it.
That's what I've been asking for... And attempted to lay out such a scenario in my lengthy post last night... But based on what's been provided thus far, I don't see how I remove that SOR risk of a rate spike destroying NAV in LTT.

I can rationalize that holding some/more/most in a shorter term fund helps to minimize that risk, especially if that 2nd fund is something like STT (the shorter the duration the less the risk at I understand it). Gives time to (hopefully) let LTT stabilize...

But at that point - at least with my set of circumstances - I question the "need" for LTT at all in my case...

While maybe not optimal but your standards, ITT might be "good enough" for me... Even just STT might be "good enough"... (Combined with my other fixed holdings.)
User avatar
anon_investor
Posts: 3501
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

vineviz wrote: Thu Apr 23, 2020 12:36 pm
anon_investor wrote: Thu Apr 23, 2020 12:21 pm
Do any long term TIPS mutual funds/ETFs exist?
The choices are not very inspiring.

PIMCO 15+ Year US TIPS ETF (LTPZ) would be the most widely available, but the ER is higher than it should be (0.2%) for a fund that only owns 11 bonds. Plus the bid-ask spreads are typically wide.

DFA LTIP Institutional (DRXIX) is cheaper (0.15%) but not very widely available.

Honestly, I'd suggest using Series I savings bonds (at least up to your annual household purchase limit) for taxable and purchasing individual TIPS for the rest (or in a tax-advantaged account).
Thanks. That is what I thought. I already maxed out his/her I Bonds for this year to lock in the 0.2% fixed rate (assuming the fixed interest rate will go down to 0% on May 1). I probably will not play the paper I Bond tax refund game this year.

It sounds like the consensus is to use EDV (since my investment time horizon is more than the 24+ year duration). I understand that VGLT/VLGSX has a shorter duration, but are those more tax efficient because they do not kick off ST/LT capital gains distributions annually? For a taxable account, I am trying to understand what makes the most sense for someone in the 24% tax bracket + NII tax sometimes; EDV or VGLT/VLGSX.
Gufomel
Posts: 548
Joined: Sat Feb 14, 2015 9:52 pm

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

Post by Gufomel »

SnowBog wrote: Thu Apr 23, 2020 12:54 pm
vineviz wrote: Thu Apr 23, 2020 12:40 pm
SnowBog wrote: Thu Apr 23, 2020 12:38 pm I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
Okay, well you can definitely get it with TWO bond funds. I'd be happy to give you some detail if you want it.
That's what I've been asking for... And attempted to lay out such a scenario in my lengthy post last night... But based on what's been provided thus far, I don't see how I remove that SOR risk of a rate spike destroying NAV in LTT.

I can rationalize that holding some/more/most in a shorter term fund helps to minimize that risk, especially if that 2nd fund is something like STT (the shorter the duration the less the risk at I understand it). Gives time to (hopefully) let LTT stabilize...

But at that point - at least with my set of circumstances - I question the "need" for LTT at all in my case...

While maybe not optimal but your standards, ITT might be "good enough" for me... Even just STT might be "good enough"... (Combined with my other fixed holdings.)
But you also said earlier:
Ideally I can survive [mostly] off a return of principal with small interest gains (to maintain purchasing power) for these years.
So you need 10 years of expenses plus interest for the years 2030-2040 (roughly). Presumably the interest gains you are ok with are roughly 0.5% annually since that’s approximately the YTM of an ITT fund currently. What happens if ITT yields drop to 0% or even negative in the coming years and stay that way through 2040? Your ITT may not maintain the purchasing power that you’re planning for. Can you still meet your goals? Maybe you can, or maybe you’re ok with that risk. But that doesn’t negate the fact that there’s reinvestment risk of holding money that you need from 2030 to 2040 in treasuries that have a significantly shorter duration.

I’ll leave it to vineviz to give his specific example, but I assume it will be to hold a combination of LTT funds and ITT/STT funds, and increasing the ratio of ITT/STT funds as you get closer to your expected timeframe of needing the money. You won’t be sitting in pure LTT by 2030. You’ll have moved a substantial portion to ITT/STT over time by then, but you won’t be subject to the reinvestment risk of holding the entire 10 years of expenses in ITT/STT today.
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

SnowBog wrote: Thu Apr 23, 2020 12:54 pm
vineviz wrote: Thu Apr 23, 2020 12:40 pm
SnowBog wrote: Thu Apr 23, 2020 12:38 pm I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
Okay, well you can definitely get it with TWO bond funds. I'd be happy to give you some detail if you want it.
That's what I've been asking for... And attempted to lay out such a scenario in my lengthy post last night... But based on what's been provided thus far, I don't see how I remove that SOR risk of a rate spike destroying NAV in LTT.
Okay, my previous posts were somewhat generalized but I'll do a real quick illustration of how it might work for someone in a position like yours.

I'm going to assume an investor who is 50 years old, wants to retire at around age 60, and intends to collect Social Security starting at age 70. For this illustration I'm going to focus exclusively on the series of cash flows starting at age 60 and ending at age 69: let's just pretend the money covering the withdrawals that start at age 70 is in a different portfolio that can be invested 60/40 or whatever. Just assume it's covered elsewhere.

Further, for maximum safety and minimum sequence risk I'm going to invest the entire portfolio funding the age 60 to age 70 cash flows in bonds. Very conservative, but not unreasonable.

So, right now the investor has in investment time horizon of approximately 14.5 years. It's an approximation because I'm using the simple formula provided upthread of

Code: Select all

(((age at start of retirement + age at end of retirement)/2) - current age)
((60+69)/2)-50) = 14.5.

If we make that same calculation each year, we can easily produce a graph showing the remaining investment horizon at each age which is the same as our target bond duration. The target bond duration starts at roughly 14.5 years and then, as you approach the last cash flow, declines steadily to zero.

Image

Now, let's set up the bond funds. I'm going to use Vanguard mutual funds but you could make lots of other choices.

For long-term Treasuries, we have Vanguard Long-Term Treasury Index Fund (VLGSX), with a duration of 18.6 years.

For our second fund, we have Vanguard Short-Term Inflation-Protected Securities Index Fund (VTAPX), with a duration of 2.4 years.



Literally the entire strategy is this: once a year look up the current "target duration" in the first graph and calculate the ratio of VLGSX to VTAPX which produces an average duration that is equal to that.

You get a glide path like this:

Image

You start with 75% in long-term nominal Treasuries and 25% in short-term TIPS, and slowly over the course of 20 years transition to 0% in long-term nominal Treasuries and 100% in short-term TIPS. The full allocation schedule is at the bottom, but in practice you'd want to check each year to make sure the fund has the same duration I used.

And because this strategy keeps the durations matched, there is NO INTEREST RATE RISK AT ALL*. No NAV shocks can throw you off course. It's the most boring plan ever, with just 15 minutes using a calculator once a year.

Image

Yields could go to 5% or -5%, tomorrow or in 10 years, and the safety of this plan is not compromised one whit.

* In this example you do end up with a bit of rate risk in the final couple fo years, since the duration of VTAPX is slightly longer than the investment time horizon. You could solve this by allocating something to a cash account (money market or CD), but the biggest 1-year decline in VTAPX is just -1.5%.

Code: Select all

Age	VLGSX	VTAPX
 50.00 	75%	25%
 51.00 	68%	32%
 52.00 	62%	38%
 53.00 	56%	44%
 54.00 	50%	50%
 55.00 	44%	56%
 56.00 	38%	63%
 57.00 	31%	69%
 58.00 	25%	75%
 59.00 	19%	81%
 60.00 	13%	87%
 61.00 	10%	90%
 62.00 	7%	93%
 63.00 	3%	97%
 64.00 	0%	100%
 65.00 	0%	100%
 66.00 	0%	100%
 67.00 	0%	100%
 68.00 	0%	100%
 69.00 	0%	100%
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Robert_007
Posts: 42
Joined: Fri May 18, 2018 1:34 pm

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

Post by Robert_007 »

Would it be smart to put LTT in a Roth if you are young and don’t have the option in your 401k? I’m guessing putting it in a taxable account is not recommended.
User avatar
Doc
Posts: 9807
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 »

Robert_007 wrote: Thu Apr 23, 2020 2:41 pm Would it be smart to put LTT in a Roth if you are young and don’t have the option in your 401k? I’m guessing putting it in a taxable account is not recommended.
In my situation Tax Efficiency: Short Treasuries > short Corporates ~ Intermediate Treasuries > Intermediate Corportates

I don't invest in long bonds but at least with today's yield curve long Treasuries are not going to be that much different than intermediates.

Also if you sell those Treasures to buy equities do you want those equites in tax advantaged or taxable?

No shoe fits all.

And don't let taxes rule the roost.

The best way to minimize taxes is to go to Vegas and put all your money on 00 and just let it ride.
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.
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

vineviz wrote: Thu Apr 23, 2020 2:33 pm
SnowBog wrote: Thu Apr 23, 2020 12:54 pm
vineviz wrote: Thu Apr 23, 2020 12:40 pm
SnowBog wrote: Thu Apr 23, 2020 12:38 pm I was trying to understand if I could get that from a bond fund (which is easier to buy/manage/etc.).
Okay, well you can definitely get it with TWO bond funds. I'd be happy to give you some detail if you want it.
That's what I've been asking for... And attempted to lay out such a scenario in my lengthy post last night... But based on what's been provided thus far, I don't see how I remove that SOR risk of a rate spike destroying NAV in LTT.
Okay, my previous posts were somewhat generalized but I'll do a real quick illustration of how it might work for someone in a position like yours.

I'm going to assume an investor who is 50 years old, wants to retire at around age 60, and intends to collect Social Security starting at age 70. For this illustration I'm going to focus exclusively on the series of cash flows starting at age 60 and ending at age 69: let's just pretend the money covering the withdrawals that start at age 70 is in a different portfolio that can be invested 60/40 or whatever. Just assume it's covered elsewhere.

Further, for maximum safety and minimum sequence risk I'm going to invest the entire portfolio funding the age 60 to age 70 cash flows in bonds. Very conservative, but not unreasonable.

So, right now the investor has in investment time horizon of approximately 14.5 years. It's an approximation because I'm using the simple formula provided upthread of

Code: Select all

(((age at start of retirement + age at end of retirement)/2) - current age)
((60+69)/2)-50) = 14.5.

If we make that same calculation each year, we can easily produce a graph showing the remaining investment horizon at each age which is the same as our target bond duration. The target bond duration starts at roughly 14.5 years and then, as you approach the last cash flow, declines steadily to zero.

Image

Now, let's set up the bond funds. I'm going to use Vanguard mutual funds but you could make lots of other choices.

For long-term Treasuries, we have Vanguard Long-Term Treasury Index Fund (VLGSX), with a duration of 18.6 years.

For our second fund, we have Vanguard Short-Term Inflation-Protected Securities Index Fund (VTAPX), with a duration of 2.4 years.



Literally the entire strategy is this: once a year look up the current "target duration" in the first graph and calculate the ratio of VLGSX to VTAPX which produces an average duration that is equal to that.

You get a glide path like this:

Image

You start with 75% in long-term nominal Treasuries and 25% in short-term TIPS, and slowly over the course of 20 years transition to 0% in long-term nominal Treasuries and 100% in short-term TIPS. The full allocation schedule is at the bottom, but in practice you'd want to check each year to make sure the fund has the same duration I used.

And because this strategy keeps the durations matched, there is NO INTEREST RATE RISK AT ALL*. No NAV shocks can throw you off course. It's the most boring plan ever, with just 15 minutes using a calculator once a year.

Image

Yields could go to 5% or -5%, tomorrow or in 10 years, and the safety of this plan is not compromised one whit.

* In this example you do end up with a bit of rate risk in the final couple fo years, since the duration of VTAPX is slightly longer than the investment time horizon. You could solve this by allocating something to a cash account (money market or CD), but the biggest 1-year decline in VTAPX is just -1.5%.

Code: Select all

Age	VLGSX	VTAPX
 50.00 	75%	25%
 51.00 	68%	32%
 52.00 	62%	38%
 53.00 	56%	44%
 54.00 	50%	50%
 55.00 	44%	56%
 56.00 	38%	63%
 57.00 	31%	69%
 58.00 	25%	75%
 59.00 	19%	81%
 60.00 	13%	87%
 61.00 	10%	90%
 62.00 	7%	93%
 63.00 	3%	97%
 64.00 	0%	100%
 65.00 	0%	100%
 66.00 	0%	100%
 67.00 	0%	100%
 68.00 	0%	100%
 69.00 	0%	100%
Thank you for the detailed explanation!

I'm my particular case, your "60" would be potentially (if I get really lucky) 53 for me. So if I revise your chart as follows:

Code: Select all

Age	VLGSX	VTAPX
53 	13%	87%
54 	10%	90%
55 	7%	93%
56 	3%	97%
57 	0%	100%
That makes sense to me! And it appears we agree holding a significant sized LTT fund is not optimal for this scenario. [edited to add] (To bridge early retirement years while delaying SS & pension.)

If we are only talking about 10% max (in "this" portfolio) of LTT at this point, I stand by my prior comment, an ITT seems a "close enough" option. (Especially when it's part of other assets like a ladder of I & EE Bonds, muni bonds, MM/CD.)
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

SnowBog wrote: Thu Apr 23, 2020 3:45 pm Thank you for the detailed explanation!

I'm my particular case, your "60" would be potentially (if I get really lucky) 53 for me. So if I revise your chart as follows:

Code: Select all

Age	VLGSX	VTAPX
53 	13%	87%
54 	10%	90%
55 	7%	93%
56 	3%	97%
57 	0%	100%
Well, no. I don't think so, unless you can make your pensions and Social Security start at age 58.

Instead, from what I can gather if you start withdrawals earlier (say, at age 53), they still have to last until age 69. Right?

So your "get really lucky" best case shortens the time horizon a little bit, but not a great deal because an earlier start partly is offset by a longer withdrawal period. Your glide path would, if I understand your plan correctly, look more like this:

Code: Select all

Age	Duration	VLGSX	VTAPX
 50 	 11.0 		53%	47%
 51 	 10.0 		47%	53%
 52 	 9.0 		41%	59%
 53 	 8.0 		34%	66%
 54 	 7.5 		31%	69%
 55 	 7.0 		28%	72%
 56 	 6.5 		25%	75%
 57 	 6.0 		22%	78%
 58 	 5.5 		19%	81%
 59 	 5.0 		16%	84%
 60 	 4.5 		13%	87%
 61 	 4.0 		10%	90%
 62 	 3.5 		7%	93%
 63 	 3.0 		3%	97%
 64 	 2.5 		0%	100%
 65 	 2.0 		0%	100%
 66 	 1.5 		0%	100%
 67 	 1.0 		0%	100%
 68 	 0.5 		0%	100%
 69 	 -   		0%	100%
You still need some long-term bonds at the beginning to negate interest rate risk completely, but you could use an intermediate TIPS fund instead of VLGSX if you were comfortable taking a little more risk.

And remember that the rest of your portfolio (the part that funds your 2% withdrawals starting at age 70) now has an EVEN LONGER time horizon than before, so even MORE interest rate risk if you don't use long-term bonds in that part.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

vineviz wrote: Thu Apr 23, 2020 4:03 pm
SnowBog wrote: Thu Apr 23, 2020 3:45 pm Thank you for the detailed explanation!

I'm my particular case, your "60" would be potentially (if I get really lucky) 53 for me. So if I revise your chart as follows:

Code: Select all

Age	VLGSX	VTAPX
53 	13%	87%
54 	10%	90%
55 	7%	93%
56 	3%	97%
57 	0%	100%
Well, no. I don't think so, unless you can make your pensions and Social Security start at age 58.

Instead, from what I can gather if you start withdrawals earlier (say, at age 53), they still have to last until age 69. Right?

So your "get really lucky" best case shortens the time horizon a little bit, but not a great deal because an earlier start partly is offset by a longer withdrawal period. Your glide path would, if I understand your plan correctly, look more like this:

Code: Select all

Age	Duration	VLGSX	VTAPX
 50 	 11.0 	53%	47%
 51 	 10.0 	47%	53%
 52 	 9.0 	41%	59%
 53 	 8.0 	34%	66%
 54 	 7.5 	31%	69%
 55 	 7.0 	28%	72%
 56 	 6.5 	25%	75%
 57 	 6.0 	22%	78%
 58 	 5.5 	19%	81%
 59 	 5.0 	16%	84%
 60 	 4.5 	13%	87%
 61 	 4.0 	10%	90%
 62 	 3.5 	7%	93%
 63 	 3.0 	3%	97%
 64 	 2.5 	0%	100%
 65 	 2.0 	0%	100%
 66 	 1.5 	0%	100%
 67 	 1.0 	0%	100%
 68 	 0.5 	0%	100%
 69 	 -   	0%	100%
In my case:
  • let's figure my EE & I Bonds start maturing at 63 - and cover roughly 70% +/- of expenses every year (wish I stated 10 years ago)
  • at 67 my spouses pension kicks in, adding another 25% coverage (with I/EE now covering 95% +/- of expenses)
  • at 70 my pensions and SS [and for now let's assume spouse SS as well, but we might do that sooner], if it makes it easier let's assume I stop the EE/I Bond ladder at 69, but I'm still covering a combined 75% +/- of expenses.
So I basically need to cover 100% of my expenses (ignoring dividends) from age 53 (if I get lucky) to 63. After that, life gets easier. But my "risk" is front loaded into these years...
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

SnowBog wrote: Thu Apr 23, 2020 4:50 pm So I basically need to cover 100% of my expenses (ignoring dividends) from age 53 (if I get lucky) to 63. After that, life gets easier. But my "risk" is front loaded into these years...
Roughly how long until you turn 53?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Noobvestor
Posts: 5644
Joined: Mon Aug 23, 2010 1:09 am

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

Post by Noobvestor »

HEDGEFUNDIE wrote: Thu Apr 23, 2020 11:47 amLet's imagine the Treasury changes the EE bond rules tomorrow that reduce the doubling feature by half, so only 50% gain over 20 years instead of 100%. My expected return in the long term (which I was counting on to retire) just got cut in half. That is not safety to me. And you can dismiss it as simply not "maximizing earnings" but the pernicious thing is that it is happening for the period which you can least afford it, when your human capital is exhausted.
I hope you're not actually suggesting the Treasury would change the rule and apply it retroactively to existing holdings. This would undermine investment confidence in the US Treasury - basically a Black Swan event with implications far beyond EE bonds. I really cannot see that happening, but if it did, the 'safe' bond game would be over - they would be free to change the rules on Treasuries, too.

Maybe you meant going forward, in which case: it is entirely possible to change the rules for new issues (there is precedent here), but shockingly enough, even with your extreme example of a 50% reduction, a 50%-after-20-years bonus will nonetheless have an annualized interest rate equivalent to over 2%, roughly twice the current interest rate of 20-year Treasuries. So if you meant going forward: EE Bonds still win.

If you mean a retroactive change: I'm not sure speculating about that kind of extreme scenario is productive (if even allowed on this forum, since it involves speculating about future legislation), but again, it would not change EE bonds being the investment with a superior expected return. If they change the rules going forward, that wouldn't impact current EE bond holders, just change the calculus on choices going forward.
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
User avatar
anon_investor
Posts: 3501
Joined: Mon Jun 03, 2019 1:43 pm

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

Post by anon_investor »

Noobvestor wrote: Thu Apr 23, 2020 5:08 pm
HEDGEFUNDIE wrote: Thu Apr 23, 2020 11:47 amLet's imagine the Treasury changes the EE bond rules tomorrow that reduce the doubling feature by half, so only 50% gain over 20 years instead of 100%. My expected return in the long term (which I was counting on to retire) just got cut in half. That is not safety to me. And you can dismiss it as simply not "maximizing earnings" but the pernicious thing is that it is happening for the period which you can least afford it, when your human capital is exhausted.
I hope you're not actually suggesting the Treasury would change the rule and apply it retroactively to existing holdings. This would undermine investment confidence in the US Treasury - basically a Black Swan event with implications far beyond EE bonds. I really cannot see that happening, but if it did, the 'safe' bond game would be over - they would be free to change the rules on Treasuries, too.

Maybe you meant going forward, in which case: it is entirely possible to change the rules for new issues (there is precedent here), but shockingly enough, even with your extreme example of a 50% reduction, a 50%-after-20-years bonus will nonetheless have an annualized interest rate equivalent to over 2%, roughly twice the current interest rate of 20-year Treasuries. So if you meant going forward: EE Bonds still win.

If you mean a retroactive change: I'm not sure speculating about that kind of extreme scenario is productive (if even allowed on this forum, since it involves speculating about future legislation), but again, it would not change EE bonds being the investment with a superior expected return. If they change the rules going forward, that wouldn't impact current EE bond holders, just change the calculus on choices going forward.
Didn't EE bonds use to double at 17 years?
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

anon_investor wrote: Thu Apr 23, 2020 5:18 pm
Noobvestor wrote: Thu Apr 23, 2020 5:08 pm
HEDGEFUNDIE wrote: Thu Apr 23, 2020 11:47 amLet's imagine the Treasury changes the EE bond rules tomorrow that reduce the doubling feature by half, so only 50% gain over 20 years instead of 100%. My expected return in the long term (which I was counting on to retire) just got cut in half. That is not safety to me. And you can dismiss it as simply not "maximizing earnings" but the pernicious thing is that it is happening for the period which you can least afford it, when your human capital is exhausted.
I hope you're not actually suggesting the Treasury would change the rule and apply it retroactively to existing holdings. This would undermine investment confidence in the US Treasury - basically a Black Swan event with implications far beyond EE bonds. I really cannot see that happening, but if it did, the 'safe' bond game would be over - they would be free to change the rules on Treasuries, too.

Maybe you meant going forward, in which case: it is entirely possible to change the rules for new issues (there is precedent here), but shockingly enough, even with your extreme example of a 50% reduction, a 50%-after-20-years bonus will nonetheless have an annualized interest rate equivalent to over 2%, roughly twice the current interest rate of 20-year Treasuries. So if you meant going forward: EE Bonds still win.

If you mean a retroactive change: I'm not sure speculating about that kind of extreme scenario is productive (if even allowed on this forum, since it involves speculating about future legislation), but again, it would not change EE bonds being the investment with a superior expected return. If they change the rules going forward, that wouldn't impact current EE bond holders, just change the calculus on choices going forward.
Didn't EE bonds use to double at 17 years?
They did. But bonds previously issued were not changed. To Noobvestor's point, only a "going forward" change is a realistic comparison.
SnowBog
Posts: 360
Joined: Fri Dec 21, 2018 11:21 pm

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

Post by SnowBog »

vineviz wrote: Thu Apr 23, 2020 4:52 pm
SnowBog wrote: Thu Apr 23, 2020 4:50 pm So I basically need to cover 100% of my expenses (ignoring dividends) from age 53 (if I get lucky) to 63. After that, life gets easier. But my "risk" is front loaded into these years...
Roughly how long until you turn 53?
10 years... Hence my comments abount in 10 years needing 10 years of safe assets! :wink:
User avatar
Uncorrelated
Posts: 1052
Joined: Sun Oct 13, 2019 3:16 pm

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

Post by Uncorrelated »

vineviz wrote: Thu Apr 23, 2020 11:05 am
CULater wrote: Thu Apr 23, 2020 10:55 am
watchnerd wrote: Thu Apr 23, 2020 10:38 am
CULater wrote: Thu Apr 23, 2020 9:00 am The main fly in the ointment of LTT + stocks is inflation. Damaging to both, as back in the 1970s. You either have to disregard that threat or try to hedge it in some way to cover all the bases.
I hedge it by having a barbell of LTT + short TIPS.
As explained above, this is a gun with no ammo:

"The same is true for short-term TIPS index funds. According to Vanguard, the correlation of ST TIPS funds to inflation has been around 0.60, which offers decent inflation protection for the money invested in such a fund. But the inflation beta is very low, so the money allocated to ST TIPS fund does very little to protect the remaining assets in the portfolio from inflation."
Both short-term TIPS and the stocks offer pretty good long-run protection from inflation (both expected and unexpected) while the LTTs offer good protection from expected inflation (and unexpected inflation too if the LTTs are inflation-indexed).

So really the investor is, in principle, mainly exposed to unexpected inflation for a small portion of their portfolio. Unless the portfolio is expected to cover a large percentage of non-discretionary expenses (i.e. the investor has a very low Social Security benefit) the overall inflation risk probably isn't extreme.
I'm quite skeptical that short-term TIPS are suitable for long-term inflation protection. The reason I say this is that we were unable to observe the yields of TIPS in an inflationary period. I think it is quite likely that, as soon as the first hints of inflation arise, the yield on fresh TIPS will plummet. No free lunch.

From an efficient market standpoint, there is little reason to expect short-term TIPS to outperform short-term nominal bonds in the average case. In fact I would expect the opposite because efficient market mechanics would demand the more risky (nominal bonds, certainly?) to have higher expected returns.
User avatar
Noobvestor
Posts: 5644
Joined: Mon Aug 23, 2010 1:09 am

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

Post by Noobvestor »

vineviz wrote: Thu Apr 23, 2020 6:07 am
Noobvestor wrote: Wed Apr 22, 2020 7:16 pmPer the post I just made prior to this one, I do think it makes sense to look at bonds from various angles and consider different risks, but that does not seem to be the mentality behind duration-matching, which focuses the entire question of risk on interest rate risk, which is only one of many. Again, too, this focus on just one risk doesn't dovetail with considering how a whole portfolio will perform as a whole.
I don't understand why this confusion persists after repeated, clear explanations about it.

LITERALLY NO ONE IS SAYING THAT OTHER RISKS DON'T EXIST, OR ARE NOT ALSO IMPORTANT.

I hate to scream, but I don't know how to make it more plain than that.
If you don't want to scream, then please just don't. It's also against forum policy: "We also require that you be considerate of our readers and ... avoid posting in ALL CAPITAL LETTERS or otherwise using distracting formatting." rules . But back to the substance of this:

I just can't reconcile your claim that other risks are important with your general advice to go long and match duration, which doesn't account for those risks. If you want to see the source of my confusion regarding your position, here is some context, starting with quotes from you (my bold):
vineviz wrote:Duration matching is a tool for reducing interest rate risk, so any discussion about duration matching is going to focus on interest rate risk.
vineviz wrote:The default solution is to match duration to your investment horizon. If you have an investment horizon of 15 years and bond duration of 15 years, you don't care what happens to interest rates between now and then: you will realize the initial yield-to-maturity.
vineviz wrote:Matching the duration and inflation-sensitivity of your bonds to the duration and inflation-sensitive of the associated expenses will virtually ALWAYS yield the most efficient portfolio. And this kind of matching will certainly be more efficient than any alternative approach.
HEDGEFUNDIE wrote: Thu Apr 23, 2020 7:39 am What vineviz and I have been trying to explain in a dozen different ways is that if your priority is safety and your time horizon is long, there is nothing more safe than holding EDV even though it is volatile
When I read the above, I come away with the following understanding of your views: duration-matching is the optimal starting point. It will almost always produce (in your words:) the most efficient portfolio. And to add HEDGEFUNDIE's more direct assertion: it is the safest option. My point stands: if we're to understand that duration-matching is the most efficient, optimal and safe option, then the focus is interest-rate risk. This means one is not, in fact, considering all parts of a portfolio, but rather focusing on one risk of one part. I don't know how to read it otherwise.

Now, as to back-testing focused on recent history, I would again like to start with a few quotes for context:
HEDGEFUNDIE wrote:The difference between an 80/20 portfolio with the 20 entirely in EDV and the 20 entirely in ITT is over 2% CAGR (or ~25% of total return) since the inception of EDV in late 2007. That’s a hell of an insurance cost.
vineviz wrote:Here are summary statistics for three hypothetical 80/20 portoflios. The first uses cash (30-day Treasury bills) as the bond allocation, the second uses anguard Total Bond Market fund (VBMFX), and the third uses Vanguard Long-Term Treasury (VUSTX). Dates are 1987 to present.
HEDGEFUNDIE: You could hardly cherry-pick a better start point for LTTs if you tried. EDV did exceptionally well in the 08/09 crisis, reasonably well the decade of declining rates and low inflation that followed, then exceptionally again in this most recent crisis. Similarly, vineviz: looking back only to the 1980s is extremely optimal for LTTs, with unprecedented conditions that can't repeat going forward. It needs to span a much longer history, including rising-rate periods of the 1940s through 70s with their associated higher inflation. I understand that's as far back as portfolio visualizer goes, but per your own modeling, a rising-rate situation (askin to the 40s-70s) would leave LTTs with very poor relative returns.

Image

I don't know what the future holds. Maybe we're at low rates forever, in which case most safe bonds look dismal. Series I and EE bonds look better than TIPS and Treasuries, at least, but they have annual purchase limits. I personally am not willing to lock up money in 20-year Treasury or TIPS bonds with their presently low yields. I see correlation advantages, to be sure, but at current rates, they hold no appeal for me. YMMV. I do continue to believe that there is a lot of risk (including: rising rates/inflation) in going long on nominal Treasuries, particularly after a big run-up and record-low rates. In short: I do not believe LTTs are necessarily optimal, efficient or safe when one considers all risks involved. /2 cents
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
_Soundwave_
Posts: 23
Joined: Sun Mar 22, 2020 8:00 pm

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

Post by _Soundwave_ »

This thread has been fantastic. I’m in my mid 30’s. With duration matching in mind, Do extended duration funds vs LTT matter all that much for someone with 20+ years to go?
Last edited by _Soundwave_ on Thu Apr 23, 2020 6:26 pm, edited 1 time in total.
User avatar
Noobvestor
Posts: 5644
Joined: Mon Aug 23, 2010 1:09 am

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

Post by Noobvestor »

anon_investor wrote: Thu Apr 23, 2020 5:18 pm Didn't EE bonds use to double at 17 years?
They did indeed. And the Treasury might well decide once again to change the doubling rules for future issues. But that does not pose a risk to existing holders of EE bonds. If they did change the doubling rules again, even radically (like: cutting the doubling to fifty-percenting) it might also still be the case that EE bonds provide higher effective rates than LTTs - that's how much more they yield right now. :idea:
"In the absence of clarity, diversification is the only logical strategy" -= Larry Swedroe
User avatar
Topic Author
vineviz
Posts: 7833
Joined: Tue May 15, 2018 1:55 pm

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

Post by vineviz »

Noobvestor wrote: Thu Apr 23, 2020 5:56 pm I just can't reconcile your claim that other risks are important with your general advice to go long and match duration, which doesn't account for those risks.
At this point I have no explanation for this except for a strong desire to not understand.

Noobvestor wrote: Thu Apr 23, 2020 5:56 pmWhen I read the above, I come away with the following understanding of your views: duration-matching is the optimal starting point.
Yes: it's the optimal starting point. Start there, then move on to deal with other risk (inflation, credit, etc.)

Or, you know what, don't start there. Start with inflation risk or credit risk and THEN duration match. The order isn't important. Do them all.

In this thread, I'm just trying to help people do this one thing.

Noobvestor wrote: Thu Apr 23, 2020 5:56 pm HEDGEFUNDIE: You could hardly cherry-pick a better start point for LTTs if you tried. EDV did exceptionally well in the 08/09 crisis, reasonably well the decade of declining rates and low inflation that followed, then exceptionally again in this most recent crisis. Similarly, vineviz: looking back only to the 1980s is extremely optimal for LTTs, with unprecedented conditions that can't repeat going forward.
Except that neither of us is telling people to use duration matching to juice their returns, and we are DEFINITELY not telling people to extrapolate past returns into the future. We're both on record saying that we have no skill at forecasting future yields and that investors should NOT try to use market timing in the bond markets.

The people who are claiming an ability to predict the future are the ones who obstinately cling to visions of rates that are more likely to go up from here than down, or a view that there is is more room for yields to increase than to decrease. If you have a beef with anyone on this count, its with the market timing folks who say things like "I'm not willing to lock up money in 20-year Treasury or TIPS bonds with their presently low yields".
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Post Reply