First 20% of bonds in long-term Treasuries

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Locked
BigJohn
Posts: 2627
Joined: Wed Apr 02, 2014 11:27 pm

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

Post by BigJohn »

vineviz wrote: Sun Apr 10, 2022 8:02 am
secondopinion wrote: Fri Mar 18, 2022 10:33 am
BigJohn wrote: Thu Mar 17, 2022 9:16 pm
secondopinion wrote: Thu Mar 17, 2022 4:21 pm It is insurance for nominal expenses in the far future; that is what the comment is suggesting. With such nominal expenses, they are risk-free even with inflation and low yields because they are hedging out the expenses completely or close to it.

Without such nominal expenses, it is a bold speculation on the long-term future of the US dollar. Diversification is true either way you look at it; the question is whether one should be taking the risks with such a speculation.
Sure but what expenses does an individual have that are in nominal dollars that far out in the future? The only one I can think of is early in a long term fixed rate mortgage. In my mind this is hardly enough for most people to justify the first 20% of bonds in LTTs.

Insurance companies have a lot of long term nominal expenses which is why they use LTTs as an appropriate risk matching strategy. However, they know that inflation is a wild card they can't predict. I think that's why you can no long purchase a COLA adjusted annuity, it's just not a risk that they want to take.

This lack of nominal expenses with the potential for purchasing power erosion if high and unexpected inflation arose is just the reason that I never bought into this strategy of using LTTs. For those that did, I hope the Fed gets inflation under control relatively quickly before the erosion gets any worse.
Nominal expenses in the far future for an individual are rare. I personally think that long-term treasuries are more speculative than a safe investment; most of the time, you are getting beta, which may work to one's advantage or disadvantage. Without any ability to adjust to inflation, long-term treasuries are also very speculative even if held to maturity. I can see the speculative value, but most here are trying to avoid speculation.
I have three thoughts on this:

One is that the topic of this thread mentions "long-term Treasuries", not "long-term nominal Treasuries". TIPS are Treasuries too.

The second thought is that the rule of thumb is to put the first 20% of the portfolio into long-term Treasuries. Any portfolio which is 80%+ invested in stocks still has virtually all of its risk tied to the performance of the stock sleeve. And historically nominal Treasuries are a better diversifier than TIPS because the total return of TIPS is partially tied to economic growth, like stocks are. Plus, generally speaking, investors with high equity concentrations tend to be younger and naturally have inflation protection elsewhere in their portfolio (e.g. human capital, fixed rate mortgages, and so forth). Investors with more than 20% of their portfolio in bonds should increasingly focus on the risk exposure of the bonds themselves. Maybe that means the next 20% (or whatever) should be long-termTIPS, or short-term corporate bonds, or total bond market, etc.

The third thought is that the idea that nominal expenses are rare is a bit of a red herring, because it mistakes the price of goods as the thing we're trying to hedge. In truth, it is the amount of consumption that we are trying to hedge. It might be easy to assume that prices and consumption have the same inflation beta, but all the empirical evidence we have about retiree spending tells us differently. Retirement consumption does NOT have an inflation beta of 1, and for many retirees it is far less than one. So once you factor Social Security benefits into the picture, the consumption that must be funded by portfolio withdrawals is - for many, though certainly no all, investors - effectively nominal at least to a large degree.
I understand and agree with your first two points. Unfortunately, in my reading of this discussion I think many made the leap to long term nominals rather than TIPS at just the wrong time. Same with some older people nearing retirement that do potential have a long enough time horizon but far less inflation protection elsewhere.

Not sure I fully understand your third point so let me repeat what I think I read in different words. You seem to be saying that losing purchasing power to inflation isn’t as big a problem in later retirement because our consumptions goes down to offset the nominal loss. If that’s your point then while I acknowledge the math as being directional valid for most, I’m not sure that will hold for multiple years of 5+% inflation. It’s also not something that I think many/most are emotionally prepared to accept.
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
hudson
Posts: 7119
Joined: Fri Apr 06, 2007 9:15 am

Many retirees are not overly affected by inflation.

Post by hudson »

vineviz wrote: Sun Apr 10, 2022 8:02 am
The third thought is that the idea that nominal expenses are rare is a bit of a red herring, because it mistakes the price of goods as the thing we're trying to hedge. In truth, it is the amount of consumption that we are trying to hedge. It might be easy to assume that prices and consumption have the same inflation beta, but all the empirical evidence we have about retiree spending tells us differently. Retirement consumption does NOT have an inflation beta of 1, and for many retirees it is far less than one. So once you factor Social Security benefits into the picture, the consumption that must be funded by portfolio withdrawals is - for many, though certainly no all, investors - effectively nominal at least to a large degree.
Thanks! Useful information!

Beta of 1: I had to look it up. I think that means that many retirees aren't overly affected by inflation.
(Beta of 1 means strongly correlated.)

I'm probably a retiree that is somewhat but not overly affected by inflation.
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

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

Post by CletusCaddy »

BigJohn wrote: Sun Apr 10, 2022 9:09 am Not sure I fully understand your third point so let me repeat what I think I read in different words. You seem to be saying that losing purchasing power to inflation isn’t as big a problem in later retirement because our consumptions goes down to offset the nominal loss. If that’s your point then while I acknowledge the math as being directional valid for most, I’m not sure that will hold for multiple years of 5+% inflation. It’s also not something that I think many/most are emotionally prepared to accept.
It’s not just that consumption goes down but also that a big part of your income is Social Security which tracks CPI
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

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

Post by CletusCaddy »

Setting TIPS aside, long term nominal bonds are still the least risky way to meet long term real obligations. Everything in investing is based on expected future outcomes. Matching more elements of expectation will always be less risky than matching fewer elements of expectation.

Here is an analogy. You need to commute to work in the morning. There is an expected time you need to arrive, an expected speed you need to drive under expected traffic conditions. Every single day you will turn out to be “wrong” about these expectations. Traffic is never the same every day, your car might break down, your boss may not show up that day and so you might arrive realizing you didn’t even need to go in at all. But none of this uncertainty means that there is not a “right” time to leave the house in the morning.

Let’s carry the analogy further. You have two cars to choose from to make (what is typically) a 30 mile commute in (what is typically) 30 minutes:

1. A car that goes a maximum of 60 mph
2. A car that has strange engine condition that changes its maximum speed every few minutes. Sometimes you’ll only get to go 10 mph, sometimes you’ll get to go 90 mph. The only thing you know for sure is that the car will start at a speed slower than 60mph

Which car is more likely to get you to your destination on time?
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 »

BigJohn wrote: Sun Apr 10, 2022 9:09 am
vineviz wrote: Sun Apr 10, 2022 8:02 am
I understand and agree with your first two points. Unfortunately, in my reading of this discussion I think many made the leap to long term nominals rather than TIPS at just the wrong time. Same with some older people nearing retirement that do potential have a long enough time horizon but far less inflation protection elsewhere.
I want to push back a little on the “ at just the wrong time” part, as it implies that there are right times and wrong times to attempt to time the market. If you’re buying the right funds for you, there’s no wrong time to do that IMHO.

Has the recent drop in the price of VGLT and EDV been unpleasant? I’m sure it has been for many. It has also been an opportunity to rebalance into bonds at a higher yield than we were getting before.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Robot Monster
Posts: 4215
Joined: Sun May 05, 2019 11:23 am

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

Post by Robot Monster »

garlandwhizzer wrote: Thu Mar 17, 2022 1:31 pm I personally believe in diversification in bonds as well as diversification in stocks in order to balance the conflicting goals of equity diversification on the one hand (LTT best) and stability of principal plus resistance to long term inflation on the other (shorter duration, TIPS best). Important to remember that LTT underperformed T-bills in real terms for the 40 years between 1940 -1980.
garlandwhizzer wrote: Thu Apr 07, 2022 6:37 pm The long duration portfolio doesn't turn over quickly to take advantage of higher bond yields as inflation rises, so you're stuck with those low yielding long bonds for many long years before they mature. When they do mature you get back your principal back but that principal is decreased in real terms by about 20 years of cumulative inflation. Not a good situation.
Reading over your posts, and taking into consideration what vineviz has said, I have some thoughts. Above you highlight the problem of inflation for long-term nominals. My understanding is, in an inflationary environment, it's not that the 20% LTT will be resilient to inflation, it's that the rest of the portfolio will: the 80% in stocks, as well as the human capital, fixed rate mortgages, and so forth which younger people have. High inflation may very well mean high economic growth, which would be good for stocks and those employed.

So, while it may be true that LTT didn't perform so nicely in the time period you mention above from 1940-1980, the S&P apparently, for the most part, did. Historical data for that time period can be seen in an excellent graphic from the New York Times. article link

Now then. What if we enter a prolonged period of high inflation coupled with low growth? I actually don't have an answer for that one.
BigJohn
Posts: 2627
Joined: Wed Apr 02, 2014 11:27 pm

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

Post by BigJohn »

vineviz wrote: Sun Apr 10, 2022 9:34 am
BigJohn wrote: Sun Apr 10, 2022 9:09 am
vineviz wrote: Sun Apr 10, 2022 8:02 am
I understand and agree with your first two points. Unfortunately, in my reading of this discussion I think many made the leap to long term nominals rather than TIPS at just the wrong time. Same with some older people nearing retirement that do potential have a long enough time horizon but far less inflation protection elsewhere.
I want to push back a little on the “ at just the wrong time” part, as it implies that there are right times and wrong times to attempt to time the market. If you’re buying the right funds for you, there’s no wrong time to do that IMHO.

Has the recent drop in the price of VGLT and EDV been unpleasant? I’m sure it has been for many. It has also been an opportunity to rebalance into bonds at a higher yield than we were getting before.
Fair point and in principle I totally agree. I have not gone the LTT route so not a worry for me no matter the direction VGLT/EDV go. I guess I was reflecting the “buyers remorse” emotion that some seem to be feeling right now. For those that made this change, stay the course and rebalance may be the right decision. However, these recent developments will make that very challenging for many.
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
Frank2012
Posts: 194
Joined: Fri Jan 25, 2013 6:16 pm

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

Post by Frank2012 »

bgf wrote: Sun Apr 10, 2022 6:33 am
slicendice wrote: Sat Apr 09, 2022 6:56 pm
garlandwhizzer wrote: Sat Apr 09, 2022 4:52 pm As for employment, at the end of 2021, there were 4.35 million fewer non farm payroll workers in America than 2 years before, pre-pandemic at the end of 2019. We lost a huge amount of jobs in the pandemic and we have yet to fully make up for that. Economic disruption, supply chain imbalances, pandemic restrictions, inflation, and the Great Resignation have been impactful and are still ongoing to some extent. These impact employment as is clear from all the help wanted signs in business windows.

In terms of GDP growth, clearly we're not in a recession, but interest rate hikes and restrictive monetary policy which is coming soon might well precipitate one. Real economic growth in the last 2 years has been sluggish (meeting my definition of stagflation derived from the word stagnant). Even with zero rate policy economic growth was a bit less than half the annual long term average, but not negative. Our economy is, however, currently IMO more vulnerable than many economists believe. I could be wrong on that, and I hope I am. It is, however, hard for me to believe that we're gong to wiggle our way out of 15 years of excessive monetary policy stimulus, massive deficit spending, QE, and the excesses of "modern monetary policy" without significant damage. Maybe I'm just old and out of touch, but no country has ever taken all these aggressive stimulatory steps simultaneously. Therefore there is no play book telling us how unwinding the mountain of debt, the loss of a FED backstop, and the impact of higher rates will all play out in the end. The base case is that we'll muddle through as always, but policy missteps now and geopolitical events may add uncertainty.

MMT was great to stimulate the economy when we needed it desperately, but its answer for what to do when rising inflation occurs is to run federal budgetary surpluses to reduce the deficit. Good luck on doing that politically. That is to say MMT has no answer for inflation like now. It merely offered us an academically created happy tune to sing when the economy tanked and kept us from worrying about mounting debt.

To reverse inflation with policy you have to raise rates, get rid of QE, shrink the FED balance sheet, reduce deficit spending, and all these things increase risks to near term economic growth. The less damaging non-policy approach is to relieve supply chain constraints, settle geopolitical issues including economically damaging sanctions, increase efficiency producing innovation and restore to international trade and globalization. All those things are anti-inflationary. Politically, that looks like a very tall order order right now. I hope we and the rest of the world are up to this challenge.
That is quite the rant. At best I find this non-actionable. It also has almost nothing to do with the topic of the thread, so I think I will just leave it alone so hopefully the thread doesn't get locked.
This is the entire forum. Everyone is now a bond expert macro economist. The boglehead maxim "nobody knows nothing" apparently only applies to public companies.

Apparently the crystal balls are clear with respect to future interest rate moves, inflation, long term bond performance, global trade, war, fiscal and monetary policy, and demographics...

If we're going to predict that stuff why not go back to future earnings. A lot fewer variables there...

It's a joke. An absolute joke.

Sure, the world is going to hell in a handbasket...but do you sell treasuries for cash, sell your equites for cash, or stay the course with your asset allocation and follow your IPS?

I'm staying the course and following my IPS, come what may....
er999
Posts: 1343
Joined: Wed Nov 05, 2008 10:00 am

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

Post by er999 »

Seems like the long term treasurys vs short / medium term is similar to the debate of total bond vs only treasurys. Some say go 100% treasurys take your risk on the stock side.

I can easily see someone closer or just starting retirement going short / medium term bonds only to avoid as much fluctuation in portfolio value even though long term treasurys might be more optimal. They could have, say, a 5-10% higher stock allocation to compensate for this (I don’t know what the percentage increase would need to be just guessing to make a point). Definitely long term treasurys are more volatile in the short run — year to date vanguard long term treasurys is down 13% compared to total bond down 8%. In the 2008 crisis people may have felt differently as I think long term treasurys were up 25%then.

I personally have 10% long term treasurys/ 90% stock but don’t know whether 100% stocks would be better for the accumulation phase. I was convinced by this thread and reading the portfolio charts articles. Once I reach my number (or close to it) and want to focus more on preservation I’m unsure whether first 20% of long term treasurys is the best vs shorter term bonds but I appreciate all of vineviz’s posts and raising interesting ideas. Should someone be in a higher stock percentage (say 70% stocks, 30% total bonds) instead of a lower stock percentage with 20% long term treasurys (say 60% stocks, 20% long term treasurys, 20% intermediate term treasurys)? Probably impossible to say in advance.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

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

Post by secondopinion »

vineviz wrote: Sun Apr 10, 2022 8:02 am
secondopinion wrote: Fri Mar 18, 2022 10:33 am
BigJohn wrote: Thu Mar 17, 2022 9:16 pm
secondopinion wrote: Thu Mar 17, 2022 4:21 pm It is insurance for nominal expenses in the far future; that is what the comment is suggesting. With such nominal expenses, they are risk-free even with inflation and low yields because they are hedging out the expenses completely or close to it.

Without such nominal expenses, it is a bold speculation on the long-term future of the US dollar. Diversification is true either way you look at it; the question is whether one should be taking the risks with such a speculation.
Sure but what expenses does an individual have that are in nominal dollars that far out in the future? The only one I can think of is early in a long term fixed rate mortgage. In my mind this is hardly enough for most people to justify the first 20% of bonds in LTTs.

Insurance companies have a lot of long term nominal expenses which is why they use LTTs as an appropriate risk matching strategy. However, they know that inflation is a wild card they can't predict. I think that's why you can no long purchase a COLA adjusted annuity, it's just not a risk that they want to take.

This lack of nominal expenses with the potential for purchasing power erosion if high and unexpected inflation arose is just the reason that I never bought into this strategy of using LTTs. For those that did, I hope the Fed gets inflation under control relatively quickly before the erosion gets any worse.
Nominal expenses in the far future for an individual are rare. I personally think that long-term treasuries are more speculative than a safe investment; most of the time, you are getting beta, which may work to one's advantage or disadvantage. Without any ability to adjust to inflation, long-term treasuries are also very speculative even if held to maturity. I can see the speculative value, but most here are trying to avoid speculation.
I have three thoughts on this:

One is that the topic of this thread mentions "long-term Treasuries", not "long-term nominal Treasuries". TIPS are Treasuries too.

The second thought is that the rule of thumb is to put the first 20% of the portfolio into long-term Treasuries. Any portfolio which is 80%+ invested in stocks still has virtually all of its risk tied to the performance of the stock sleeve. And historically nominal Treasuries are a better diversifier than TIPS because the total return of TIPS is partially tied to economic growth, like stocks are. Plus, generally speaking, investors with high equity concentrations tend to be younger and naturally have inflation protection elsewhere in their portfolio (e.g. human capital, fixed rate mortgages, and so forth). Investors with more than 20% of their portfolio in bonds should increasingly focus on the risk exposure of the bonds themselves. Maybe that means the next 20% (or whatever) should be long-termTIPS, or short-term corporate bonds, or total bond market, etc.

The third thought is that the idea that nominal expenses are rare is a bit of a red herring, because it mistakes the price of goods as the thing we're trying to hedge. In truth, it is the amount of consumption that we are trying to hedge. It might be easy to assume that prices and consumption have the same inflation beta, but all the empirical evidence we have about retiree spending tells us differently. Retirement consumption does NOT have an inflation beta of 1, and for many retirees it is far less than one. So once you factor Social Security benefits into the picture, the consumption that must be funded by portfolio withdrawals is - for many, though certainly no all, investors - effectively nominal at least to a large degree.
Most of the time, people use the term "treasuries" and "TIPS" for the nominal and inflation-adjusted securities respectively. I am not going to argue that either one is bad.

As far as nominal treasuries being better for high stock allocations, I can total agree with that since too many things are market driven. I am not disagreeing of the hedge. Given the studies, I entirely see that the lower beta means buying nominal bonds; consumption is probably the right thing to hedge. In light of that, I would give my approval to long-term nominal bonds as a better hedge for consumption. I retract my statement of long-term nominal bonds as being mostly speculative.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

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

Post by secondopinion »

er999 wrote: Sun Apr 10, 2022 11:06 am Seems like the long term treasurys vs short / medium term is similar to the debate of total bond vs only treasurys. Some say go 100% treasurys take your risk on the stock side.

I can easily see someone closer or just starting retirement going short / medium term bonds only to avoid as much fluctuation in portfolio value even though long term treasurys might be more optimal. They could have, say, a 5-10% higher stock allocation to compensate for this (I don’t know what the percentage increase would need to be just guessing to make a point). Definitely long term treasurys are more volatile in the short run — year to date vanguard long term treasurys is down 13% compared to total bond down 8%. In the 2008 crisis people may have felt differently as I think long term treasurys were up 25%then.

I personally have 10% long term treasurys/ 90% stock but don’t know whether 100% stocks would be better for the accumulation phase. I was convinced by this thread and reading the portfolio charts articles. Once I reach my number (or close to it) and want to focus more on preservation I’m unsure whether first 20% of long term treasurys is the best vs shorter term bonds but I appreciate all of vineviz’s posts and raising interesting ideas. Should someone be in a higher stock percentage (say 70% stocks, 30% total bonds) instead of a lower stock percentage with 20% long term treasurys (say 60% stocks, 20% long term treasurys, 20% intermediate term treasurys)? Probably impossible to say in advance.
I am actually short-term and long-term fixed-income (I do not carry the intermediate) for the bond convexity. I do not reject corporate bonds since I still feel that there is a solid case for them versus holding stock and treasuries only.

What is "safe" for the investor is often different from what is "best"; that is where we hope that taking the extra risk will better help us meet our goals.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
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 »

Since this thread was started in August 2019 the yield on 20-year US Treasury bonds has increased roughly 110bps, from roughly 2.1% to 3.2%, so I thought it'd be worth comparing the returns since then of three 80/20 portfolios: one using a total bond market fund for the 20% one using long-term nominal Treasuries, and one using long-term TIPS.

It was not unusual at the time to hear two refrains: 1) bond yields have nowhere to go but up; and 2) increasing yields will crush long-term bond investors. Bond yields have, it happens, gone up so let's take a look at outcomes.

From 8/1/2019 through 4/20/2022 the three different portfolios have produced the following results.

80% VT 20% BND: 10.40% CAGR, 14.22% std. deviation, 0.73 Sharpe ratio.
80% VT 20% VGLT: 10.48% CAGR, 13.49% std. deviation, 0.77 Sharpe ratio.
80% VT 20% LTPZ: 11.97% CAGR, 14.68% std. deviation, 0.81 Sharpe ratio.

Despite a significant increase in yields, the portfolios using long-term bonds have arguably outperformed the portfolio using BND. The portfolio using long-term nominal Treasuries had slightly better growth and less volatility. The portfolio using long-term TIPS had notably higher growth and slightly more volatility. In both cases the volatility-adjusted returns (i.e. Sharpe ratios) were improved modestly.

Image
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
er999
Posts: 1343
Joined: Wed Nov 05, 2008 10:00 am

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

Post by er999 »

Interesting result and extending back the same portfolio to 2010 (earliest date for VGLT) and the VGLT portfolio beats BND more significantly. I changed to annual rebalancing from monthly as I feel that’s more typical for a boglehead portfolio and $10k start Feb 2010 would be $27,958 for 20% BND, $30,153 for 20% VGLT, and $30,555 for 20% LTPZ (sorry I don’t know how to add graphs). An investor would have 7.8% more after 12 years if using VGLT rather than BND.

To get the same return (from 2/2010 - 3/31/22) as a 80% VT / 20% VGLT investor as BND investor would have to be 90% VT / 10% BND (89% would have been just under the return, 90% beats it up $42 so about the same.
ambient
Posts: 10
Joined: Tue Jan 25, 2022 10:48 pm

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

Post by ambient »

er999 wrote: Thu Apr 21, 2022 11:32 amI changed to annual rebalancing from monthly as I feel that’s more typical for a boglehead portfolio
Long-term bonds are great, if annually is also when you even look at your portfolio.

Holding a mix of stocks and long-term bonds is like riding a bronco, who keeps saying, "I promise this will be all right in the end!" For your entire life.

I used to wonder why more people didn't invest in long-term bonds. People said, "Bonds are the part of my portfolio that help me not panic-sell." With long-term bonds, that is harder. They seem just as volatile as stocks sometimes. Actually this year to date, VGLT is down 17%, while VT and VTI are down only 8%.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Thu Apr 21, 2022 10:03 am Despite a significant increase in yields, the portfolios using long-term bonds have arguably outperformed the portfolio using BND. The portfolio using long-term nominal Treasuries had slightly better growth and less volatility. The portfolio using long-term TIPS had notably higher growth and slightly more volatility. In both cases the volatility-adjusted returns (i.e. Sharpe ratios) were improved modestly.
Isn't that because BND is also a nominal intermediate fund, and also got hit by both inflation being unexpectedly high and the yield curve steepening?

Also, long-term rates dropped from 2019 to 2020, and I think only got consistently above where they were in summer of 2019 as of the beginning of 2022. Not coincidentally, inflation was also below 2019 levels in 2020, and only started spiking above in Q2 2021.

So, this is testing a relatively small allocation of bonds, with a relatively small difference in bond type, over a period in which rate and inflation trends were mostly still favorable for longer-term nominal bonds.

If you instead tested 20% VGLT against something like 20% TIP, and starting a year later around the bottom of rates/inflation, I assume 20% TIP would do a lot better. And 20% VTIP better still.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

ambient wrote: Thu Apr 21, 2022 11:48 am They seem just as volatile as stocks sometimes. Actually this year to date, VGLT is down 17%, while VT and VTI are down only 8%.
The idea was US stocks and LT nominal USD bonds were supposed to move oppositely.

And I gather a lot of people overlooked that is only likely to be true in deflationary/disinflationary stock crises.

In an inflationary US stock crisis, LT nominal USD bonds are likely to be heading the same direction as US stocks. As we are now seeing.
garlandwhizzer
Posts: 3565
Joined: Fri Aug 06, 2010 3:42 pm

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

Post by garlandwhizzer »

Backtesting is not IMO the most reliable way to figure out future volatility adjusted returns of LT bonds versus ST or IT bonds. There is no doubt whatsoever that LT outperformed ST and IT consistently and also provided greater diversification to equity volatility for the 38 years from 1982 to 2020. That is not the issue. LTT backtests beautifully for decades prior to 2020.

The issue is: after 38 years of ever decreasing inflation and bond yields--a period in which LT clearly outperformed and also offer better equity diversification relative to shorter durations--are we now entering an entirely new period? A period of increasing inflation and increasing rates--a scenario is which duration exposure reduces real returns and increases volatility relative to ST or IT? Quite clearly we are in that latter situation now and the only question is how long it will last? Those holding 100% LTT have been savaged in the last 2 years, a hard lesson that a big time long term winner can quickly become a big time loser.

Negative real returns for LTT lasted for 40 years between 1942 and 1982 as inflation and higher rate momentum became self sustaining. LTT underperformed MMF during that period. Past history shows that it can go on for decades. Most observers now believe it will only be temporary, just a head fake, lasting years but slowly receding. Whether we will get back to the former 2% FED inflation goal, however, is unclear. Currently the labor market is very tight and wages as well as commodity prices are rising. The point is that no one knows reliably how the future will play out. In the case of bonds, backtesting now is the IMO the least reliable place to seek answers. I personally believe the current high level of uncertainty is a good reason for wide diversification in duration and the inclusion of some TIPS in the bond portfolio rather than loading up on LTT because it did so well in past decades. Just my point of view.

Garland Whizzer
SnowBog
Posts: 4699
Joined: Fri Dec 21, 2018 10:21 pm

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

Post by SnowBog »

er999 wrote: Thu Apr 21, 2022 11:32 am Interesting result and extending back the same portfolio to 2010 (earliest date for VGLT) and the VGLT portfolio beats BND more significantly. I changed to annual rebalancing from monthly as I feel that’s more typical for a boglehead portfolio and $10k start Feb 2010 would be $27,958 for 20% BND, $30,153 for 20% VGLT, and $30,555 for 20% LTPZ (sorry I don’t know how to add graphs). An investor would have 7.8% more after 12 years if using VGLT rather than BND.

To get the same return (from 2/2010 - 3/31/22) as a 80% VT / 20% VGLT investor as BND investor would have to be 90% VT / 10% BND (89% would have been just under the return, 90% beats it up $42 so about the same.
Interesting. So much of this seems to depend on timing...

In my 401k, I've been mostly FXNAX (aka BND) since moving to a "3-fund" portfolio in 2018-2019. Last year, started moving some to FNBGX (aka VGLT). At current, my FXNAX is down roughly 5.6% (from basis) and FNBGX is down roughly 17.4% (from basis).

Had I moved to LTT sooner or later (or not at all) I'd be ahead...
ambient wrote: Thu Apr 21, 2022 11:48 am Holding a mix of stocks and long-term bonds is like riding a bronco, who keeps saying, "I promise this will be all right in the end!" For your entire life.
I like this analogy!!

In hindsight, I wasn't ready to ride this bronco... It still seems unreal to me that my LTT often goes up-or-down more in a day than my stocks do... And there's no way I could sustain that ride if these were my first bonds...

But, LTT is only part of my bonds (and my last part). I can look at my I & EE Bonds for complete stability and my "total bond" for more "normal" (if such exists) bonds.

And giving my investing horizon is long term, and LTT is just part of the overall, i remain optimistic it works out favorably in the end (provided I can hold the reigns).
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 »

NiceUnparticularMan wrote: Thu Apr 21, 2022 1:03 pm
vineviz wrote: Thu Apr 21, 2022 10:03 am Despite a significant increase in yields, the portfolios using long-term bonds have arguably outperformed the portfolio using BND. The portfolio using long-term nominal Treasuries had slightly better growth and less volatility. The portfolio using long-term TIPS had notably higher growth and slightly more volatility. In both cases the volatility-adjusted returns (i.e. Sharpe ratios) were improved modestly.
Isn't that because BND is also a nominal intermediate fund, and also got hit by both inflation being unexpectedly high and the yield curve steepening?
I’m sure that partly explains the outcome.

The main point I’m hoping to make is that many people seemed to be SURE that long bonds would be “crushed” if yields increased significantly, but it’s usually not that simple.

As you note, the curve flattened but that’s TYPICALLY what happens since short term yields are much more volatile than long-term yields. In other words, it’s usually not that simple.

I don’t want to dwell on short term results, and I know that YTD bond losses are freaking some people out as well.

But I hope this illustrates that you can be RIGHT in predicting that yields will go up and still get the market timing call (eg “shorten duration”) wrong.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Thu Apr 21, 2022 2:01 pm The main point I’m hoping to make is that many people seemed to be SURE that long bonds would be “crushed” if yields increased significantly, but it’s usually not that simple.
Yeah, the not simple part is understanding exactly why nominal yields on long-term USD bonds are increasing, and what can happen next.

But some of the scenarios like that are bad for long-term nominal bonds. Like the scenario since 2020.
As you note, the curve flattened but that’s TYPICALLY what happens since short term yields are much more volatile than long-term yields.
Well, in this case, over the last couple years the real yield curve has moved more at the long end. This is why, say, ST TIPS are doing better than LT TIPS over this period.

Volatility, though, is not necessarily the right term for this, because these real yield curve dynamics can more look like secular trends. At least with the benefit of hindsight . . . .
But I hope this illustrates that you can be RIGHT in predicting that yields will go up and still get the market timing call (eg “shorten duration”) wrong.
Indeed. There was always a risk of an inflationary stock crisis. There was always a risk of a steepening real yield curve. But I personally had no idea in advance when this would happen, and I also have no idea now what will happen next.

So, good for people to be reminded of these risks. Bad if they think this makes a case for market-timing.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

vineviz wrote: Thu Apr 21, 2022 10:03 am Since this thread was started in August 2019 the yield on 20-year US Treasury bonds has increased roughly 110bps, from roughly 2.1% to 3.2%, so I thought it'd be worth comparing the returns since then of three 80/20 portfolios: one using a total bond market fund for the 20% one using long-term nominal Treasuries, and one using long-term TIPS.

It was not unusual at the time to hear two refrains: 1) bond yields have nowhere to go but up; and 2) increasing yields will crush long-term bond investors. Bond yields have, it happens, gone up so let's take a look at outcomes.

From 8/1/2019 through 4/20/2022 the three different portfolios have produced the following results.

80% VT 20% BND: 10.40% CAGR, 14.22% std. deviation, 0.73 Sharpe ratio.
80% VT 20% VGLT: 10.48% CAGR, 13.49% std. deviation, 0.77 Sharpe ratio.
80% VT 20% LTPZ: 11.97% CAGR, 14.68% std. deviation, 0.81 Sharpe ratio.

Despite a significant increase in yields, the portfolios using long-term bonds have arguably outperformed the portfolio using BND. The portfolio using long-term nominal Treasuries had slightly better growth and less volatility. The portfolio using long-term TIPS had notably higher growth and slightly more volatility. In both cases the volatility-adjusted returns (i.e. Sharpe ratios) were improved modestly.

Image
100% VT crushed all of them, albeit only in terms of performance, not sharpe. The longer the time horizon, the more "100% equities" dominates everything else. For shorter horizons or leveraged portfolios, the sharpe matters more. For non-leveraged-constrained portfolios, equities + leveraged ITT (intermediate term treasuries) outperforms other combinations including 80/20 on risk-adjusted basis. There is an mHFEA thread in this forum.
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 »

comeinvest wrote: Thu Apr 21, 2022 8:16 pm 100% VT crushed all of them, albeit only in terms of performance, not sharpe.
Color me unsurprised: stocks usually have higher returns than bonds.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Morik
Posts: 1344
Joined: Tue Nov 25, 2014 11:26 am

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

Post by Morik »

How does rebalancing affect the overall portfolio in a rising rate environment when holding LTTs?

E.g., in a steadily rising market I read that holding for 2x the duration minus one year 'breaks even' with the capital losses of the constant rising rates. If you are also taking money out of stocks (with a higher expected return) to buy more bonds as rates rise, how does that affect the overall portfolio?

Rebalancing into bonds as capital losses occur would mean getting in at the higher rates, but sacrificing the higher expected returns of stock. How does this compare with non-leveraged ITTs over a long time period (say, 40-50 years) where rates are generally rising?
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 »

Morik wrote: Thu Apr 21, 2022 8:56 pm How does rebalancing affect the overall portfolio in a rising rate environment when holding LTTs?
The illustration I provided shows precisely this: rebalancing a portfolio with LTTs when rates rose.

When bond yields go up, all else equal, the expected future returns of both stocks and bonds go up. The process is noisy, of course, and unexpected returns (either positive or negative) are likely. But rebalancing keeps the risks of the portfolio where you want them, so it's smart to keep rebalancing regardless of whether you think your'e in a rising rate environment or a falling rate environment.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
One More Thing
Posts: 107
Joined: Mon Sep 20, 2021 10:02 pm

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

Post by One More Thing »

Pardon me for skipping straight to the last page thus far but tell me if I understand the logic of this strategy. Is the point to set up your bond allocation for as much safe compounding returns as possible while you are accumulating wealth? As you age you increase your investment horizon shortens and your bond allocation rises so you invest in shorter and shorter duration treasuries in order to preserve your wealth over compounding interest. Is that the idea?
000
Posts: 8211
Joined: Thu Jul 23, 2020 12:04 am

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

Post by 000 »

comeinvest wrote: Thu Apr 21, 2022 8:16 pm 100% VT crushed all of them, albeit only in terms of performance, not sharpe. The longer the time horizon, the more "100% equities" dominates everything else. For shorter horizons or leveraged portfolios, the sharpe matters more. For non-leveraged-constrained portfolios, equities + leveraged ITT (intermediate term treasuries) outperforms other combinations including 80/20 on risk-adjusted basis. There is an mHFEA thread in this forum.
Is it really true that stocks have less risk the longer the time horizon?

Longer time horizon -> more opportunities for a black swan to come along and wipe the slate clean.
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

000 wrote: Thu Apr 21, 2022 10:03 pm
comeinvest wrote: Thu Apr 21, 2022 8:16 pm 100% VT crushed all of them, albeit only in terms of performance, not sharpe. The longer the time horizon, the more "100% equities" dominates everything else. For shorter horizons or leveraged portfolios, the sharpe matters more. For non-leveraged-constrained portfolios, equities + leveraged ITT (intermediate term treasuries) outperforms other combinations including 80/20 on risk-adjusted basis. There is an mHFEA thread in this forum.
Is it really true that stocks have less risk the longer the time horizon?

Longer time horizon -> more opportunities for a black swan to come along and wipe the slate clean.
That's perfectly debatable, and people smarter than myself have debated it before, and there have been threads in this forum. I would think if your investment horizon is a few decades, and if and when there is a black swan event in your lifetime that you would not recover from during those few decades, that could only happen if that event triggered a systemic change, like the October revolution in Russia. In which case it doesn't matter if you held stocks, bonds, real estate, or anything else. Equities survived the world wars better than cash or bonds. Barring systemic changes, the exponential function embedded in the equity risk premium eventually outpaces valuation changes and other (mostly mean-reverting) forces that control equity pricing, with a likelihood approaching 1 (on a scale from 0 to 1) as your time horizon increases. Basic high school math 101. One of the few free lunches in investing. Simulations show that the range of terminal outcomes ("dispersion") gets wider with increasing time; but that range will almost entirely sit above the range of outcomes of lower return assets, as time increases.
But all that I think is a bit off-topic.
000
Posts: 8211
Joined: Thu Jul 23, 2020 12:04 am

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

Post by 000 »

comeinvest wrote: Thu Apr 21, 2022 10:45 pm That's perfectly debatable, and people smarter than myself have debated it before, and there have been threads in this forum. I would think if your investment horizon is a few decades, and if and when there is a black swan event in your lifetime that you would not recover from during those few decades, that could only happen if that event triggered a systemic change, like the October revolution in Russia. In which case it doesn't matter if you held stocks, bonds, real estate, or anything else. Equities survived the world wars better than cash or bonds. Barring systemic changes, the exponential function embedded in the equity risk premium eventually outpaces valuation changes and other (mostly mean-reverting) forces that control equity pricing, with a likelihood approaching 1 as your time horizon increases. Basic high school math 101. One of the few free lunches in investing. Simulations show that the range of outcomes ("dispersion") gets wider with increasing time; but that range will almost entirely sit above the range of outcomes of lower yielding assets, as time increases.
But all that I think is a bit off-topic.
I think it is on topic, because if stocks have as much or even *more* risk as time increases, that would be a supporting reason for long bonds.

I don't think all black swans have to be like that. In fact, Taleb suggested a portfolio predominantly invested in Treasuries. Many have gotten this idea into their heads, which I call Selective Doomism, that it is not worth worrying about the risk in stocks because "oh well everything else would just crash too". I believe that is incorrect. All that is required is for the economic regime to be less rewarding of risk taking, as it was for most temporal and geographic periods, for long term stock risk to manifest.

Anyway, while researching exactly what happened to Tsarist bonds after the Russian revolution I found this:
After the collapse of the Soviet Union, the newly formed Russian Federation had to not only come up with a new financial strategy for its future, but also had to consider repaying the billions of dollars the Soviet Union borrowed from abroad. In 1996, Paris and Moscow signed an accord for Russia to repay a nominal value of between $80 and $100 for each of the 4 million czarist bonds believed to remain in circulation in France, for a total payout of around $400 million.[6] Russia paid but not nearly as generously as the descendants of French bond buyers hoped.[7]
That's what I call staying the course! :!:
comeinvest
Posts: 2709
Joined: Mon Mar 12, 2012 6:57 pm

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

Post by comeinvest »

000 wrote: Thu Apr 21, 2022 11:00 pm
comeinvest wrote: Thu Apr 21, 2022 10:45 pm That's perfectly debatable, and people smarter than myself have debated it before, and there have been threads in this forum. I would think if your investment horizon is a few decades, and if and when there is a black swan event in your lifetime that you would not recover from during those few decades, that could only happen if that event triggered a systemic change, like the October revolution in Russia. In which case it doesn't matter if you held stocks, bonds, real estate, or anything else. Equities survived the world wars better than cash or bonds. Barring systemic changes, the exponential function embedded in the equity risk premium eventually outpaces valuation changes and other (mostly mean-reverting) forces that control equity pricing, with a likelihood approaching 1 as your time horizon increases. Basic high school math 101. One of the few free lunches in investing. Simulations show that the range of outcomes ("dispersion") gets wider with increasing time; but that range will almost entirely sit above the range of outcomes of lower yielding assets, as time increases.
But all that I think is a bit off-topic.
I think it is on topic, because if stocks have as much or even *more* risk as time increases, that would be a supporting reason for long bonds.

I don't think all black swans have to be like that. In fact, Taleb suggested a portfolio predominantly invested in Treasuries. Many have gotten this idea into their heads, which I call Selective Doomism, that it is not worth worrying about the risk in stocks because "oh well everything else would just crash too". I believe that is incorrect. All that is required is for the economic regime to be less rewarding of risk taking, as it was for most temporal and geographic periods, for long term stock risk to manifest.

Anyway, while researching exactly what happened to Tsarist bonds after the Russian revolution I found this:
After the collapse of the Soviet Union, the newly formed Russian Federation had to not only come up with a new financial strategy for its future, but also had to consider repaying the billions of dollars the Soviet Union borrowed from abroad. In 1996, Paris and Moscow signed an accord for Russia to repay a nominal value of between $80 and $100 for each of the 4 million czarist bonds believed to remain in circulation in France, for a total payout of around $400 million.[6] Russia paid but not nearly as generously as the descendants of French bond buyers hoped.[7]
That's what I call staying the course! :!:
Hmm. So nearly a century later, somebody got some fraction back for the bonds of their ancestors ;) Lol
But I think both theoretical and empirical evidence for both "regular" and black swan scenarios like world wars, supports what I said before. Bonds basically add a layer of risk: inflation, that may not be naturally mean reverting: One period of hyperinflation will wipe you out. By contrast, the effect of the higher volatility of equities from valuation changes will diminish with increasing time horizon. You are rewarded for the equity risk with the equity risk premium, but the associated risk diminishes with time. That is the free lunch and the luxury of having a long investment horizon.

You are correct that it is very conceivable that investment returns and risk premia become smaller, or that risk-free investment returns even disappear or become negative in absolute terms. Expected investment returns are governed by supply and demand for "storage" of capital, i.e. delayed consumption by savers vs. immediate consumption by borrowers. But it would appear that the investment returns of various asset classes tend to move in tandem, like expected return of equities becomes smaller, all the while expected (real) returns of treasuries become negative. That's why equities for example get a boost when interest rates drop.

EDIT: Thinking of it, there is also a debate around the justification for the equity risk premium ("ERP puzzle"). You could argue that longer duration assets are safer, not more risky, than shorter duration assets, depending on the duration of investors' liabilities that they try to match. You could also argue that real assets are safer than bonds or cash and therefore don't deserve a premium (true for long horizons). But I personally would rely on the empirical evidence of a positive ERP that has persisted for generations and centuries I believe, rather than betting on the opposite to happen during my limited lifetime, based on exotic theories. The average investor seems to be averse to short-term volatility. I think currently the ERP can be estimated at around 4% p.a. (real) globally. A lot would have to change to make that negative all of a sudden; and even then, I guess at least in part it would be accomplished with astronomically surging valuations and asset prices, to precede decades of negative ERP, i.e. current investors would not really suffer. Think of the Japanese stock market of the 1980ies before it skyrocketed.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Thu Apr 21, 2022 9:24 pm When bond yields go up, all else equal, the expected future returns of both stocks and bonds go up.
As always, one should be cautious about statements like this to make sure one is keeping track of nominal versus real rates.

For example, this would be false:

"When [nominal bond] yields go up, all else equal [including the real yield curve remaining unchanged], the expected future [real] returns of . . . [nominal] bonds go up."

This of course is definitional--if the real yield curve has not changed, then the expected future real returns of nominal bonds has not changed.

But it is rather important to know this can happen! Among other things, it explains why nominal bonds can experience permanent real losses, and have done so in the past.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 5:24 am
But it is rather important to know this can happen! Among other things, it explains why nominal bonds can experience permanent real losses, and have done so in the past.
Note my emphasis on the word “both”. Unexpectedly high inflation erodes the value of all nominal assets.

Including stocks, which gave also suffered real losses.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Fri Apr 22, 2022 5:39 am
NiceUnparticularMan wrote: Fri Apr 22, 2022 5:24 am
But it is rather important to know this can happen! Among other things, it explains why nominal bonds can experience permanent real losses, and have done so in the past.
Note my emphasis on the word “both”. Unexpectedly high inflation erodes the value of all nominal assets.
So my point was your statement was potentially misleading about bonds to the extent it could fairly be taken as saying higher nominal rates meant higher real returns on nominal bonds going forward. It is very important not to mislead people about that issue, and therefore I think it is very important to be clear when making such statements.
Including stocks, which gave also suffered real losses.
Stocks are not bonds and are not what I would call "nominal assets". Stocks are ownership shares in companies. Companies produce returns for their owners through organizing productive activities using various assets which generate profits. Stock shares are then priced based on applying risk- and time-related discount rates to their expected future profits, all in light of relevant capital supply dynamics.

So as always, what is happening with stocks when nominal interest rates increase depends on why that is happening. The same with what is happening with stocks when inflation is unexpectedly high. Both of those conditions can be associated with a variety of macroeconomic conditions, and a variety of changes as to how profitable those companies are expected to be, how risky those expectations have become, discount rates, and so on.

So, it is true that the last time we saw a prolonged period of unexpectedly high USD inflation, which became a very bad period for nominal USD bonds, this was also a bad period for US stocks in terms of real returns. But that is because that specific period was associated with bad things for US stocks.

And for sure, the same thing could happen again. Indeed, it just DID happen again. But whether that continues going forward, we don't know.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 7:43 am So my point was your statement was potentially misleading about bonds to the extent it could fairly be taken as saying higher nominal rates meant higher real returns on nominal bonds going forward.
I think that my statement could only be taken that way if someone was intentionally trying to misinterpret it.

Why would anyone in good faith assume I was using nominal terms in half of the sentence and real terms in the other half?

NiceUnparticularMan wrote: Fri Apr 22, 2022 7:43 am And for sure, the same thing could happen again. Indeed, it just DID happen again. But whether that continues going forward, we don't know.
The future is always uncertain. That's what makes investing interesting. Unexpected returns happen all the time, but my statement specifically and intentionally referred to expected returns

Every asset pricing model I've seen models expected stock returns as some sort of equity risk premium(s) plus a risk-free rate. Empirically, the best fit for that risk-free rate is a nominal US Treasury note. And there are very sound theoretical reasons that this should be the case.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Fri Apr 22, 2022 7:58 am Why would anyone in good faith assume I was using nominal terms in half of the sentence and real terms in the other half?
For various reasons, including because you actually did change your terminology. Again, here is what you said, with emphasis of the key terms added:
When bond yields go up, all else equal, the expected future returns of . . . bonds go up.
Conventionally, most personal investors talk about bond YIELDS in nominal terms, and yet talk about future bond RETURNS in real terms.

This makes sense, because the former are transparent and reported widely, and the latter are what most personal investors care about.

And so while I doubt you intended to actually mislead anyone, the way you phrased this very much ran a large risk of misleading people who think about both yield and returns in conventional ways.
Unexpected returns happen all the time, but my statement specifically and intentionally referred to expected returns
Right, and it is inaccurate as to expected real returns. Stock and nominal bond expected real returns can change differently depending on the reasons for a nominal yield increase.
Every asset pricing model I've seen models expected stock returns as some sort of equity risk premium(s) plus a risk-free rate.
I agree that is a good working model.
Empirically, the best fit for that risk-free rate is a nominal US Treasury note. And there are very sound theoretical reasons that this should be the case.
I am aware of no such theoretical reasons.

One risk is default risk, and U.S. Treasuries have extremely little default risk. So that makes sense so far.

But most stock investors are interested in real returns, so it would make no sense for the risk-free rate to be given by a nominal bond subject to inflation risk.

Further, standard models of equity risk premiums tend to assume short-term volatility is at least one of those risks. So, it would make no sense for the risk-free rate to be given by longer-term bonds.

So, short-term TIPS are the natural choice for as close to a risk-free asset as exists (although even those are subject to some liquidity risk).

The appropriate risk-free rate is therefore the expected rate on rolling short-term TIPS.

As a final thought, all this is fine so far, except there is no good reason to expect the equity risk premium to be fixed. So, even as the real rate on short-term TIPS is changing, that does not mean the expected real return on stocks is changing in lockstep. That all depends on what is happening with the equity risk premium as well, and indeed is quite likely the dominant factor.

As one should expect. Stocks would not be so volatile if the equity risk premium was fixed.
muffins14
Posts: 5528
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

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

Post by muffins14 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 8:22 am
Conventionally, most personal investors talk about bond YIELDS in nominal terms, and yet talk about future bond RETURNS in real terms.
For what it’s worth, I always assume people mean nominal. And certainly why would you assume he meant one thing in half the sentence and another in the other half?

When people talk about expected returns in a mathematical sense, it is almost universally nominal returns, like in an asset pricing model.

I do not think anyone aside from yourself is taking such a big issue from this
Crom laughs at your Four Winds
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 8:22 am Conventionally, most personal investors talk about bond YIELDS in nominal terms, and yet talk about future bond RETURNS in real terms.
If you think "most personal investors" talk about returns in real terms you're meeting very different investors than I am.
NiceUnparticularMan wrote: Fri Apr 22, 2022 8:22 am Stock and nominal bond expected real returns can change differently depending on the reasons for a nominal yield increase.
This would be a reasonable comment if I were talking about nominal yields but real returns. I already pointed out that I was doing no such thing, and no reasonable person would think I was.

You've posted nearly 2,000 words attempting to clarify a single sentence I wrote that no one (including you) seems to have actually misunderstood. Can we drop this and stick to the actual topic of the thread?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

muffins14 wrote: Fri Apr 22, 2022 8:32 amWhen people talk about expected returns in a mathematical sense, it is almost universally nominal returns
I think the exact opposite!

In a context like retirement savings, I think when people are talking about future rates of return, withdrawal rates, and so on, they are almost talking about them in real terms.

Or at least should be. Inflation has been consistently low enough in recent years that perhaps some people have forgotten that real rates can be quite different from nominal rates. And so maybe to some, it is coming as an unpleasant surprise that they can depart so much.

But I certainly hope that most people thinking about these issues have been thinking in real terms.
I do not think anyone aside from yourself is taking such a big issue from this
What harm would be caused if people were always careful to be clear?
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Fri Apr 22, 2022 8:45 am If you think "most personal investors" talk about returns in real terms you're meeting very different investors than I am.
Maybe we need a survey.
Can we drop this and stick to the actual topic of the thread?
Again, I don't see the harm in being careful to be clear. But nor am I looking for any sort of "victory". I just really do wish people would be more careful when speaking about this issue.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 9:46 am Again, I don't see the harm in being careful to be clear. But nor am I looking for any sort of "victory". I just really do wish people would be more careful when speaking about this issue.
And I wish people wouldn't sidetrack the thread with false complaints about lack of clarity.

But it's the internet, so we don't always get what we want.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
BigJohn
Posts: 2627
Joined: Wed Apr 02, 2014 11:27 pm

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

Post by BigJohn »

NiceUnparticularMan wrote: Fri Apr 22, 2022 9:41 am
muffins14 wrote: Fri Apr 22, 2022 8:32 amWhen people talk about expected returns in a mathematical sense, it is almost universally nominal returns
I think the exact opposite!

In a context like retirement savings, I think when people are talking about future rates of return, withdrawal rates, and so on, they are almost talking about them in real terms.

Or at least should be. Inflation has been consistently low enough in recent years that perhaps some people have forgotten that real rates can be quite different from nominal rates. And so maybe to some, it is coming as an unpleasant surprise that they can depart so much.

But I certainly hope that most people thinking about these issues have been thinking in real terms.
I do not think anyone aside from yourself is taking such a big issue from this
What harm would be caused if people were always careful to be clear?
What “should be” isn’t alway the case. I agree with vineviz on this. While I always work in real returns, in my experience most people think in nominal terms. A lot of very smart people I know don’t even know the definition of the term real return. Once explained, they still struggle to change they way they think, it’s just too abstract a concept. Even on this forum I see it over and over again in discussions of TIPS were many people focus on the nominal return numbers.
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

BigJohn wrote: Fri Apr 22, 2022 12:45 pm What “should be” isn’t alway the case. I agree with vineviz on this. While I always work in real returns, in my experience most people think in nominal terms. A lot of very smart people I know don’t even know the definition of the term real return. Once explained, they still struggle to change they way they think, it’s just too abstract a concept. Even on this forum I see it over and over again in discussions of TIPS were many people focus on the nominal return numbers.
Assuming this is true, then I would view it as an overwhelming imperative to try to educate as many personal investors as possible about real terms, how they are different from nominal terms, and why for most investors it is real terms they should care about.

Like, I think in another thread someone was literally arguing you cannot lose money if you hold a long-term nominal bond to maturity (absent default, presumably). If this person does understand the difference between nominal and real rates, and understands then how little this may mean to a personal investor for practical purposes, then that may technically be true as they are defining "lose money." But then they should know what they said just doesn't mean much as an observation. Indeed, it is just literally defining what it means for a nominal bond not to default, and nothing more that would actually be of use to retirement savers planning how to invest.

But if they actually don't understand that all they are doing is reiterating the definition of not defaulting, if they actually think they are saying something of practical use to investors--I think that would indicate they don't really understand the distinction, and they are in serious danger of planning mistakes.

So if a lot of that is going on, I would hope that people who understood these terms would not just feed this problem by saying things which are technically true but also much more trivial than they might sound to people who did not understand these terms so well. I hope instead they would make sure such people understood the important practical concepts and issues.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 1:09 pm I hope instead ….
Can we PLEASE stick to the topic of this thread?
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
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 »

I just want to chime in to say this thread has been very helpful to myself and others and do not want to see it continue to be derailed.
"Contentment", the only thing you ever truly need more of!
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Fri Apr 22, 2022 1:12 pm
NiceUnparticularMan wrote: Fri Apr 22, 2022 1:09 pm I hope instead ….
Can we PLEASE stick to the topic of this thread?
I think I am.

This is the core argument from the OP:
However, it seems to me that precious few of these same Bogleheads are allocating their fixed income allocations in a manner congruent with modern financial knowledge. This is especially true for young accumulators, who seem just as prone as retirees to rely on milquetoast short- and intermediate-term bond funds when they should almost certainly be favoring long-term bonds instead.

Long-term bond funds not only offer superior diversification, but they also minimize interest rate risk for investors with long-term investment horizons (including virtually all investors who are currently accumulating retirement savings).
Starting at the end, "virtually all investors who are currently accumulating retirement savings" should be thinking in real terms.

And once you start thinking in real terms, long-term NOMINAL bonds do not minimize real return risk for such investors, which is a very important form of risk.

But long-term IPS do not necessarily offer superior diversification to a stock-heavy portfolio.

So it is not true such investors "should almost certainly be favoring" long-term NOMINAL bonds, nor almost certainly favoring long-term IPS.

And I think that is where modern financial knowledge, including understanding the distinction between real returns and nominal returns, and long-term nominal bonds and long-term IPS, and how they interact with different stock scenarios, actually leads.

Of course some might disagree. But to have a reasonable and informative discussion about those issues, we all need to make sure we are being clear about what we are talking about, and why.
User avatar
LadyGeek
Site Admin
Posts: 95691
Joined: Sat Dec 20, 2008 4:34 pm
Location: Philadelphia
Contact:

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

Post by LadyGeek »

Let's stay focused on the OP's proposal, which is:
vineviz wrote: Wed Aug 07, 2019 7:24 pm ...Leaving aside any question of what the stock/bond allocation should be for an investor, I propose a simple rule of thumb that gets the bond allocation in the right ballpark. This rule is designed specifically for retirement portfolios that are in their accumulation phase (i.e. pre-retirement), and while it may not be strictly optimal in every regard it is easy to remember and to codify into an IPS:

The first 20% of a portfolio allocated to bonds should be allocated to long-term US Treasury bonds.

If your portfolio is 90% stocks, then the other 10% should be long-term Treasuries.

If your portfolio is 80% stocks, then the other 20% should be long-term Treasuries.

If your portfolio is 70% stocks, then 20% should be long-term Treasuries and 10% in intermediate bonds (e.g. total bond market or intermediate-term Treasuries).

If your portfolio is 60% stocks, then 20% should be long-term Treasuries and 20% in intermediate bonds (e.g. total bond market or intermediate-term Treasuries).

And so on.
=========================================================
NiceUnparticularMan wrote: Fri Apr 22, 2022 1:23 pm Starting at the end, "virtually all investors who are currently accumulating retirement savings" should be thinking in real terms.

And once you start thinking in real terms, long-term NOMINAL bonds do not minimize real return risk for such investors, which is a very important form of risk.

But long-term IPS do not necessarily offer superior diversification to a stock-heavy portfolio.

So it is not true such investors "should almost certainly be favoring" long-term NOMINAL bonds, nor almost certainly favoring long-term IPS.

And I think that is where modern financial knowledge, including understanding the distinction between real returns and nominal returns, and long-term nominal bonds and long-term IPS, and how they interact with different stock scenarios, actually leads.

Of course some might disagree. But to have a reasonable and informative discussion about those issues, we all need to make sure we are being clear about what we are talking about, and why.
Since investors seem to misunderstand the terminology, may I suggest you start a new thread explaining what, exactly, nominal and real-return means?

My not-so hidden intent is that the wiki has nothing which discusses this important concept. Perhaps we can create a new wiki page on this topic?

Adding to the confusion is that we have a wiki page for: Nominal bond, which mentions real-return bonds are I-Bonds and TIPS (inflation linked).

If you're interested, start a new thread and post the link here.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 1:23 pm
Starting at the end, "virtually all investors who are currently accumulating retirement savings" should be thinking in real terms.

And once you start thinking in real terms, long-term NOMINAL bonds do not minimize real return risk for such investors, which is a very important form of risk.
If the question is whether nominal bonds immunize the investor against inflation risk, the answer is pretty clearly "no". If you think that's the right answer then we agree, and can move on. Right?

A typical 80/20 investor who is "currently accumulating retirement savings" is 15+ years from retirement, however, and has a considerable amount of built-in inflation protection in both their human capital and their equity holdings. For such investors the amount of residual inflation risk is generally manageable without the need for inflation-indexed bonds in their retirement portfolio, particularly if their non-retirement portfolio (e.g. emergency fund) contains some inflation-indexed assets like TIPS or Series I savings bonds. In fact, the amount of inflation protection they have is SO great that they probably benefit from the additional diversification power of long-term nominal Treasuries rather long-term TIPS.

But either form of long-term bonds (TIPS or nominal) serve the same function of minimizing interest rate risk, which is distinct from inflation risk as I have repeatedly reminded folks.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

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

Post by NiceUnparticularMan »

vineviz wrote: Fri Apr 22, 2022 1:55 pm A typical 80/20 investor who is "currently accumulating retirement savings" is 15+ years from retirement, however, and has a considerable amount of built-in inflation protection in both their human capital and their equity holdings. For such investors the amount of residual inflation risk is generally manageable without the need for inflation-indexed bonds in their retirement portfolio, particularly if their non-retirement portfolio (e.g. emergency fund) contains some inflation-indexed assets like TIPS or Series I savings bonds.
I actually think the typical investor who is 15+ years from retirement and also has a healthy emergency plan (I prefer that term over emergency fund) probably has no need for additional fixed-income at all. Indeed, they might well be best advised to be mildly leveraged.

So if they do want to add fixed income anyone, I would ask why. And then I could recommend a fixed-income strategy to accomplish that goal, or maybe try to persuade them they don't need one.
they probably benefit from the additional diversification power of long-term nominal Treasuries rather long-term TIPS.
Diversification logic doesn't apply to all risky assets. Meaning the efficient frontier can include none of certain risky assets. Indeed, almost all personal investors are excluding MANY types of risky assets.

Whether to include a given risky asset in your best guess as to an efficient portfolio typically requires some complex forecasting involving both expected returns and expected correlations. I am skeptical that there are any decent forward-looking models that make a strong case for including long-term nominal Treasuries at the current low real expected returns. But I would be willing to consider them.

Still, I think a good default position is diversified stocks is maybe enough in terms of risky assets for most personal investors far from retirement, with a reasonably high burden of proof on anyone who wants to argue adding some other type of risky asset is likely to make that risky portfolio more efficient going forward.
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 »

NiceUnparticularMan wrote: Fri Apr 22, 2022 2:06 pm
I actually think the typical investor who is 15+ years from retirement and also has a healthy emergency plan (I prefer that term over emergency fund) probably has no need for additional fixed-income at all. Indeed, they might well be best advised to be mildly leveraged.
Again, completely off-topic.

The suggestions was that most investors WHO CHOOSE TO OWN BONDS in their retirement portfolio would benefit from allocating to long-term Treasury bonds for the first 20% of bonds they own.

If you want to talk people out of buying bonds at all, then perhaps you can find another thread in which to do that.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
anoop
Posts: 3930
Joined: Tue Mar 04, 2014 12:33 am

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

Post by anoop »

Lacy Hunt who advises WHOSX and is typically a big long term bond bull ended his most recent newsletter with this.
"These and many other harbingers of
recession constitute a favorable environment for
long-term bond investors. However, should the
Federal Reserve cease in their efforts to calm
inflation before it has been fully restrained, bond
investors should be wary."
https://hoisington.com/pdf/HIM2022Q1NP.pdf

There are many that think the fed has already lost control of inflation.
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

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

Post by secondopinion »

anoop wrote: Fri Apr 22, 2022 2:16 pm Lacy Hunt who advises WHOSX and is typically a big long term bond bull ended his most recent newsletter with this.
"These and many other harbingers of
recession constitute a favorable environment for
long-term bond investors. However, should the
Federal Reserve cease in their efforts to calm
inflation before it has been fully restrained, bond
investors should be wary."
https://hoisington.com/pdf/HIM2022Q1NP.pdf

There are many that think the fed has already lost control of inflation.
Without diving into politics, there are ways they can control it; they have not lost control yet.

"Wary" is different from "do not do it". As long as people are buying long-term bonds for the right reasons, then this should not matter.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
Locked