So... the best predictor of future bond returns?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

So... the best predictor of future bond returns?

Post by HomerJ »

I've read (even Bogle said it) that the best predictor of future bond returns is actually today's current yield.

Well, 10 months ago it was like 1.5%, and now it's like 3.5%...

So, was the 10 month ago number wrong or is today's number wrong?

I actually have a spreadsheet where I pull money from my investments for the next 30 years, and things work even with the 1.5% dividends. But most of the money is gone in 30 years.

If I can plug 3.5% into there now, I can last 30 years and still have my original portfolio intact.

So which number is the best predictor for future bond returns?
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
dcabler
Posts: 4543
Joined: Wed Feb 19, 2014 10:30 am
Location: TX

Re: So... the best predictor of future bond returns?

Post by dcabler »

They're both right because you also need to pay attention to the duration and not just the yield.

The current yield is reasonably accurate only over the duration of a bond fund. If you have a bond fund with, say, a 7 year duration and the current yield is 2%, then you'll likely get something pretty close to 2% CAGR at the end of 7 years. But the trajectory of the dollar value of the fund can be just about anything between now and 7 years from now.

Withdrawing along the way, though is a different story. Duration matching with 2 (or more) bond funds and using the PMT function in excel will get you a pretty smooth withdrawal along the way.

Cheers.
Last edited by dcabler on Wed Sep 21, 2022 5:49 pm, edited 1 time in total.
jebmke
Posts: 25475
Joined: Thu Apr 05, 2007 2:44 pm
Location: Delmarva Peninsula

Re: So... the best predictor of future bond returns?

Post by jebmke »

Almost all forecasts are wrong. I’d still use the YTM of bonds, knowing it is likely wrong but better than anything else. But I don’t use predictions so take that in account. Usually I just pick a range of values that bracket a reasonable estimate. I never use anything but round numbers - no digits to the right of the decimal point.

I use Tips yields to approximate real rates and then round up or down. These days it seems to be 1-2% real for longer durations.
Don't trust me, look it up. https://www.irs.gov/forms-instructions-and-publications
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

Re: So... the best predictor of future bond returns?

Post by Marseille07 »

HomerJ wrote: Wed Sep 21, 2022 5:32 pm Well, 10 months ago it was like 1.5%, and now it's like 3.5%...

So, was the 10 month ago number wrong or is today's number wrong?
Both are right.

10 months ago your NAV was a lot higher, from which you could expect to make 1.5% CAGR. Today, your NAV is much lower, but you can expect to make 3.5% CAGR.

This is the "self-correcting" mechanism you talked about elsewhere.
Last edited by Marseille07 on Wed Sep 21, 2022 6:13 pm, edited 5 times in total.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: So... the best predictor of future bond returns?

Post by dbr »

It is possible that the best predictor is still a really lousy predictor.

A retirement spending model that casts the future as a wild range of probabilities imagined to be somewhere in the spread of all outcomes observed over the last century might give a better idea of how things might go. That is the methodology of the typical SWR study, FireCalc model, etc. If a person does not like that kind of modeling then a Monte Carlo model might be a better choice assuming you know how to be astute about the input distributions. If one likes neither, the problem of estimating can be dodged by not owning variable assets. The alternative is things like Social Security, pensions, fixed annuities, long term bond ladders, and so on. It may not be easy to realize those things in inflation indexed instruments. There is an approach to dodging future uncertainty by way of duration matching.
Last edited by dbr on Wed Sep 21, 2022 5:51 pm, edited 1 time in total.
rockstar
Posts: 6326
Joined: Mon Feb 03, 2020 5:51 pm

Re: So... the best predictor of future bond returns?

Post by rockstar »

Whatever yield you buy your bonds at is what you should expect to get back in return.
User avatar
Lee_WSP
Posts: 10401
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: So... the best predictor of future bond returns?

Post by Lee_WSP »

The full quote from Bogle as taken from his books is something similar to the following:

The expected return of a bond fund if held for the duration of the fund is the current yield.

So if you had a 1 year bond fund 10 months ago, it should have yielded you close to 1.5% as you approach 12 months if you add back in dividends. Likewise, you should be able to expect 3.5% from a 10 year bond fund if you hold it for ten years.

It makes sense to me and I'm okay with believing it.
dcabler
Posts: 4543
Joined: Wed Feb 19, 2014 10:30 am
Location: TX

Re: So... the best predictor of future bond returns?

Post by dcabler »

rockstar wrote: Wed Sep 21, 2022 5:50 pm Whatever yield you buy your bonds at is what you should expect to get back in return.
Over the life of the bond, yep. Bonds returns are about as deterministic as it gets if held to maturity.

Cheers.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: So... the best predictor of future bond returns?

Post by dbr »

rockstar wrote: Wed Sep 21, 2022 5:50 pm Whatever yield you buy your bonds at is what you should expect to get back in return.
Yes, that is certainly true.

But it doesn't help for a future in which one has not yet bought most of the bonds involved. It works fine if you can buy all the bonds now.

I take the OP to mean that he is looking at a plan for a retirement that starts some years from now and will last 30 years. It is hard to buy bonds for 30-50 years out.

There is also the problem that what you address assumes waiting to maturity to get back the return. That works in a long future only if you hold a ladder of maturities spanning the entire future. The problem is that to obtain income you have to match the maturities/durations to the timing of the withdrawals.

At this point assuming a person has therefore solved the predictability problem he is still at the mercy of history regarding what yields he might have managed to buy depending on when he can buy and how much money he has to invest at that time. Even buying a ladder or an SPIA has very different outcomes depending on when you can buy it. Once bought, of course, one knows what one has.
Tamalak
Posts: 1989
Joined: Fri May 06, 2016 2:29 pm

Re: So... the best predictor of future bond returns?

Post by Tamalak »

Neither is wrong. Investing in bonds then gave you an expected 1.5% return. Investing in them now gives an expected 3.5%. The price drop in bonds from then to now bridges the gap between those expectations.
km91
Posts: 1388
Joined: Wed Oct 13, 2021 12:32 pm

Re: So... the best predictor of future bond returns?

Post by km91 »

The best predictor of a bond's return is it's starting yield. We know what the final value of the bond will be so given the the starting price and coupon structure we can be reasonably certain in the path an individual bond will take. What we don't know with certainty, and I think this is what you are alluding to, is forward expected starting yields. The term structure of interest rates tells us what the market expects forward rates to be. For example, given the 1yr Treasury yield and 2yr Treasury yield we can back into the market implied 1y 1y forward yield that satisfies the no arbitrage condition, but the research tells us that market implied forward yields are not very good predictors of forward realized rates
Logan Roy
Posts: 1838
Joined: Sun May 29, 2022 10:15 am

Re: So... the best predictor of future bond returns?

Post by Logan Roy »

I think investing in the form of a bond ladder is by far the most reliable way to know and control where you'll be. You know exactly what a 10 yr Treasury will return in nominal terms, and exactly what 10yr TIPS will return in real terms.

As inflation's the big unknown, it's going to express itself in the form of how much you have to withdraw to meet expenses. For this reason, one might consider TIPS the only return you can really forecast in a portfolio. And I think debt's a very big issue – as the more of it we have, the less we can afford to have positive real yields on government bonds. So by laddering, you can at least decide whether buying bonds (TIPS) at today's prices makes sense.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: So... the best predictor of future bond returns?

Post by skierincolorado »

If you are investing with a 30 year horizon, but the bonds that you own are less than 30 years in duration, then yes your outlook improves when real yields increase (real yields have gone from -1 to +1.3%).

For a more accurate picture be sure to factor in the implied inflation rate in current yields. The 10 year breakeven inflation rate between nominal and tips is 2.5%.
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

I guess my point was not clear...

Yes, if you buy a 10-year bond, you know exactly what the return will be in 10 years.

I was thinking more along the lines of the "2% is the new 4%" crowd like Wade Pfau (not trying to single him out, just an example) who have written papers stating that today's low bond yields predict low bond yields forever...
But the biggest driver for what I'm talking about right now is the low interest-rate environment. Low bond yields mean low bond returns in the future. And there's not really any controversy about that. It's a very close mathematical relationship. If interest rates don't change, today's bond yields will be the bond returns. And then, of course, if you're holding bond mutual funds, well, if interest rates go up, you're going to have capital losses, which make things even worse. Or vice versa, if interest rates decrease further, you could have capital gains. But effectively, future bond returns are going to be very close to today's bond yields. And that means spending from bonds is going to be lower mathematically.
The problem with the above statement is that interest rates have changed. I might have planned around 1.5% just 10 months ago, but should I now be planning around 3.5% instead?

I don't understand the comment "Low bond yields now means low bond returns in the future".

Interest rates do change, especially over 30 years...

I never understood the idea that "4% is more likely to fail" just because THIS YEAR, with 30 years of changes ahead of us, bond yields are low.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
rossington
Posts: 1823
Joined: Fri Jun 07, 2019 2:00 am
Location: Florida

Re: So... the best predictor of future bond returns?

Post by rossington »

HomerJ wrote: Sat Sep 24, 2022 1:25 pm

Low bond yields mean low bond returns in the future.
The way I read it is the above statement is explained by this:
And there's not really any controversy about that. It's a very close mathematical relationship. If interest rates don't change, today's bond yields will be the bond returns
.

----------------------------------------
The problem with the above statement is that interest rates have changed. I might have planned around 1.5% just 10 months ago, but should I now be planning around 3.5% instead?

Interest rates do change, especially over 30 years...

I never understood the idea that "4% is more likely to fail" just because THIS YEAR, with 30 years of changes ahead of us, bond yields are low.

If you're holding bond funds since they are constantly rolling the bonds (and if you are still investing in them or reinvesting dividends) I would use the SEC yield to gauge what current return at best may be. Other than that a fund is going to be difficult to predict vs. individual bonds over the time frame since most of the bonds in the fund won't be held to maturity and your basis is constantly changing year to year.

I never lend any credence to the percentage predictions, everyone is different and what works best for one person does not work as well for another.
"Success is going from failure to failure without loss of enthusiasm." Winston Churchill.
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

rossington wrote: Sun Sep 25, 2022 1:28 am
HomerJ wrote: Sat Sep 24, 2022 1:25 pm

Low bond yields mean low bond returns in the future.
The way I read it is the above statement is explained by this:
And there's not really any controversy about that. It's a very close mathematical relationship. If interest rates don't change, today's bond yields will be the bond returns
.
Yes, I guess it's explained by that.

If NOTHING CHANGES, then everything will remain the same.

But interest rates DO change, especially over the entire course of a 30-year retirement... even in just 10 months like we've just seen.

So it's a ridiculously STUPID phrase for a PhD economist to make, and for many here to think helps when they try to calculate expected returns.

That's my point.

Why do people say low bond yields today indicate low bond yields in the FUTURE?

It's only true if interest rates don't change. But interest rates do change, so the statement is not true.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

Re: So... the best predictor of future bond returns?

Post by Marseille07 »

HomerJ wrote: Sun Sep 25, 2022 1:41 pm Yes, I guess it's explained by that.

If NOTHING CHANGES, then everything will remain the same.

But interest rates DO change, especially over the entire course of a 30-year retirement... even in just 10 months like we've just seen.

So it's a ridiculously STUPID phrase for a PhD economist to make, and for many here to think helps when they try to calculate expected returns.

That's my point.

Why do people say low bond yields today indicate low bond yields in the FUTURE?

It's only true if interest rates don't change. But interest rates do change, so the statement is not true.
If you had 1M of bond ETF 10 months ago, your expected return calculation would be 1M * 1.5%.

Today, bond NAV dropped and you have 900K. Your expected return calculation would be 900K * 3.5%.

It doesn't seem stupid to me.
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: So... the best predictor of future bond returns?

Post by SimpleGift »

HomerJ wrote: Sun Sep 25, 2022 1:41 pm Why do people say low bond yields today indicate low bond yields in the FUTURE?
My sense is that people say today's starting bond yields generally forecast future bond returns for the approximate duration of the bond or the bond fund. For example, since the 1970s the history of the U.S. Aggregate Bond Index — with an approximate duration of 5 years — seems to bear this out (chart below, R^2 = 0.88):
Another way of looking at this same data:
Note that this relationship is not automatic, and exceptions can be offered. For example, periods of aggressive rate hikes, like in the 1970s, confound the relationship. And the relationship only appears to hold for the approximate duration of the bond or bond fund — i.e., not 30 years in the future for the starting yield of a bond fund with only a 5-year duration.
Last edited by SimpleGift on Sun Sep 25, 2022 3:35 pm, edited 1 time in total.
CletusCaddy
Posts: 2678
Joined: Sun Sep 12, 2021 4:23 am

Re: So... the best predictor of future bond returns?

Post by CletusCaddy »

SPLB corporate long term bond ETF is yielding a juicy 5.5% right now. Very tempting
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

SimpleGift wrote: Sun Sep 25, 2022 3:31 pm
HomerJ wrote: Sun Sep 25, 2022 1:41 pm Why do people say low bond yields today indicate low bond yields in the FUTURE?
My sense is that people say today's starting bond yields generally forecast future bond returns for the approximate duration of the bond or the bond fund. For example, since the 1970s the history of the U.S. Aggregate Bond Index — with an approximate duration of 5 years — seems to bear this out (chart below, R^2 = 0.88):
That's fair... But they say that when doing 30-year retirement SWR numbers, not just talking about 6 years from an intermediate bond fund.

"Interest rates are real low today so you can't count on 4% SWR for retirement"

Except that interest rates may not be low in 5 years (or in 10 months), so maybe making proclamations about 4% failing over 30 years due to low interest rates TODAY is incorrect.

That's my point.
Note that this relationship is not automatic, and exceptions can be offered. For example, periods of aggressive rate hikes, like in the 1970s, confound the relationship. And the relationship only appears to hold for the approximate duration of the bond or bond fund — i.e., not 30 years in the future for the starting yield of a bond fund with only a 5-year duration.
Exactly.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

HomerJ wrote: Sun Sep 25, 2022 1:41 pm Why do people say low bond yields today indicate low bond yields in the FUTURE?

It's only true if interest rates don't change. But interest rates do change, so the statement is not true.
The statement is true even though yields change because those changes in yields are not predictable.

Current yield curves reflect the market's aggregate expectations about what future economic conditions will be. Pretty much by definition, they are as likely to be better than expected as they are to be worse than expected.

So low bond yields translate into high expected returns, and vice versa. Surprises are certain, but their direction and magnitude are not.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Beensabu
Posts: 5657
Joined: Sun Aug 14, 2016 3:22 pm

Re: So... the best predictor of future bond returns?

Post by Beensabu »

HomerJ wrote: Wed Sep 21, 2022 5:32 pm So which number is the best predictor for future bond returns?
3.5% of current value for the duration of what is held.
"The only thing that makes life possible is permanent, intolerable uncertainty; not knowing what comes next." ~Ursula LeGuin
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: So... the best predictor of future bond returns?

Post by abc132 »

I think the bond return calculations are fairly accurate because all your old bonds are still earning those old rates. It's only a big change if you are going from lots of stocks and buying into a bunch of bonds at one point in time.

I think the SWR calculations can change too quickly to be useful. Not many of us are out there spending 1.5 times what we did 2 years ago because of updated SWR calculations. The predictors tend to linearize the present to a degree that makes their predictions less accurate.

That being said, it's tough to predict even the mean returns for the next X years so finding the 90% or 95% case from that mean is never going to be accurate. There are some tricks like curve fitting annual returns, but what we care about is annual returns to the 20th or 30th power. 3% annual error over 20 years becomes something more like (1.03)^20 which gives you a net worth estimate that is almost totally useless. The average annual estimate is even worse when you include withdrawals and deviation.
Last edited by abc132 on Sun Sep 25, 2022 4:10 pm, edited 1 time in total.
User avatar
TheTimeLord
Posts: 12130
Joined: Fri Jul 26, 2013 2:05 pm

Re: So... the best predictor of future bond returns?

Post by TheTimeLord »

HomerJ wrote: Wed Sep 21, 2022 5:32 pm I've read (even Bogle said it) that the best predictor of future bond returns is actually today's current yield.

Well, 10 months ago it was like 1.5%, and now it's like 3.5%...

So, was the 10 month ago number wrong or is today's number wrong?

I actually have a spreadsheet where I pull money from my investments for the next 30 years, and things work even with the 1.5% dividends. But most of the money is gone in 30 years.

If I can plug 3.5% into there now, I can last 30 years and still have my original portfolio intact.


So which number is the best predictor for future bond returns?
Are those number real returns, because if they are nominal you would have to know the future inflation rate to know how long your money would last wouldn't you?
IMHO, Investing should be about living the life you want, not avoiding the life you fear. | Run, You Clever Boy! [9085]
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

HomerJ wrote: Sun Sep 25, 2022 3:50 pm
That's fair... But they say that when doing 30-year retirement SWR numbers, not just talking about 6 years from an intermediate bond fund.
The same nature of relationship holds when using longer-term bonds: the best predictor of bond returns over the next 30 years would be the yield on a bond with a 15 year duration , which is pretty close to the duration on a 20-year bond.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Beensabu
Posts: 5657
Joined: Sun Aug 14, 2016 3:22 pm

Re: So... the best predictor of future bond returns?

Post by Beensabu »

HomerJ wrote: Sun Sep 25, 2022 3:50 pm But they say that when doing 30-year retirement SWR numbers, not just talking about 6 years from an intermediate bond fund.

"Interest rates are real low today so you can't count on 4% SWR for retirement"
The 4% SWR came from the Trinity study which used long-term investment grade corporate bonds. So perhaps they were talking about 30 years from a long-term investment grade bond fund with whatever yields were at that time.

Edit: And Bengen's original study used 5-year treasuries. So perhaps they were talking about how 5-year treasury yields were currently much lower than what was used there.
"The only thing that makes life possible is permanent, intolerable uncertainty; not knowing what comes next." ~Ursula LeGuin
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: So... the best predictor of future bond returns?

Post by SimpleGift »

vineviz wrote: Sun Sep 25, 2022 4:12 pm
HomerJ wrote: Sun Sep 25, 2022 3:50 pm That's fair... But they say that when doing 30-year retirement SWR numbers, not just talking about 6 years from an intermediate bond fund.
The same nature of relationship holds when using longer-term bonds: the best predictor of bond returns over the next 30 years would be the yield on a bond with a 15 year duration , which is pretty close to the duration on a 20-year bond.
Right. The chart below shows starting monthly, long-term government bonds yields and their subsequent 20-year real returns for the 64-year period from January 1926 to September 1990. The basic relationship between starting yields and eventual returns for long-term bonds still appears to hold — however with a bit wider dispersion:
Note there are historical periods of rapidly changing inflation and bond yields when the forecast relationship breaks down.
Last edited by SimpleGift on Sun Sep 25, 2022 5:25 pm, edited 2 times in total.
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

vineviz wrote: Sun Sep 25, 2022 4:12 pm
HomerJ wrote: Sun Sep 25, 2022 3:50 pm
That's fair... But they say that when doing 30-year retirement SWR numbers, not just talking about 6 years from an intermediate bond fund.
The same nature of relationship holds when using longer-term bonds: the best predictor of bond returns over the next 30 years would be the yield on a bond with a 15 year duration , which is pretty close to the duration on a 20-year bond.
Ugh... so 1 year ago, the yield was 1.5% on a 10-year note, and today it's nearly 4%... so was the best long-term 30-year prediction made by you and other economists 1 year ago wrong? or it today's 4% best long-term 30-year prediction wrong?

Because they can't both be right.

I assume you are using the phrase "best predictor" in very subjective manner.

Just like if I have a saw and screwdriver to drive in a nail, the screwdriver (handle) is the "best tool" for the job, but it's still a pretty terrible tool.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

abc132 wrote: Sun Sep 25, 2022 4:08 pm I think the SWR calculations can change too quickly to be useful.
SWR-based withdrawal estimates are a lot less volatile than you might think, in large part because of the largely self-correcting relationship between prices and expected returns that we are discussing here.

Withdrawal rates tend to go up when yields go up, but increases in yield are mechanistically linked to decreasing prices. A 20% drop in portfolio value is not troubling if the SWR has increased from 2% to 2.5% because of the corresponding increase in expected return.

In other words, the amount of the sustainable withdrawal changes much less quickly or vigorously than the rate of sustainable withdrawal does.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
finite_difference
Posts: 3633
Joined: Thu Jul 09, 2015 7:00 pm

Re: So... the best predictor of future bond returns?

Post by finite_difference »

I would assume 0% real for bonds over the long term, like 10+ years.

In Monte Carlo simulations, maybe go with a range of values but such that the long term average is 0% real.

You are using Monte Carlo simulations, right?
The most precious gift we can offer anyone is our attention. - Thich Nhat Hanh
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: So... the best predictor of future bond returns?

Post by abc132 »

vineviz wrote: Sun Sep 25, 2022 4:34 pm
abc132 wrote: Sun Sep 25, 2022 4:08 pm I think the SWR calculations can change too quickly to be useful.
SWR-based withdrawal estimates are a lot less volatile than you might think, in large part because of the largely self-correcting relationship between prices and expected returns that we are discussing here.
They do not, because I understand this and think the withdrawal rate changes too much.

If I ignored the prior calculation I would still be on track today.


vineviz wrote: Sun Sep 25, 2022 4:34 pm Withdrawal rates tend to go up when yields go up, but increases in yield are mechanistically linked to decreasing prices. A 20% drop in portfolio value is not troubling if the SWR has increased from 2% to 2.5% because of the corresponding increase in expected return.

In other words, the amount of the sustainable withdrawal changes much less quickly or vigorously than the rate of sustainable withdrawal does.
If you are suggesting we are at 2.5% today (we are not), that would make sense.

What's the increase if we are at 4.0% SWR today?

X*(2.2%) = 2.2% withdrawal
0.8*X*(4.0%) = 3.2% of the original amount

(3.2/2.2 - 1) = 45% increase in spending

Feel free to tell us what you think today's SWR is and we can compute the change in spending. I'll bet it's close to +50% in the 2.5 years time since the SWR = 2.2% prediction.
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

HomerJ wrote: Sun Sep 25, 2022 4:30 pm
Ugh... so 1 year ago, the yield was 1.5% on a 10-year note, and today it's nearly 4%... so was the best long-term 30-year prediction made by you and other economists 1 year ago wrong? or it today's 4% best long-term 30-year prediction wrong?

Because they can't both be right.
For starters, I will again encourage you distinguish an "estimate" from a "prediction". It will help you, I promise.

For another thing, the 10-year bond yield was never the "best long-term 30-year prediction" because 10 years and 30 years aren't the same thing.

Finally, they CAN "both be right" and probably are. The price of a 30 year zero coupon bond purchased last year has fallen by 25-30%, so obviously its yield today is higher. Held for 30 years, a person who bought that bond last year will indeed see a return of about 1.9% whereas someone buying that same bond today will see a return of 3.1% from here.
Last edited by vineviz on Sun Sep 25, 2022 4:52 pm, edited 1 time in total.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

abc132 wrote: Sun Sep 25, 2022 4:41 pm They do not, because I understand this and think the withdrawal rate changes too much.
I'm pretty confident you are wrong on both counts.

abc132 wrote: Sun Sep 25, 2022 4:41 pm
vineviz wrote: Sun Sep 25, 2022 4:34 pm Withdrawal rates tend to go up when yields go up, but increases in yield are mechanistically linked to decreasing prices. A 20% drop in portfolio value is not troubling if the SWR has increased from 2% to 2.5% because of the corresponding increase in expected return.

In other words, the amount of the sustainable withdrawal changes much less quickly or vigorously than the rate of sustainable withdrawal does.
If you are suggesting we are at 2.5% today (we are not), that would make sense.

What's the increase if we are at 4.0% SWR today?
Please just follow the math in my example. I made it as simple as I could.

2.5% of $80 = 2% of $100.

In both cases, the amount of the sustainable withdrawal is the same $2.00.
Last edited by vineviz on Sun Sep 25, 2022 4:51 pm, edited 1 time in total.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Marseille07
Posts: 16054
Joined: Fri Nov 06, 2020 12:41 pm

Re: So... the best predictor of future bond returns?

Post by Marseille07 »

vineviz wrote: Sun Sep 25, 2022 4:46 pm Finally, they CAN "both be right" and probably are. The price of a 30 year zero coupon bond purchased last year has fallen by 25-30%, so obviously its yield today is higher. Held for 30 years, a person who bought that bond last year will indeed see a return of about 1.9% whereas someone buying that same bond today will see a return of 3.1% from there.
Pretty much this. Not difficult.
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

vineviz wrote: Sun Sep 25, 2022 4:46 pm For another thing, the 10-year bond yield was never the "best long-term 30-year prediction" because 10 years and 30 years aren't the same thing.
So are you saying all SWR calculations (which implies 30+ years) are based on 30-year yields?

But what if I, like almost everyone else, use intermediate bond funds instead? Or, at least, a mixture?

I don't think anyone retires with JUST 30-year bonds in their portfolio.
Finally, they CAN "both be right" and probably are. The price of a 30 year zero coupon bond purchased last year has fallen by 25-30%, so obviously its yield today is higher. Held for 30 years, a person who bought that bond last year will indeed see a return of about 1.9% whereas someone buying that same bond today will see a return of 3.1% from here.
Sure, but that's not the question... I already agreed that if you buy a 30-year bond or a 10-year bond, you know what they are worth based on their starting yield.

But unless someone retires with JUST 30-year bonds (and almost no does), how do you extrapolate intermediate term bond yields across the full 30 years?

The REASON people were saying just last year (and for years before that) that 4% is not likely to work at this time in history is because bond yields were so low. But last year's bond yields were unlikely to stay that low for the entire 30 years.

And in fact, they didn't stay low.
Last edited by HomerJ on Sun Sep 25, 2022 5:08 pm, edited 1 time in total.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

Marseille07 wrote: Sun Sep 25, 2022 4:51 pm
vineviz wrote: Sun Sep 25, 2022 4:46 pm Finally, they CAN "both be right" and probably are. The price of a 30 year zero coupon bond purchased last year has fallen by 25-30%, so obviously its yield today is higher. Held for 30 years, a person who bought that bond last year will indeed see a return of about 1.9% whereas someone buying that same bond today will see a return of 3.1% from there.
Pretty much this. Not difficult.
Again, not the subject. I agree that a 30-year bond will pay out around it's starting yield, no matter when you cash it in.

But I don't own any 30-year bonds. But I will have bonds in my portfolio for 30+ years.

How do I determine what yield I am likely to get from intermediate and short-term bonds over the 30+ years?

If the answer is, "you can't", that's fine. Because things could be different in 10 years (or 10 months, as we've just seen).

That's what I expect. But I know that the people talking about 4% failing the last couple of years were saying that because they thought bond yields will likely be low for 30 years straight.

Which doesn't jive with "you can't"
Last edited by HomerJ on Sun Sep 25, 2022 5:07 pm, edited 1 time in total.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
Topic Author
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: So... the best predictor of future bond returns?

Post by HomerJ »

finite_difference wrote: Sun Sep 25, 2022 4:35 pm I would assume 0% real for bonds over the long term, like 10+ years.
Not a bad solution. :beer

Better than trying to estimate the next 30 years of intermediate and short-term yields based on just today's yield.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
Beensabu
Posts: 5657
Joined: Sun Aug 14, 2016 3:22 pm

Re: So... the best predictor of future bond returns?

Post by Beensabu »

HomerJ wrote: Sun Sep 25, 2022 5:05 pm How do I determine what yield I am likely to get from intermediate and short-term bonds over the 30+ years?
You can only know what yield you will (most likely) get over the durations of the bonds you hold.
"The only thing that makes life possible is permanent, intolerable uncertainty; not knowing what comes next." ~Ursula LeGuin
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: So... the best predictor of future bond returns?

Post by abc132 »

vineviz wrote: Sun Sep 25, 2022 4:49 pm
Please just follow the math in my example. I made it as simple as I could.

2.5% of $80 = 2% of $100.

In both cases, the amount of the sustainable withdrawal is the same $2.00.
I followed your post, Please try reading mine.

When SWR prediction goes from 2.2% to 4.0% with a 20% drop, you will be spending 45% more per year.

Yes 2.2% to 2.5% with a 20% drop results in no change.

Please try an example that matches what has actually happened with SWR predictions.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: So... the best predictor of future bond returns?

Post by Kevin M »

HomerJ wrote: Wed Sep 21, 2022 5:32 pm I've read (even Bogle said it) that the best predictor of future bond returns is actually today's current yield.

Well, 10 months ago it was like 1.5%, and now it's like 3.5%...

So, was the 10 month ago number wrong or is today's number wrong?

I actually have a spreadsheet where I pull money from my investments for the next 30 years, and things work even with the 1.5% dividends. But most of the money is gone in 30 years.

If I can plug 3.5% into there now, I can last 30 years and still have my original portfolio intact.

So which number is the best predictor for future bond returns?
If you bought a zero-coupon 10-year bond 10 months ago, you would earn the initial yield if held to maturity. If you buy one now, same thing. Funds are less predictable, as are rolling bond ladders.
If I make a calculation error, #Cruncher probably will let me know.
Triple digit golfer
Posts: 10433
Joined: Mon May 18, 2009 5:57 pm

Re: So... the best predictor of future bond returns?

Post by Triple digit golfer »

Here's what I gather. You can safely predict the exact returns of a bond held to maturity. Everything else (not held to maturity or bond funds) is unpredictable because of interest rate or market fluctuations.

Is that right?
User avatar
Beensabu
Posts: 5657
Joined: Sun Aug 14, 2016 3:22 pm

Re: So... the best predictor of future bond returns?

Post by Beensabu »

abc132 wrote: Sun Sep 25, 2022 5:19 pm When SWR prediction goes from 2.2% to 4.0% with a 20% drop, you will be spending 45% more per year.
No, you won't.

The safe withdrawal rate is the percentage of your portfolio that you will withdraw the first year of withdrawal. You will then withdraw that same amount (except adjusted for inflation) each year thereafter. And doing this should last you 30 years.

You don't reset SWR each year that you are withdrawing. You adjust the initial amount that was withdrawn the first year based on inflation over time.
"The only thing that makes life possible is permanent, intolerable uncertainty; not knowing what comes next." ~Ursula LeGuin
Northern Flicker
Posts: 15363
Joined: Fri Apr 10, 2015 12:29 am

Re: So... the best predictor of future bond returns?

Post by Northern Flicker »

HomerJ wrote: Wed Sep 21, 2022 5:32 pm I've read (even Bogle said it) that the best predictor of future bond returns is actually today's current yield.

Well, 10 months ago it was like 1.5%, and now it's like 3.5%...

So, was the 10 month ago number wrong or is today's number wrong?
Neither. This is a predictor of future return over a time period. If a bond or bond fund had a negative return in year 1 and a 3.5% annual return for the next (N-1) years then it certainly could have an aggregate return of 1.5% over N years starting at the beginning of year 1.
User avatar
Kevin M
Posts: 15787
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: So... the best predictor of future bond returns?

Post by Kevin M »

Triple digit golfer wrote: Sun Sep 25, 2022 5:33 pm Here's what I gather. You can safely predict the exact returns of a bond held to maturity. Everything else (not held to maturity or bond funds) is unpredictable because of interest rate or market fluctuations.

Is that right?
Basically, yes. You know the exact internal rate of return (IRR) of a bond, assuming no default, as it is equal to the initial yield to maturity (or very close).

However, the investment return depends on the reinvestment rates of the coupon payments. A zero-coupon bond (bills or STRIPS) has no coupon payments, so the investment return is known in advance, as well as the IRR.

There are some defined maturity bond ETFs, but I haven't researched them at all.
If I make a calculation error, #Cruncher probably will let me know.
Triple digit golfer
Posts: 10433
Joined: Mon May 18, 2009 5:57 pm

Re: So... the best predictor of future bond returns?

Post by Triple digit golfer »

Kevin M wrote: Sun Sep 25, 2022 5:39 pm
Triple digit golfer wrote: Sun Sep 25, 2022 5:33 pm Here's what I gather. You can safely predict the exact returns of a bond held to maturity. Everything else (not held to maturity or bond funds) is unpredictable because of interest rate or market fluctuations.

Is that right?
Basically, yes. You know the exact internal rate of return (IRR) of a bond, assuming no default, as it is equal to the initial yield to maturity (or very close).

However, the investment return depends on the reinvestment rates of the coupon payments. A zero-coupon bond (bills or STRIPS) has no coupon payments, so the investment return is known in advance, as well as the IRR.

There are some defined maturity bond ETFs, but I haven't researched them at all.
Yes, makes sense. Thanks for confirming!
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: So... the best predictor of future bond returns?

Post by abc132 »

Beensabu wrote: Sun Sep 25, 2022 5:34 pm
abc132 wrote: Sun Sep 25, 2022 5:19 pm When SWR prediction goes from 2.2% to 4.0% with a 20% drop, you will be spending 45% more per year.
No, you won't.

The safe withdrawal rate is the percentage of your portfolio that you will withdraw the first year of withdrawal. You will then withdraw that same amount (except adjusted for inflation) each year thereafter. And doing this should last you 30 years.

You don't reset SWR each year that you are withdrawing. You adjust the initial amount that was withdrawn the first year based on inflation over time.
Technically, yes, but that means no SWR calculation is ever relevant after you start withdrawing, and no new information post-retirement ever matters.

The new calculation is saying you can "safely" spend 45% more each year for the next 27.5 years.

If you believed the old calculation, why would you not believe the new calculation?
Last edited by abc132 on Sun Sep 25, 2022 5:43 pm, edited 1 time in total.
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: So... the best predictor of future bond returns?

Post by vineviz »

abc132 wrote: Sun Sep 25, 2022 5:19 pm When SWR prediction goes from 2.2% to 4.0% with a 20% drop, you will be spending 45% more per year.
A 20% drop in portfolio is not sufficient to move the SWR from 2.2% to 4.0%. You need a much bigger drop in price or a smaller increase in SWR to get the numbers to match.

You're assuming your conclusion, in other words.

abc132 wrote: Sun Sep 25, 2022 5:19 pm Please try an example that matches what has actually happened with SWR predictions.
From August 2021 to August 2022 the SWR (assuming 90% probability of success) for a 60/40 portfolio increased roughly 19%, from about 2.6% to about 3.1%. Over that same time the value of a 60/40 portfolio dropped by 14%.

That's a change in expected withdrawal, expressed in dollars, of only about 2-3%.

As I said, month to month volatility is just not that great.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
abc132
Posts: 2435
Joined: Thu Oct 18, 2018 1:11 am

Re: So... the best predictor of future bond returns?

Post by abc132 »

vineviz wrote: Sun Sep 25, 2022 5:43 pm
abc132 wrote: Sun Sep 25, 2022 5:19 pm When SWR prediction goes from 2.2% to 4.0% with a 20% drop, you will be spending 45% more per year.
A 20% drop in portfolio is not sufficient to move the SWR from 2.2% to 4.0%. You need a much bigger drop in price or a smaller increase in SWR to get the numbers to match.

You're assuming your conclusion, in other words.

abc132 wrote: Sun Sep 25, 2022 5:19 pm Please try an example that matches what has actually happened with SWR predictions.
From August 2021 to August 2022 the SWR (assuming 90% probability of success) for a 60/40 portfolio increased roughly 19%, from about 2.6% to about 3.1%. Over that same time a the value of a 60/40 portfolio dropped by 14%.

That's a change in expected withdrawal, expressed in dollars, of only about 2-3%.

As I said, month to month volatility is just not that great.
That's cute you found a period of time it didn't change much.

Now try Spring 2020 (~2.2% SWR) to today and tell me what you think the percent change is.

I concede that it sometimes doesn't change much and can self correct - that was never even a point of discussion on my end.

I'm letting you use your own calculation to prove your point.

You just have to be able to respond to my posts - that "safe" spending changed a lot (too much) in the last 2.5 years.

The target date is the date from the Wade Pfau 2.2% prediction/publication (Spring 2020) to today.


I also may need to check your 3.1% number today - that seems quite low given we can guarantee 4.0% SWR (100% success) with TIPS.

Can you share the expected stock and bond returns and AA you used to get that number?

I believe Wade has a site where we can also compare his SWR number today (a good apples-to-apples comparison).
Last edited by abc132 on Sun Sep 25, 2022 6:01 pm, edited 5 times in total.
alluringreality
Posts: 1511
Joined: Tue Nov 12, 2019 9:59 am

Re: So... the best predictor of future bond returns?

Post by alluringreality »

Crestmont Research suggests: "...interest rates somewhere along the yield curve tend to move more than 50 basis points during the subsequent six month period." They indicate the statement for Friday values has been true: "Since 1968... more than 97% of the weeks", although they suggest that zero interest rate policy reduced the frequency to: "85%". One item they call attention to with their presentation is how in a lower interest rate environment a tendency towards moves of more than 50 basis points over 6 months represents a larger percentage of the initial rate than in a higher rate environment.
https://www.crestmontresearch.com/docs/ ... hanges.pdf
45% US Indexes, 25% Ex-US Indexes, 30% Fixed Income - Buy & Hold
Parkinglotracer
Posts: 3947
Joined: Fri Dec 20, 2019 2:49 am
Location: Upstate NY

Re: So... the best predictor of future bond returns?

Post by Parkinglotracer »

We all know that quote about predictions and forecasts … they are usually good as long as it isn’t about the future.
Post Reply