A bond duration glide path for retirement investing

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

sls239 wrote: Wed Jun 24, 2020 4:28 pm Well doesn't credit risk have a duration component? I mean isn't it inherently more difficult to evaluate credit risk on a longer duration bond?
Credit risk does have a term structure, but that term structure is considered to be independent the US Treasury yield curve.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Triple digit golfer
Posts: 10430
Joined: Mon May 18, 2009 5:57 pm

Re: A bond duration glide path for retirement investing

Post by Triple digit golfer »

I haven't read the entire thread and always enjoy the perspective of vineviz, but the more I learn, accumulate, and experience market swings, the more I think I may forever be happy with an 80/20 allocation.

While I'm youngish (35) it keeps a nice bond cushion in case of long term job loss, etc. Simply put, it helps me sleep at night.

As I age, the portfolio gets larger, as does the risk of job loss, long term unemployment, or forced early retirement. If those things happen, 20% of a larger portfolio is a larger dollar amount. The bond portion grows in dollars along with the risk of bad things happening.

In retirement, assuming 25+ years of expenses, I will have 5+ years of bonds going in. That seems like enough for me to be comfortable.

At this point, I see myself sleeping well at night at 80/20 in all scenarios. I've seen posts where people recommend 75/25 in retirement. Close enough. I would be comfortable with that in all scenarios as well.

All of this assumes no dedicated emergency fund outside of the AA.
pascalwager
Posts: 2312
Joined: Mon Oct 31, 2011 8:36 pm

Re: A bond duration glide path for retirement investing

Post by pascalwager »

vineviz wrote: Wed Jun 24, 2020 7:54 am
pascalwager wrote: Wed Jun 24, 2020 7:01 am
Ben Mathew wrote: Tue Jun 23, 2020 10:39 pm
pascalwager wrote: Tue Jun 23, 2020 7:47 pm
Swivelguy wrote: Tue Jun 23, 2020 7:19 pm

If you will be spending your portfolio at a roughly constant rate over the next 10 years, the average duration to the expenditure is 5 years.
You may not make it to 87, or you may live longer, so a ten-year duration makes more sense to me as an average.
You need to consume 1 year from now, 2 years from now, and so on. So you would need 1 year bonds, 2 year bonds, etc.. That brings the average duration across all your bonds closer to the middle of expected years left rather than towards the end of expected years left.
Okay, thanks, probably makes sense. I'm only one duration-year above the curve for my age, but previously thought I was three years below. Also, I'm atypical in my use of the portfolio and heir situation. So, I've got some rethinking to do.
I wouldn't obsess over small differences, honestly.

For one thing, I applied some smoothing to the graph I posted and was also using a mortality table slightly different than the one used by SSA.gov.

More importantly, unless you've got strong longevity insurance (i.e. a deferred income annuity) in place you're almost certainly going to want to plan for a retirement longer than your remaining life expectancy. The challenge is that your calculated (estimated, really) remaining investment time horizon becomes increasingly dependent on a subjective evaluation (i.e. what probability of outliving your wealth are you comfortable with) as you move into your 80s and early 90s.

Personally, I'm with you in your original evaluation: something close to a 8 to 10-year duration is about what I'd say is prudent for a 77-year old investor, because that investor has a non-trivial (i.e. 10% or better) chance of living to at least age 95. In other words, "remaining life expectancy" is a pretty good ballpark estimate of "remaining investment horizon".
OK, good. Right now my SS life expectancy is ten (10) years and my bonds duration is seven (7) years. I'd like to raise the duration a bit, but my available account funds don't permit that without reducing my inflation protection. So, I'm inclined to let my life expectancy diminish until it's down to seven (7) years while maintaining my bond duration at seven (7) years for this period. After that, I'll just continuously reduce the bonds duration to match reduced life expectancy.

Note: I don't have SS and it's inflation protection, so I take my 50% TIPS allocation seriously.

Thanks for your input!
VT 60% / VFSUX 20% / TIPS 20%
Angst
Posts: 2968
Joined: Sat Jun 09, 2007 11:31 am

Re: A bond duration glide path for retirement investing

Post by Angst »

vineviz wrote: Wed Jun 24, 2020 2:37 pm Someone suggested that it might be helpful to show how an investor might use the duration glide path with a little more specificity, so I constructed this example glide path using an assortment of bond funds and a generic stock allocation.

Image

Don't fixate on the particular bond funds I chose, or even the number of funds. The duration glide path can easily be managed using just two bond funds at a time: the only effort required is to periodically adjust the ratio of the longer-term fund and the shorter-term fund(s) to keep the weighted average duration close your target.
Thank you for the additional graphic - I think this visualization helps some people conceptualize what you're talking about. From my own perspective, the portfolio (LMP) of LT TIPS I've been buying for about a decade now at auction could be thought of as somewhat akin to the green area of your graph, although my generic graph's green portion would elongate or stretch out more horizontally to span from 40 years to 100 years when the final 30 yr TIPS rung bought at age 70 matures.
Robot Monster
Posts: 4215
Joined: Sun May 05, 2019 11:23 am

Re: A bond duration glide path for retirement investing

Post by Robot Monster »

Okay, to recap some of what I've learned from vineviz. Investors who keep their bond duration matched to their time horizon have eliminated their interest rate risk. So, for example, if you have $10,000 that you don't plan on touching for 6.5 years, you could buy Intermediate-Term Bond (VBILX) which has an average duration of 6.5 years, therefore matching duration to time horizon. (Obviously, if your time horizon doesn't exactly match the duration of a single bond fund, two bond funds can be utilized.)

Now to play this out with an example.

1. Jerry, purchases $10,000 of VBILX, having a 6.5 year time horizon on that money. At the time he purchases the fund, the fund's yield is 1%.
2. Interest rates shoot up 2% overnight. Jerry's VBILX shares are now worth $8,700.

The fund's yield is now 3%. Let's say it stays stable for the next 6.5 years.

3. After the first year has ended, Jerry's shares are worth $8,961 (8700*1.03).
4. After the second year, his shares are worth $9,229, year three at $9,506, year four at $9,791, year five at $10,084, year six at $10,386, and six months later $10,541.

Am I getting this right so far? Not asking you to check my math, just the general thought process.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Robot Monster wrote: Tue Jul 21, 2020 11:13 am Okay, to recap some of what I've learned from vineviz. Investors who keep their bond duration matched to their time horizon have eliminated their interest rate risk. So, for example, if you have $10,000 that you don't plan on touching for 6.5 years, you could buy Intermediate-Term Bond (VBILX) which has an average duration of 6.5 years, therefore matching duration to time horizon. (Obviously, if your time horizon doesn't exactly match the duration of a single bond fund, two bond funds can be utilized.)
. . .
Am I getting this right so far? Not asking you to check my math, just the general thought process.
I think the only thing you might be missing is that the investment horizon (the number to which you're matching the bond duration) typically isn't static.

To make it more clear I think I'd restate your premise slightly: " ... you have $10,000 that you plan on spending in exactly 6.5 years."

So today (7/21/20202) your investment horizon is 6.5 years. In 12 months (i.e. 7/21/2021) your investment horizon is 5.5 years. Etc.

The reason you need two funds is that you need to maintain the equality between bond duration and investment horizon if you want to totally eliminate interest rate risk.

So while you might want 100% in VBILX today, one year from now you want 85% in VBILX and 15% in cash (which has a duration of 0). The reason is that [(.85 x 6.5) + (.15 x 0.0) = 5.5] and your investment horizon one year from now will be 5.5 years.

This is the concept of a glide path: you'll "glide" from 100% VBILX today to 85% in 1 year, 69% in 2 years, ...., 8% in six years, and finally 0% VBILX in 6.5 years.

If you do that, you'll have $10,668 at the end no matter what happens to bond yields in the meantime. There are some nuances in the calculations but basically it's just the yield compounded over 6.5 years: (1.1^6.5) * $10,000 = $10,668.
"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: A bond duration glide path for retirement investing

Post by Robot Monster »

vineviz wrote: Tue Jul 21, 2020 12:26 pm To make it more clear I think I'd restate your premise slightly: " ... you have $10,000 that you plan on spending in exactly 6.5 years." So today (7/21/20202) your investment horizon is 6.5 years. In 12 months (i.e. 7/21/2021) your investment horizon is 5.5 years. Etc.
Oh, okay, I'll revise my example.

1. Jerry, purchases $10,000 of VBILX, having a 6.5 year time horizon on that money. At the time he purchases the fund, the fund's yield is 1%.
2. Interest rates shoot up 2% overnight. Jerry's VBILX shares are now worth $8,700.

The fund's yield is now 3%. Let's say it stays stable for the next 6.5 years. Let's say yield on money markets jumps to 1%.

Year 1 -- 3% turns $8700 into $8961
Year 2 -- 85% VBILX 15% cash yields 2.7% which turns $8961 into $9202
Year 3 -- 2.38% yields $9421
Year 4 -- 2.08% yields $9617
Year 5 -- 1.76% yields $9786
Year 6 -- 1.46% yields $9928
Year 6.5 -- 1.16% yields $10007

Wow, that actually worked out...hopefully I did the calculations correctly and didn't just land on that accidentally.
User avatar
Steve Reading
Posts: 2959
Joined: Fri Nov 16, 2018 9:20 pm

Re: A bond duration glide path for retirement investing

Post by Steve Reading »

Robot Monster wrote: Tue Jul 21, 2020 2:40 pm
vineviz wrote: Tue Jul 21, 2020 12:26 pm To make it more clear I think I'd restate your premise slightly: " ... you have $10,000 that you plan on spending in exactly 6.5 years." So today (7/21/20202) your investment horizon is 6.5 years. In 12 months (i.e. 7/21/2021) your investment horizon is 5.5 years. Etc.
Oh, okay, I'll revise my example.

1. Jerry, purchases $10,000 of VBILX, having a 6.5 year time horizon on that money. At the time he purchases the fund, the fund's yield is 1%.
2. Interest rates shoot up 2% overnight. Jerry's VBILX shares are now worth $8,700.

The fund's yield is now 3%. Let's say it stays stable for the next 6.5 years. Let's say yield on money markets jumps to 1%.

Year 1 -- 3% turns $8700 into $8961
Year 2 -- 85% VBILX 15% cash yields 2.7% which turns $8961 into $9202
Year 3 -- 2.38% yields $9421
Year 4 -- 2.08% yields $9617
Year 5 -- 1.76% yields $9786
Year 6 -- 1.46% yields $9928
Year 6.5 -- 1.16% yields $10007

Wow, that actually worked out...hopefully I did the calculations correctly and didn't just land on that accidentally.
Just fyi, matching duration to investment horizon technically works only if there is a parallel shift in interest rates. If the yield curve actually changes, things are different. For instance, if the yield of VBILX jumps to 3% but cash yields stay right where they are, then as you reinvest part of your portfolio onto cash, you don't actually make up the money.

Your numbers work because whatever money you lose on VBILX, you technically make back up with the higher yields over the following 6.5 years. But if those yields aren't actually any better (so the 6.5 year yield is the only one that moved up), then you just lose money and nothing else.

Of course, the yield curve might be as likely to flatten as it could get steeper so you still want to match duration to horizon as it's the best you could do. I'm just saying it won't work out perfectly. Vineviz will feel free to correct me if I'm wrong but I believe what I'm saying is correct.
"... so high a present discounted value of wealth, it is only prudent for him to put more into common stocks compared to his present tangible wealth, borrowing if necessary" - Paul Samuelson
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Steve Reading wrote: Tue Jul 21, 2020 2:49 pm Vineviz will feel free to correct me if I'm wrong but I believe what I'm saying is correct.
No, that's right: duration calculations include some embedded assumptions, which typically include parallel shifts in the yield curve as well as the ability to reinvest coupon payments at the bond's prevailing yield. Like many simplifying assumptions these make the problem computationally efficient at the expense of a little precision, getting you 95% of the way to the right answer with 5% of the effort it would take to get 100% of the way there.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
GoneOnTilt
Posts: 1841
Joined: Fri Sep 28, 2018 4:59 pm

Re: A bond duration glide path for retirement investing

Post by GoneOnTilt »

mhc wrote: Tue Jun 23, 2020 3:59 pm If a person only knew when he would die, planning would be so easy.
^This. Intermediate bonds and call it a day.
Wrench
Posts: 1046
Joined: Sun Apr 28, 2019 10:21 am

Re: A bond duration glide path for retirement investing

Post by Wrench »

Vineviz has done a superb analysis on using bond fund allocations to reduce interest rate risk in retirement. But, wouldn't it be easier just to buy a bond that produces the income desired on the date you want it? If I am 65 and want $10,000 in 10 years, buy a TIPS bond (or a corporate or treasury bond if you don't want inflation protection) that expires in 10 years and then I will have my (real if TIPS) $10,000 at that time. If you want it every year for a period of years, build a TIPS bond ladder. Done. No rebalancing required, no need to do anything except spend the money when each bond expires. Alternatively, buy iBonds, though of course you are limited in the amount to $10K per year. (Although not really - if you are married, it's $10K for you and your spouse and you set up a trust (or an LLC) and purchase another $10K/year, and if you have kids you trust give each of them $10K to buy one for you, etc.). None of this at all diminishes what Vineviz proposes - it is brilliant - but rather just another way to skin the cat! (De gustibus non est disputandum).
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Wrench wrote: Sun Jul 26, 2020 7:38 am Vineviz has done a superb analysis on using bond fund allocations to reduce interest rate risk in retirement. But, wouldn't it be easier just to buy a bond that produces the income desired on the date you want it?
For investors over the age and 55 or 60 this is almost certainly the case.

However, many investors start getting serious about retirement planning in their 40s or early 50s. The longest maturity Treasury bonds are currently 30 years, so these “transitional” investors will need either a DIY rolling ladder or a long-term bond fund.

Buying Treasuries (nominal or TIPS) at auction is cheap and easy, but I’m less enthusiastic about selling small lots of bonds into the secondary market when low cost bond index funds can do that well.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
17outs
Posts: 222
Joined: Thu Feb 13, 2020 3:03 pm

Re: A bond duration glide path for retirement investing

Post by 17outs »

I'm glad that this thread led to exactly what I am looking for. Is there a guideline for matching the duration with 2 funds? For example I understand the guideline is to use LTT and then slowly introduce inflation protected bonds as I get closer to retirement.

I am calculating duration for the investment horizon that starts in 10 years and lasts for 20. Age in 10 years (50) + Age in 30 years (70) /2 = 60. 60 - TODAYAGE = 20 avg bond duration

How does one take a duration and equate it to 2 funds? Is the guideline to start with LTT adding in TIPS to reach the duration? At which point does ratio in the 2 funds flip flop? Like others said it would be cool to have the glide path not by age and more towards years until retirement and years of retirement but I understand all are different.

Thanks for reading and allowing me to revive an old thread.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

17outs wrote: Mon Jan 04, 2021 12:39 pm How does one take a duration and equate it to 2 funds? Is the guideline to start with LTT adding in TIPS to reach the duration? At which point does ratio in the 2 funds flip flop? Like others said it would be cool to have the glide path not by age and more towards years until retirement and years of retirement but I understand all are different.

Yes, really all you need to know is the target duration and the duration of the two funds. I made a simple Google Sheets spreadsheet that calculates the allocation based on those three inputs: https://docs.google.com/spreadsheets/d/ ... sp=sharing

You have copy the spreadsheet or download it for Excel to change the inputs, but feel free to do that if you want to experiment.

The following formula will tell the what portion of the bonds should be in the long-duration fund:

Let TD = the target duration (e.g. 20 years);
Let DL = the duration of the longest duration fund;
Let DS = the duration of the shortest duration fund;

Code: Select all

(TD - DS) / (DL - DS)
So if the target duration is 20 years, the "long" fund has a duration of 24 years (for instance, Vanguard Extended Duration Treasury ETF), and the short fund has a duration of 8 years (e.g. an intermediate TIPS fund) then the formula is:

(20 - 8)/(24-8) = 12/16 = 75% in the longer fund, 25% in the shorter fund.

If your target duration changes to 18 years, just change the numbers: (18-8)/(24-8) = 10/16 = 63% in the longer fund, 37% in the shorter fund.

And so forth.

If you make a column in Excel with our target duration for each year, that formula will generate your "glide path" if you just copy it down a column next to it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
garlandwhizzer
Posts: 3562
Joined: Fri Aug 06, 2010 3:42 pm

Re: A bond duration glide path for retirement investing

Post by garlandwhizzer »

vineviz wrote:

Actively taking on interest rate risk (by choosing shorter duration bonds than your investment horizon,for example) is the speculative bet.

Just like the stock/bond glide path is a tool for avoiding market timing, so too is the bond duration glide path.
First of all no one knows beforehand with actionable accuracy what his investment horizon is going to be. We only guess. You can die much sooner than expected or much later than expected. Also you can encounter unanticipated emergencies and/or changes in your financial circumstances before you anticipated horizon including divorce, lawsuits, children in financial distress needing money, loss of job, disability, early dementia, long term care costs, rising inflation that increases nominal costs of living, etc.,). Any of these can dramatically increase your need for liquid stable nominal assets at an unexpected and unplanned for time.

Holding all LTB may look nice on backtesting charts and make a convincing academic argument but it's important to remember are not dots on a chart. We have variable, not fixed, financial needs and these can change quickly, unexpectedly, and dramatically. With all the uncertainty in our financial futures I personally believe there is a place for multiple bond portfolios: bond barbells (ST + LT), TBM that averages out to intermediate term or a single high quality intermediate term bond fund, all of which reduce principal volatility relative to LTB alone. Intermediate term takes a mid position between yield (higher with LTT) and stability of principal (higher with STB). The higher yield with LTB which comes at the price of greater principal volatility. In short, I don't think that it's a once size (LTT) fits all situation.

Another reason there is so much love now for LTT is that we're at the end of 39 massive bull market in bonds, especially LTT which have massively benefited from their enormous principal value appreciation as both rates and inflation have been declining from 1982 to now (LTT yields 15% to 1.66%). We've seen the optimal side of long duration during that bull market, but that train is not going to run the same way for the next 39 years. If instead of declining inflation and rates begin a steady climb higher, you'll see the ugly side of long duration as investors witnessed from 1940 -1980. Backtesting over the last 4 decades of LTB relative to STB or ITB is IMO a joke relative to future expectations. No one knows the future but one thing is clear. The bond market and the economy today is lot more like it was in 1940 than in 1980.

The following article from Inestopedia discusses interest rate risk. Vineviz has suggested that LTB reduces interest rate risk relative to shorter term bonds. Investopedia has the exact opposite opinion on this question.

https://www.investopedia.com/terms/i/in ... terisk.asp Quote from article:
Bond Price Sensitivity
The value of existing fixed-income securities with different maturity dates declines by varying degrees when market interest rates rise. This phenomenon is referred to as “price sensitivity” and is measured by the bond's duration.

For instance, suppose there are two fixed-income securities, one that matures in one year and another that matures in 10 years. When market interest rates rise, the owner of the one-year security can reinvest in a higher-rate security after hanging onto the bond with a lower return for only one year at most. But the owner of the 10-year security is stuck with a lower rate for nine more years.

That justifies a lower price value for the longer-term security. The longer a security's time to maturity, the more its price declines relative to a given increase in interest rates.

Note that this price sensitivity occurs at a decreasing rate. A 10-year bond is significantly more sensitive than a one-year bond but a 20-year bond is only slightly less sensitive than a 30-year one.

The Maturity Risk Premium
A long-term bond generally offers a maturity risk premium in the form of a higher built-in rate of return to compensate for the added risk of interest rate changes over time. The larger duration of longer-term securities means higher interest rate risk for those securities. To compensate investors for taking on more risk, the expected rates of return on longer-term securities are typically higher than rates on shorter-term securities. This is known as the maturity risk premium.
Garland Whizzer
17outs
Posts: 222
Joined: Thu Feb 13, 2020 3:03 pm

Re: A bond duration glide path for retirement investing

Post by 17outs »

Thanks! And the "long" fund will always be the treasuries or does it really matter as long as the duration is matched?
Horton
Posts: 1479
Joined: Mon Jan 21, 2008 2:53 pm
Contact:

Re: A bond duration glide path for retirement investing

Post by Horton »

The DFA Target Retirement Income glidepath is an alternative for this discussion - see below. They do not begin adding “Income Risk Management” (i.e., duration matched TIPS) until an individual is 20 years from retirement. Ostensibly, this is because an individual faces interest rate risk prior to this point that cannot be managed by the long TIPS. Another way to say it - an individual faces reinvestment risk prior to this point because there are no nominal or real government bonds with maturities more than 30 years out.

I also consider it instructive to think in terms of cash flows - both the cash flows of the underlying assets as well as cash flows of the liability (i.e., spending in retirement, which is subject to volatility). If an accumulator is 35 years old and planning to retire at age 65, there there is no asset that generates cash flows to match the expected retirement cash flows. So, why bother?

As an alternative, one could be agnostic to duration - use whatever you want - up until you get within 20ish years from retirement and then begin considering duration matched TIPS or nominal Treasuries. As a benchmark or rule of thumb, the DFA funds provide the percentages to allocate.

Image
80% global equities (faith-based tilt) + 20% TIPS (LDI)
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

17outs wrote: Mon Jan 04, 2021 4:03 pm Thanks! And the "long" fund will always be the treasuries or does it really matter as long as the duration is matched?
For purposes of duration, it doesn't really matter.

I'd normally recommend that the longer fund have as little credit risk as possible, which normally means Treasury bonds (either nominal or TIPS), but there's nothing ironclad about that preference.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

garlandwhizzer wrote: Mon Jan 04, 2021 3:57 pm First of all no one knows beforehand with actionable accuracy what his investment horizon is going to be.
I see claims like this a lot, but they just don't hold up.

There's no requirement that we know for certain the date of our retirement and the date of our death before we can begin to plan for retirement. I MIGHT die two years after I retire, but I'm not building my financial plan with that as the EXPECTED length of my retirement. In fact, a prudent investor will typically need to plan on a retirement that lasts beyond their life expectancy.

The investment horizon is a planning horizon: if you have a financial goal then you have an implicit planning horizon. Obviously that time horizon will have some uncertainty surround it, since all estimates naturally do, but people are forced to make decisions in the face of uncertainty all the time.

The problem is that people aren't naturally very good at it, which is why heuristics like the bond duration glide path can be helpful as planning tools.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Horton
Posts: 1479
Joined: Mon Jan 21, 2008 2:53 pm
Contact:

Re: A bond duration glide path for retirement investing

Post by Horton »

vineviz wrote: Mon Jan 04, 2021 7:00 pm
garlandwhizzer wrote: Mon Jan 04, 2021 3:57 pm First of all no one knows beforehand with actionable accuracy what his investment horizon is going to be.
I see claims like this a lot, but they just don't hold up.

There's no requirement that we know for certain the date of our retirement and the date of our death before we can begin to plan for retirement. I MIGHT die two years after I retire, but I'm not building my financial plan with that as the EXPECTED length of my retirement. In fact, a prudent investor will typically need to plan on a retirement that lasts beyond their life expectancy.

The investment horizon is a planning horizon: if you have a financial goal then you have an implicit planning horizon. Obviously that time horizon will have some uncertainty surround it, since all estimates naturally do, but people are forced to make decisions in the face of uncertainty all the time.

The problem is that people aren't naturally very good at it, which is why heuristics like the bond duration glide path can be helpful as planning tools.
The two of you may be able to find some common ground:

viewtopic.php?p=4529781#p4529781
80% global equities (faith-based tilt) + 20% TIPS (LDI)
Horton
Posts: 1479
Joined: Mon Jan 21, 2008 2:53 pm
Contact:

Re: A bond duration glide path for retirement investing

Post by Horton »

vineviz wrote: Tue Jun 23, 2020 7:28 pm
marcopolo wrote: Tue Jun 23, 2020 6:57 pm Makes sense.

Are you aware of any funds that do this rebalancing for you?

I am thinking something like a target date fund that shortens the maturity as it gets closer to your predicted date of demise?
The TDF series that get closest to the ideal setup IMHO are Dimensional Fund Advisors (DFA) and State Street (SSGA), but neither is widely available to retail investors.
Looks like State Street will be removing long bonds from their TDFs this April:
Effective April 1, 2021, State Street may adjust the mix of Bond Investments for investors prior to retirement by reducing the allocation to Long-term U.S. Treasury Bonds and introducing a new allocation to Intermediate U.S. Treasury Bonds. The overall allocation to U.S. Treasury Bonds is expected to remain the same.
Replacing a portion of the existing Long-
term U.S. Treasury Bond exposure with a new exposure to Intermediate U.S. Treasury Bonds potentially would improve return expectations for the Target Retirement Funds in the event of rising interest rates. Importantly, the overall exposure to U.S. Treasury Bonds is not expected to change as a result of this enhancement, so the Target Retirement Funds may still benefit from the historically strong diversification characteristics of U.S. Treasury Bonds.
https://www.ssga.com/library-content/pr ... l-fund.pdf
80% global equities (faith-based tilt) + 20% TIPS (LDI)
User avatar
Scott S
Posts: 1937
Joined: Mon Nov 24, 2008 2:28 am
Location: building my position

Re: A bond duration glide path for retirement investing

Post by Scott S »

Horton wrote: Tue Feb 09, 2021 7:00 pmLooks like State Street will be removing long bonds from their TDFs this April:
Effective April 1, 2021, State Street may adjust the mix of Bond Investments for investors prior to retirement by reducing the allocation to Long-term U.S. Treasury Bonds and introducing a new allocation to Intermediate U.S. Treasury Bonds. The overall allocation to U.S. Treasury Bonds is expected to remain the same.
Replacing a portion of the existing Long-
term U.S. Treasury Bond exposure with a new exposure to Intermediate U.S. Treasury Bonds potentially would improve return expectations for the Target Retirement Funds in the event of rising interest rates. Importantly, the overall exposure to U.S. Treasury Bonds is not expected to change as a result of this enhancement, so the Target Retirement Funds may still benefit from the historically strong diversification characteristics of U.S. Treasury Bonds.
https://www.ssga.com/library-content/pr ... l-fund.pdf
Suppose SSA is hedging their bets? Target retirement funds ought to take the long view, but they'll always be compared to similar funds in the short term. If interest rates do rise, then LTTs will hurt the SSA funds for now. But if they move to intermediate term bonds now, they won't do any worse than other funds that were already in ITTs. Right?
"Old value investors never die, they just get their fix from rebalancing." -- vineviz
guppyguy
Posts: 463
Joined: Tue Jan 30, 2018 3:24 pm

Re: A bond duration glide path for retirement investing

Post by guppyguy »

I find this thread very interesting.

46yo, approximately 80/20 split with bonds split between about 5% I-Bonds and the rest Intermediate Govt (VGIT).

How would I implement a bond duration glide path and what mix of nominal/inflation bonds should I achieve?

Just thinking out loud....
I-Bonds have a duration of zero so that would average any LTT duration down?
I originally was considering a 50/50 mix between short term TIP/IBond and VGIT. A bond allocation completely in LTT seems to be devoid of an unexpected inflation protection...am I wrong here?
hudson
Posts: 7098
Joined: Fri Apr 06, 2007 9:15 am

Re: A bond duration glide path for retirement investing FROM JUNE 24 2020

Post by hudson »

vineviz wrote: Wed Jun 24, 2020 7:54 am
pascalwager wrote: Wed Jun 24, 2020 7:01 am
Ben Mathew wrote: Tue Jun 23, 2020 10:39 pm
pascalwager wrote: Tue Jun 23, 2020 7:47 pm
Swivelguy wrote: Tue Jun 23, 2020 7:19 pm

If you will be spending your portfolio at a roughly constant rate over the next 10 years, the average duration to the expenditure is 5 years.
You may not make it to 87, or you may live longer, so a ten-year duration makes more sense to me as an average.
You need to consume 1 year from now, 2 years from now, and so on. So you would need 1 year bonds, 2 year bonds, etc.. That brings the average duration across all your bonds closer to the middle of expected years left rather than towards the end of expected years left.
Okay, thanks, probably makes sense. I'm only one duration-year above the curve for my age, but previously thought I was three years below. Also, I'm atypical in my use of the portfolio and heir situation. So, I've got some rethinking to do.
I wouldn't obsess over small differences, honestly.

For one thing, I applied some smoothing to the graph I posted and was also using a mortality table slightly different than the one used by SSA.gov.

More importantly, unless you've got strong longevity insurance (i.e. a deferred income annuity) in place you're almost certainly going to want to plan for a retirement longer than your remaining life expectancy. The challenge is that your calculated (estimated, really) remaining investment time horizon becomes increasingly dependent on a subjective evaluation (i.e. what probability of outliving your wealth are you comfortable with) as you move into your 80s and early 90s.

Personally, I'm with you in your original evaluation: something close to a 8 to 10-year duration is about what I'd say is prudent for a 77-year old investor, because that investor has a non-trivial (i.e. 10% or better) chance of living to at least age 95. In other words, "remaining life expectancy" is a pretty good ballpark estimate of "remaining investment horizon".
OLD DISCUSSION FROM JUNE 24, 2020...
Therefore, one could use the Social Security Actuarial Life Table (Life Expectancy)?

https://www.ssa.gov/oact/STATS/table4c6.html.

Age 77 Life Expectancy = 10.12 years.
Age 58: 23.31
Age 40: 38.75

Bottom Line: Life expectancy is used to figure average duration of bond holdings for the vineviz Bond Duration Glide Path for Retirement Investing.
Did I get that right?
withrye
Posts: 169
Joined: Mon Feb 29, 2016 12:48 pm

Re: A bond duration glide path for retirement investing

Post by withrye »

I just wanted to thank vineviz for making this thread. I had previously read their thread about "first 20% of bonds in LTT" and was wondering how to actually implement a duration glide path. I was on the cusp of posing the question in a new thread when I ran into this one. This thread, and in particular the linked sheet that does the simple algebra for you, is immensely useful as I mull the logistics of the approach further.

I will need to take a look at my pre-tax accounts to see what Treasury bond funds exist for mix-and-match purposes. I am in a high tax bracket and I expect that would be the simplest way to implement the approach given the accessibility of tax-free rebalancing in those accounts.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing FROM JUNE 24 2020

Post by vineviz »

hudson wrote: Fri Sep 17, 2021 7:31 am Therefore, one could use the Social Security Actuarial Life Table (Life Expectancy)?

https://www.ssa.gov/oact/STATS/table4c6.html.

Age 77 Life Expectancy = 10.12 years.
Age 58: 23.31
Age 40: 38.75

Bottom Line: Life expectancy is used to figure average duration of bond holdings for the vineviz Bond Duration Glide Path for Retirement Investing.
Did I get that right?
Yes, that's the right way to use that method. It provides a reasonably easy and accurate approximation.

The actuarial table used by the SSA may not accurately reflect YOUR longevity risk, so I typically suggest using a longevity estimation which takes advantage of some demographic and lifestyle variables.

E.g. https://www.blueprintincome.com/tools/l ... ll-i-live/ (note this site sells annuities, but the calculator itself is evidence-based)

or

https://www.longevityillustrator.org
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
hudson
Posts: 7098
Joined: Fri Apr 06, 2007 9:15 am

Re: A bond duration glide path for retirement investing FROM JUNE 24 2020

Post by hudson »

vineviz wrote: Wed Dec 08, 2021 7:22 am
hudson wrote: Fri Sep 17, 2021 7:31 am Therefore, one could use the Social Security Actuarial Life Table (Life Expectancy)?

https://www.ssa.gov/oact/STATS/table4c6.html.

Age 77 Life Expectancy = 10.12 years.
Age 58: 23.31
Age 40: 38.75

Bottom Line: Life expectancy is used to figure average duration of bond holdings for the vineviz Bond Duration Glide Path for Retirement Investing.
Did I get that right?
Yes, that's the right way to use that method. It provides a reasonably easy and accurate approximation.

The actuarial table used by the SSA may not accurately reflect YOUR longevity risk, so I typically suggest using a longevity estimation which takes advantage of some demographic and lifestyle variables.

E.g. https://www.blueprintincome.com/tools/l ... ll-i-live/ (note this site sells annuities, but the calculator itself is evidence-based)

or

https://www.longevityillustrator.org
Thanks again vineviz!
I liked the blueprintincome calculator best because it went into more detail and let me live longer. :)
balbrec2
Posts: 653
Joined: Mon Nov 13, 2017 2:03 pm

Re: A bond duration glide path for retirement investing

Post by balbrec2 »

So hypothetically and investor would trend from 80/20 stocks/long term bonds
to 20/80 stocks/short term bonds?
djm2001
Posts: 382
Joined: Fri May 29, 2020 6:23 am

Re: A bond duration glide path for retirement investing

Post by djm2001 »

balbrec2 wrote: Wed Dec 08, 2021 7:51 am So hypothetically and investor would trend from 80/20 stocks/long term bonds
to 20/80 stocks/short term bonds?
Raises an interesting question: Should long-term bond allocation go to zero before stock allocation goes to zero? From a purely term-risk perspective, this seems wrong. Are there other compensated risks that overpower term risk to make it preferable to hold stocks over long-term bonds even when approaching end of life?
AA = global stocks & bonds @ market weight (~60/40); EF = i-bonds; WR = -PMT(1%, 100-age, 1, 0, 1)
pascalwager
Posts: 2312
Joined: Mon Oct 31, 2011 8:36 pm

Re: A bond duration glide path for retirement investing

Post by pascalwager »

djm2001 wrote: Wed Dec 08, 2021 9:41 am
balbrec2 wrote: Wed Dec 08, 2021 7:51 am So hypothetically and investor would trend from 80/20 stocks/long term bonds
to 20/80 stocks/short term bonds?
Raises an interesting question: Should long-term bond allocation go to zero before stock allocation goes to zero? From a purely term-risk perspective, this seems wrong. Are there other compensated risks that overpower term risk to make it preferable to hold stocks over long-term bonds even when approaching end of life?
Overall fixed-income duration would not be long-term deep into retirement. The shortened investment horizon would not allow this: duration = calculated investment horizon.
User avatar
quisp65
Posts: 245
Joined: Wed Apr 10, 2019 7:44 am
Location: San Diego CA

Re: A bond duration glide path for retirement investing

Post by quisp65 »

I failed in thinking about this adequately. I balance one way into fixed income at market highs and let my approximate 8 years of fixed income run out in a long bear. I shouldn't be thinking of including intermediate bond funds given my short & shrinking fixed income horizon. Will let my 1.5 years of intermediates run their course and won't replace. Short term bonds & cash for me in the future and if I want to get any duration, then I'll do a TIPS ladder or a CD ladder with early withdrawal options.

Edited later: Back to intermediate being ok in my plan. Picking a bond strat has been difficult for me. Going all intermediates and just accepting loss on periods of rate increase. In my long term plan... intermediates will pay more than lose.
Last edited by quisp65 on Mon Jan 24, 2022 6:01 am, edited 1 time in total.
international001
Posts: 2746
Joined: Thu Feb 15, 2018 6:31 pm

Re: A bond duration glide path for retirement investing

Post by international001 »

I'm within the ones that don't get it, what is the point of minimizing rate risk by just (essentially) buying a bond?
You have much more risk in the stock side
sycamore
Posts: 6309
Joined: Tue May 08, 2018 12:06 pm

Re: A bond duration glide path for retirement investing

Post by sycamore »

international001 wrote: Sun Jan 23, 2022 7:34 pm I'm within the ones that don't get it, what is the point of minimizing rate risk by just (essentially) buying a bond?
You have much more risk in the stock side
Stocks are risky for sure.

Do you think a bond duration glide path is pointless for all investors?

I would imagine it's helpful for a certain investor profile. Like if your AA is heavily tilted to bonds and/or you're retired and withdrawing from your portfolio.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

international001 wrote: Sun Jan 23, 2022 7:34 pm I'm within the ones that don't get it, what is the point of minimizing rate risk by just (essentially) buying a bond?
You have much more risk in the stock side
The goal of the technique I outlined is to give investors a (relatively) easy way to use bond funds in a way that achieves the same benefit that individual bonds provide: certainty.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
international001
Posts: 2746
Joined: Thu Feb 15, 2018 6:31 pm

Re: A bond duration glide path for retirement investing

Post by international001 »

I guess I'm not striving for absolute certainty. Just relative certainty. In other words, small enough variability of results. But in my whole portfolio. As soon as I get any stocks in my portfolio, the absolute certainty of a bond is meaningless.
hudson
Posts: 7098
Joined: Fri Apr 06, 2007 9:15 am

Re: A bond duration glide path for retirement investing

Post by hudson »

vineviz wrote: Mon Jan 04, 2021 2:53 pm
17outs wrote: Mon Jan 04, 2021 12:39 pm How does one take a duration and equate it to 2 funds? Is the guideline to start with LTT adding in TIPS to reach the duration? At which point does ratio in the 2 funds flip flop? Like others said it would be cool to have the glide path not by age and more towards years until retirement and years of retirement but I understand all are different.

Yes, really all you need to know is the target duration and the duration of the two funds. I made a simple Google Sheets spreadsheet that calculates the allocation based on those three inputs: https://docs.google.com/spreadsheets/d/ ... sp=sharing

You have copy the spreadsheet or download it for Excel to change the inputs, but feel free to do that if you want to experiment.

The following formula will tell the what portion of the bonds should be in the long-duration fund:

Let TD = the target duration (e.g. 20 years);
Let DL = the duration of the longest duration fund;
Let DS = the duration of the shortest duration fund;

Code: Select all

(TD - DS) / (DL - DS)
So if the target duration is 20 years, the "long" fund has a duration of 24 years (for instance, Vanguard Extended Duration Treasury ETF), and the short fund has a duration of 8 years (e.g. an intermediate TIPS fund) then the formula is:

(20 - 8)/(24-8) = 12/16 = 75% in the longer fund, 25% in the shorter fund.

If your target duration changes to 18 years, just change the numbers: (18-8)/(24-8) = 10/16 = 63% in the longer fund, 37% in the shorter fund.

And so forth.

If you make a column in Excel with our target duration for each year, that formula will generate your "glide path" if you just copy it down a column next to it.
Most useful spreadsheet! Thanks!
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

vineviz wrote: Wed Jun 24, 2020 10:15 am Choosing a bond duration to match your investment horizon is, in actuality, the “I’m not smarter than the market” approach precisely because it eliminates interest rate risk.

Actively taking on interest rate risk (by choosing shorter duration bonds than your investment horizon,for example) is the speculative bet.
Matching your bond duration to your investment horizon doesn't "eliminate interest rate risk." For a longer term investment horizon of, say 30 years, you are actually increasing interest rate risk by locking in a fixed rate for the full term. You'd be better hedged using a blended portfolio of short and long term bonds.

By locking in long term rates today, you are making the very speculative bet that today's rates are better than what you'll see over the next 30 years. If, for example, you lock in 2.95% on the 30 year today, and next Monday rates go up to 5% you lose. That's a real risk of lost income - and just because you're not selling your (now underwater) bonds doesn't mean you are avoiding the consequences of that risk. You will lose 2% per annum on your "risk free" bet for the next 30 years.

How does a huge loss like that eliminate interest rate risk?

Now obviously if you invest a portion of your bonds short term you are taking reinvestment risk. So there is risk either way, but don't kid yourself that buying bonds whose duration matches your liabilities eliminates interest rate risk. You are just taking on a different - and no less significant - risk.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Mon Apr 18, 2022 6:44 pm Matching your bond duration to your investment horizon doesn't "eliminate interest rate risk."
It absolutely does. In fact, one of the characteristics of duration is that it reflects the point at which income uncertainty due to interest rate changes is zeroed out.

When duration of your assets (i.e. bonds) is equal to the duration of your liabilities (i.e. consumption) you have a duration gap of zero, which means that changes in interest rates/bond yields will have no impact on the cash flows you receive.

This is why keeping bond duration matched with the investment horizon is the way to eliminate interest rate risk.

This link takes you a primer on the topic from the Federal Reserve Bank of Chicago. It's a bit academic and is written from the perspective of bank balance sheets, but the concepts apply equally to a household balance sheet.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

vineviz wrote: Mon Apr 18, 2022 8:13 pm
Jaylat wrote: Mon Apr 18, 2022 6:44 pm Matching your bond duration to your investment horizon doesn't "eliminate interest rate risk."
It absolutely does.
I've shown in my above post a very simple example that clearly demonstrates how you can incur a substantial interest rate loss using your advice to buy LT treasuries to match your future liabilities. If you had indeed "eliminated interest rate risk" this would not be possible.

What is the risk you are trying to avoid by buying LTT? The risk is that interest rates can go down and you're stuck with a below market coupon. You are incurring the flip side of this same risk by locking yourself into a fixed rate for 30 years. Rates can go up and you're left holding the bag with a below market coupon. How is that different?

Regarding your link, it's a very different situation if (1) a bank borrows to match its future liabilities vs (2) Joe Bogle trying to fund his retirement. Just as Joe shouldn't have only one stock in his portfolio, neither should he place all his bets on a 30 year treasury and ignore shorter duration investments.

BTW I've worked for major money center banks for decades and am quite familiar with the concept of duration matching.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Mon Apr 18, 2022 10:37 pm
What is the risk you are trying to avoid by buying LTT?
The risk you avoid with duration matching is interest rate risk: the risk that an unexpected change in bond yields will cause a bond investor to experience a loss.

Jaylat wrote: Mon Apr 18, 2022 10:37 pm The risk is that interest rates can go down and you're stuck with a below market coupon.
Duration matching allows an investor, whether an institution or an individual, to cover their liabilities without risk of coming up short due to changes in yields. It's what makes the "fixed" aspect of "fixed income" work. It's not about trying to make a speculative bet in which you "win" or "lose" relative to other investors: it's about taking interest rate risk out of the picture.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

You are trying to avoid my arguments by dogmatically repeating your mantras about interest rate risk. Just because you say you’ve eliminated interest rate risk doesn’t mean you’ve actually done so, as I have shown in the examples above.

To state the obvious: people aren’t banks. Banks have tens of thousands of assets and liabilities with varying durations, so matching duration risk makes sense for them.

Investors trying to fund their retirement face very different risks. The challenge for an investor is to get the highest return on his investment without exceeding his risk profile. That is why Bogleheads prefer diversified investments like the 3 fund portfolio.

Putting all your fixed income into LTT significantly increases your interest rate risk. You are making a very speculative bet that today’s rates will exceed those for the next 30 years.

The other factor is that individuals are much less diversified than banks. Stuff happens – people get divorced, injured, lose jobs, their house burns down, unexpected kids, etc. Any of these events might require you to liquidate your investment portfolio.

If you are sitting on a pile of fixed rate LTT and have to sell them due to unexpected circumstances, you are going to get a very visceral lesson in interest rate risk.

Quick thought experiment: If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk? If not, why not?
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Tue Apr 19, 2022 9:18 am Putting all your fixed income into LTT significantly increases your interest rate risk. You are making a very speculative bet that today’s rates will exceed those for the next 30 years.
I'm not suggesting "putting all your fixed income into LTT": I'm suggesting that matching duration with investment horizon eliminates interest rate risk because, well, it does. By definition.

Duration matching is the opposite of speculation. If I have a liability of $D at time T, I know TODAY how much money I must put in a bond portfolio with duration T that will give me $D at time T. I'm not counting on a favorable change in bond yields in the future to cover that liability, because I've immunized the liability against such changes.

A "speculative bet" would be investing in short-term bonds with a hope that the yield curve will shift in the future in a way that is favorable. It might turn out better than the duration matching approach and it might turn out worse. The outcome is uncertain with the speculative bet, but known with duration matching.
Jaylat wrote: Tue Apr 19, 2022 9:18 am The other factor is that individuals are much less diversified than banks. Stuff happens – people get divorced, injured, lose jobs, their house burns down, unexpected kids, etc. Any of these events might require you to liquidate your investment portfolio.
All of those events (specifically, the probability-weighted value of them occurring) are inputs in determining the investment horizon as well, at least indirectly, the overall asset allocation.

If "anything can happen" is an argument for never owning long-term bonds it is also an argument for never owning stocks. In truth, it's not a good argument for either one.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

vineviz wrote: Tue Apr 19, 2022 10:12 am Duration matching is the opposite of speculation. If I have a liability of $D at time T, I know TODAY how much money I must put in a bond portfolio with duration T that will give me $D at time T.
You are conflating two very different concepts: (1) certainty of cash flows vs (2) freedom from interest rate risk. Your proposal gives you the first but exposes you to the second, as interest rates change every day.

Using your logic, stuffing your mattress full of cash also eliminates interest rate risk.
Jaylat wrote: Tue Apr 19, 2022 9:18 am Quick thought experiment: If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk? If not, why not?
This strikes to the heart of the argument.

I can see that you're heavily invested in the concept that people are just like banks. I feel it's a mistake to focus exclusively on matching duration as it results in a fixed income portfolio that is not well diversified. It's better to have a mix of durations, in the same way it's better to have a mix of stocks.

I doubt we will come to agreement.
Robot Monster
Posts: 4215
Joined: Sun May 05, 2019 11:23 am

Re: A bond duration glide path for retirement investing

Post by Robot Monster »

Jaylat wrote: Tue Apr 19, 2022 10:46 am I feel it's a mistake to focus exclusively on matching duration as it results in a fixed income portfolio that is not well diversified. It's better to have a mix of durations, in the same way it's better to have a mix of stocks.
Doesn't BLV, or EDV, or any long-term bond fund have a mix of durations, though?

Vanguard Long-Term Bond ETF (BLV)
Vanguard Extended Duration Treasury ETF (EDV)
Jaylat wrote: Tue Apr 19, 2022 10:46 am I doubt we will come to agreement.
It's a very good discussion, nevertheless. I thank you for it.
HootingSloth
Posts: 1041
Joined: Mon Jan 28, 2019 2:38 pm

Re: A bond duration glide path for retirement investing

Post by HootingSloth »

Jaylat wrote: Tue Apr 19, 2022 10:46 am
vineviz wrote: Tue Apr 19, 2022 10:12 am Duration matching is the opposite of speculation. If I have a liability of $D at time T, I know TODAY how much money I must put in a bond portfolio with duration T that will give me $D at time T.
You are conflating two very different concepts: (1) certainty of cash flows vs (2) freedom from interest rate risk. Your proposal gives you the first but exposes you to the second, as interest rates change every day.

Using your logic, stuffing your mattress full of cash also eliminates interest rate risk.
Jaylat wrote: Tue Apr 19, 2022 9:18 am Quick thought experiment: If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk? If not, why not?
This strikes to the heart of the argument.

I can see that you're heavily invested in the concept that people are just like banks. I feel it's a mistake to focus exclusively on matching duration as it results in a fixed income portfolio that is not well diversified. It's better to have a mix of durations, in the same way it's better to have a mix of stocks.

I doubt we will come to agreement.
Jaylat, I do not know if you are thinking about and referring to different concepts, but vineviz is referring to well-understood conceptions of interest rate risk, duration, and investment horizon where what he is saying more-or-less follows from the definitions. Although there are perhaps some hidden assumptions, like uniform changes in the yield curve and no changes in the volatility of interest rates, it does not seem like you are talking about those assumptions. Here is an example of how such definitions work.

When you say "If LTT go to 0%" are you referring to yields falling to 0%? (Nominal or real?) A (uniform) fall in yields should not effect how these definitions work. If yields fall, then coupon reinvestment risk will outweigh price risk for investment horizons shorter than the (Macaulay) duration, the reverse will be true for investment horizons longer than the duration, and the two will offset (and so there will be no interest rate risk) for an investment horizon equal to the duration.

If instead, you mean "If LTT's become worthless," then this should not happen because of a change in interest rates. It could happen if there were a default, but credit risk is distinct from interest rate risk.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Tue Apr 19, 2022 10:46 am
vineviz wrote: Tue Apr 19, 2022 10:12 am Duration matching is the opposite of speculation. If I have a liability of $D at time T, I know TODAY how much money I must put in a bond portfolio with duration T that will give me $D at time T.
You are conflating two very different concepts: (1) certainty of cash flows vs (2) freedom from interest rate risk. Your proposal gives you the first but exposes you to the second, as interest rates change every day.
There are two basic ways to immunize a portfolio against interest rate risk: cash flow matching and duration matching. The former is a specific application of the latter.
Jaylat wrote: Tue Apr 19, 2022 10:46 am Using your logic, stuffing your mattress full of cash also eliminates interest rate risk.
It doesn't eliminate interest rate risk unless your investment horizon is infinitesimally short. For an investor with a long-term investment horizon, hoarding cash magnifies interest rate risk. It doesn't eliminate it.
Jaylat wrote: Tue Apr 19, 2022 9:18 am
Jaylat wrote: Tue Apr 19, 2022 9:18 am Quick thought experiment: If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk? If not, why not?
This strikes to the heart of the argument.
I think it strikes to the heart of your misunderstanding about what interest rate risk is. If the yield on bonds goes to zero AFTER you've duration matched (i.e. after you've immunized your portfolio against interest rate changes), there is no longer any interest rate risk left. It won't matter whether yields go to 0% or 50%, since you've already locked in your return.
Jaylat wrote: Tue Apr 19, 2022 9:18 amIt's better to have a mix of durations, in the same way it's better to have a mix of stocks.
I think it strikes to the heart of another misunderstanding you have about bonds or, more particularly, about the risks associated with them. For any investor, the interest rate risk they face is not related to the duration of the bonds they face but instead to the gap (or difference) between the bond duration and their investment horizon.

Having a "mix of stocks" is valuable because we want to diversify away as much idiosyncratic risk as possible, since such risks are uncompensated. Interest rate risk is much more like a systematic risk, which can't be avoided by diversification and instead must be managed by hedging it. And the way to do that is through an immunization strategy, like duration matching and/or cash flow matching.

If the goal is to minimize interest rate risk then having "a mix of durations" moves you away from the goal, not towards it.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

vineviz wrote: Tue Apr 19, 2022 10:12 am
Jaylat wrote: Tue Apr 19, 2022 9:18 am The other factor is that individuals are much less diversified than banks. Stuff happens – people get divorced, injured, lose jobs, their house burns down, unexpected kids, etc. Any of these events might require you to liquidate your investment portfolio.
All of those events (specifically, the probability-weighted value of them occurring) are inputs in determining the investment horizon as well, at least indirectly, the overall asset allocation.
Can you explain how this would work in real life? Here's how I would envision this kind of logical thought process:

"I expect to retire in 30 years, but I have a 15% chance of getting a divorce in year 10, plus a 5% chance of my house burning down in year 20, plus a 30% chance of getting run over by a truck in year 25, resulting in a debilitating spinal injury."

The idea of putting all this stuff in a spreadsheet sounds very entertaining.

Again, the point is that people are not banks. Banks have clearly defined liabilities, people do not.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Tue Apr 19, 2022 12:07 pm
vineviz wrote: Tue Apr 19, 2022 10:12 am
Jaylat wrote: Tue Apr 19, 2022 9:18 am The other factor is that individuals are much less diversified than banks. Stuff happens – people get divorced, injured, lose jobs, their house burns down, unexpected kids, etc. Any of these events might require you to liquidate your investment portfolio.
All of those events (specifically, the probability-weighted value of them occurring) are inputs in determining the investment horizon as well, at least indirectly, the overall asset allocation.
Can you explain how this would work in real life? Here's how I would envision this kind of logical thought process:

"I expect to retire in 30 years, but I have a 15% chance of getting a divorce in year 10, plus a 5% chance of my house burning down in year 20, plus a 30% chance of getting run over by a truck in year 25, resulting in a debilitating spinal injury."
That's one way to do it, but I suspect this approach will appeal to only to the overly analytical. I'm definitely try not to be dogmatic about being overly precise, since financial planning inherently involves working with a great deal of uncertainty. But I do think most people who are actively managing their personal finances have a pretty good idea about whether most of their future portfolio-funded consumption lies in the next five years or is more like 10+ years away.

I think most people will prefer to aggregate things into mental accounts or "buckets", which is one reason people tend to treat their so-called "emergency funds" somewhat differently than they might create their retirement accounts. For instance, even the youngest investors might keep most of their "emergency fund" in short-term bonds, money market funds, or bank accounts while simultaneously investing their "retirement fund" 90% or more in stocks or long-term bonds.

I'll note that the latter is specifically the subject of this thread, though of course the principle would apply to any sort of expected future consumption.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Jaylat
Posts: 349
Joined: Sat Mar 12, 2016 10:11 am

Re: A bond duration glide path for retirement investing

Post by Jaylat »

vineviz wrote: Tue Apr 19, 2022 12:02 pm
Jaylat wrote: Tue Apr 19, 2022 9:18 am If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk?
If the yield on bonds goes to zero AFTER you've duration matched (i.e. after you've immunized your portfolio against interest rate changes), there is no longer any interest rate risk left. It won't matter whether yields go to 0% or 50%, since you've already locked in your return.
The market would strongly disagree with you - just take a look at the plunge in value of those low coupon LTT in your account after rates spike. And, no, it doesn't matter if you don't sell them, you are still experiencing a loss in investment income over time equal to the drop in bond price.
User avatar
Topic Author
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: A bond duration glide path for retirement investing

Post by vineviz »

Jaylat wrote: Tue Apr 19, 2022 2:15 pm
vineviz wrote: Tue Apr 19, 2022 12:02 pm
Jaylat wrote: Tue Apr 19, 2022 9:18 am If LTT go to 0%, as has happened in other countries, would you still maintain that buying LTT will eliminate interest rate risk?
If the yield on bonds goes to zero AFTER you've duration matched (i.e. after you've immunized your portfolio against interest rate changes), there is no longer any interest rate risk left. It won't matter whether yields go to 0% or 50%, since you've already locked in your return.
The market would strongly disagree with you - just take a look at the plunge in value of those low coupon LTT in your account after rates spike. And, no, it doesn't matter if you don't sell them, you are still experiencing a loss in investment income over time equal to the drop in bond price.
The market can't disagree with me, since it has no way of knowing either the average duration of my portfolio or the length of my investment horizon.

When the yield on a bond goes up, it's true that its price will go down. But the increase in yield tells us, by definition, the the future return has gone up by an offsetting amount.

Duration is nothing more or less than the estimate for how long it will take for those future returns to exactly offset the current drop in price.

You don't have to believe me: the CFA refresher readings explain it too.
The investment horizon is at the heart of understanding interest rate risk and return. There are two offsetting types of interest rate risk that affect the bond investor: cou- pon reinvestment risk and market price risk. The future value of reinvested coupon payments (and, in a portfolio, the principal on bonds that mature before the horizon date) increases when interest rates rise and decreases when rates fall. The sale price on a bond that matures after the horizon date (and thus needs to be sold) decreases when interest rates rise and increases when rates fall. Coupon reinvestment risk matters more when the investor has a long-term horizon relative to the time-to-maturity of the bond. For instance, a buy-and-hold investor only has coupon reinvestment risk. Market price risk matters more when the investor has a short-term horizon relative to the time-to-maturity. For example, an investor who sells the bond before the first coupon is received has only market price risk. Therefore, two investors holding the same bond (or bond portfolio) can have different exposures to interest rate risk if they have different investment horizons.
In short, when the investor's average bond duration is less than their investment horizon the risk is that yields will decline over their holding period. When the investor's average bond duration is more than their investment horizon the risk is that yields will rise over their holding period. And when the investor's average bond duration is equal to their investment horizon there is no interest rate risk because their duration gap is zero.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Post Reply