Question on mechanics of bond funds

[StillGoing]

While there are some differences in the mechanics of a DIY rolling bond ladder and an equivalent (in terms of maturity range) bond fund (see viewtopic.php?t=438577), the overall duration is a more important factor.

Hello StillGoing,
I would say that interest rate movements are the primary determining factor of price/yield and how this affects bonds/funds of different duration.

The duration represents the sensitivity of the fund to changes in interest rates, so the change in price (NAV) is a combination of both the duration and the change in interest rates. For a parallel change in yields (i.e., the same change across all maturities), the largest change in NAV will be observed for the longest duration fund. For non-parallel changes, the result depends on how the yield curve is changing.

The point I was making, that whether you use a DIY rolling bond ladder or a bond fund of a similar duration is largely immaterial (a few basis points in terms of returns).

cheers
StillGoing

[dbr]

I would guess that the average Boglehead like myself thinks of bond funds as something similar to a MMF or savings account where it has a bunch of bonds in it, generates interest, is pretty safe money. We know it's not exactly like that, but that's the general idea.

I think so, and that misunderstanding is what resulted in wailing and gnashing of teeth when a 3% points increase in interest rates resulted in a 15% crash in some bond fund NAV or another a couple of years ago. The fallout in some cases has been a misplaced aversion to holding bond funds at all and lots of thrashing around looking for something else.

Equities are easy to understand. You buy it at price A, it goes to price B due to what the market thinks it should be, B-A is your gains or loss.

Bond funds...so it is a bundle of bonds at different durations, but the bonds are not always held to its full duration as the managers buy and sell them behind the scenes, and the value is affected by interest rates...all of this is confusing.

Yes, it is not simple.

But then again, is understanding them necessary to effectively invest like a Boglehead? We're told to park our fixed income allocation in a total bond fund, which is maybe just fine.

Of course bond funds are just fine and to start with were/are the basic choice for a portfolio of stocks and bonds in which one accumulates wealth from postponed consumption and from which one eventually withdraws and consumes. If a basic investment background needs to be applied it would just be portfolio theory of return of combinations of assets in long term holdings. Bond funds were/are never appropriate for targeted short term holdings, cash reserves, of for specific goals with specific timing and scheduled redemption. And they are not savings accounts and never were.

[jeffyscott]

Equities are easy to understand. You buy it at price A, it goes to price B due to what the market thinks it should be, B-A is your gains or loss.

Bond funds...so it is a bundle of bonds at different durations, but the bonds are not always held to its full duration as the managers buy and sell them behind the scenes, and the value is affected by interest rates...all of this is confusing.

Yes, it is not simple.

But a stock fund is a bundle of individual stocks with different characteristics. The values of stocks are also affected by interest rates, as well as a whole lot of other confusing factors, with different stocks responding differently.

Stocks also pay out dividends, so there's that in addition to price going from a to b. You can also buy a bond at price A and it goes to price B due to what the market thinks it should be, B-A is your gains or loss (with coupon payments in addition, just like stock dividends).

The difference is we take that bond price change, do some math, and then say the yield is now X instead of Y. And then we reverse this mathematical relationship between yield and price and say "the price changed because interest rates changed", as if the interest rate just appeared on the scene independently, rather than it just being an alternate way of stating the price that the market has set.

The fact that bonds mature and coupon payments are fixed (in either real or nominal terms) is, of course, one real difference. And this difference is what allows for the more exact math to be used to convert price to yield.

(Of course, there are many different types of bonds that add other complexities, but for purposes of this discussion and comparison, Bonds = Treasuries. Except where I mentioned fixed real coupon payments.)

[Beensabu]

But then again, is understanding them necessary to effectively invest like a Boglehead?

Understanding them to the extent necessary to avoid behavioral error is probably a good idea.

[typical.investor]

So in summary, you could (higher number is more ideal):

1) spend from the intermediate
2) switch to a shorter bond fund
3) switch to an individual bond whose maturity matches the fund duration (simplified approximation)
4) switch to an individual bond whose duration matches the fund duration (more accurate but calculation required). Or use a zero coupon bond.

What is the point of 4)? If the duration of a bond (or a bond fund) describes its risk, wouldn't the two accomplish the same risk?

Yes, and that is the point. To move from a bond fund to an individual bond, you eliminate the NAV risk by choosing an individual bond with the same duration as that of the fund. If you choose a treasury zero, which pays no coupon, as your individual bond, it will mature exactly at the fund duration mark. As the coupon rises on bonds with that maturity date, the duration will shorten. It is not as safe to replace a fund with a 10 year duration with a bond that has, for example, a seven year duration. If rates rose yesterday, and you sell the fund today to fund the individual purchase, the individual bond with a seven year duration will not have suffered an equivalent NAV loss as the 10 year duration fund. So you will be selling low and buying at a higher price. A treasury zero or an individual bond with the same 10 year duration will be selling at the same price as your fund with a 10 year duration. That is what you want.

To see this look at Vanguard Long-Term Bond (BLV) for example. As of 9.2.2024 it has:

Code: Select all

Vanguard Long-Term Bond
Average duration 13.7 years	 
Average effective maturity 22.5 years 
Average coupon	3.9% 

To spend from this fund with no NAV risk, you would ideal purchase something like:

Code: Select all

US Treasury STRIP (Coupon 0%)
Maturity 05/15/2038
CUSIP: 912803DD2
Yield-to-Maturity 4.242%
Duration: 13.7 years

If you were to purchase an individual treasury whose maturity matched Vanguard Long-Term Bond's duration of 13.7 years, you would be getting something like this:

Code: Select all

US Treasury (Coupon) 4.5% 
Maturity 05/15/2038. <-- maturity matches BLV's duration
CUSIP: 912810PX0
Yield-to-Maturity 3.978%
Duration: 10.5 years. <--- duration is slightly shorter

As you can see, the duration of the individual bond whose maturity (in 05/15/2038) matches the fund duration of 13.7 years has a very different maturity than the fund. The 4.5% coupon of the individual bond reduces it's duration. You'd likely need something around a 22.5 year maturity to have a 13.7 year duration with that coupon.

In any case, a general rule of thumb is that NAV moves 1% per year of duration per 1% interest rate movement. So if rates at the 13.7 year mark on the curve rose by 1% yesterday (which is a large and probably unrealistic movement ... but just for example), you would be realizing around a 3.2% loss by moving from Vanguard Long-Term Bond (duration of 13.7 years) to the 4.5% US Treasury which matures in 13.7 years. Nobody would want to realize a 3.2% loss but keep in mind, you could lose 13.7% if you hold the fund until the spending date and rates rise 1% the day before. Obviously a treasury zero (with maturity matching the fund's duration) would be much preferred for your anticipated spending 13.7 years from now if you wanted to prevent rate (NAV) risk as you would be switching from a fund to a maturing individual bond with 0% NAV risk no matter what rates have done or will do.

Of course with funds with a shorter duration, switching to an individual bond whose maturity matches the fund's duration isn't nearly as bad:

Code: Select all

Vanguard Total Bond Market (BND)
Average duration 6.0 years	 
Average effective maturity 8.4 years 
Average coupon	3.5% 

Code: Select all

US Treasury (Coupon) 4.125% 
Maturity 08/31/2030. <-- maturity matches BND's duration
CUSIP: 91282CHW4
Yield-to-Maturity 3.769%
Duration: 5.4 years. <--- duration is short

If rates rose by 1% at the six year mark yesterday (again a pretty large and probably unrealistic movement ... but just for example), one would realize a 0.6% NAV loss by choosing an individual treasury whose maturity matches the fund's duration. One could, of course, instead choose a zero coupon bond with a maturity closest to 6 years from now to fund expenses in six years and eliminate all rate (NAV LOSS) risk.

By the way, to calculate the duration of an individual bond, you can use googlesheet and the duration function.

Code: Select all

DURATION(settlement, maturity, rate, yield, frequency, [day_count_convention])

* for settlement I use 'today()+1' as today's trades settle tomorrow
* rate is the coupon percent
* yield is the yield maturity at today's price
* frequency would be 1 if it pays interest yearly or 2 if every six months

[typical.investor]

The point I was making, that whether you use a DIY rolling bond ladder or a bond fund of a similar duration is largely immaterial (a few basis points in terms of returns).

cheers
StillGoing

For the period you are rolling your bond ladder, yes there is little difference between that an a bond fund.

What about when you need to fund expenses though? I don't believe your work has examined that. I was looking at viewtopic.php?t=438577, but did you do it someplace else?

For example, results are presented using bond funds of different durations in the Simba data set for backtesting in the thread at viewtopic.php?p=7463791 (towards the end of that thread there is also some testing for duration matching using two funds, with no significant improvements over holding a single fund). Over a retirement period of 30 years, sometimes the shortest durations (T-bills) did best and sometimes the longest durations (over 10 years) did best depending on whether yields are generally rising or falling, respectively. STT (maturities 1 to 5 years) and TBM (maturities 3 to 10 years) did best overall since they were neither worst nor (usually) best performing.

It makes sense that holding duration will perform poorly when rates are generally rising, as you would suffer NAV loss in order to fund your expenses. And it makes sense that holding duration will do well when rates are generally constant falling as you will benefit from a term premium.

Doesn't holding a ladder of individual bonds let you do both - avoiding NAV loss while maximizing the term premium?

I found ladder construction and maintenance to being a bit of a hassle especially since I didn't start 30 years in advance. In any case, I think using a bond fund for convenience and then switching to zero coupon bonds in advance (fund duration matching zero bond maturity) of your spending needs to perhaps be an ideal solution. It's too much work for me to test it historically over different interest rate environments. I don't have access to good data.

Any by the way StillGoing, your work on using funds with different durations to try and match your spending horizon was really good.

[firebirdparts]

To me the biggest difference is that in retirement, to get spending money, you can’t just sell the mature bonds in a fund and let the ladder get a foot shorter. The fund is going to let that bottom rung go, but they are going to buy bonds at the top of the duration range. With each passing year, more and more of what you’re selling for spending money is effectively way longer than you want to be selling.

You could benefit of course. Everybody talks about risk but sometimes you win when you can’t or won’t avoid the risk.

[beyou]

A fund gets you the expertise of fund managers (for active funds), the diversification of many sectors and issuers and industries (active or passive funds) and generally a higher yield.

Why higher yield, because most comments posted are comparing to ladders of TREASURY BONDS only. Bond funds such as BND and many others can usually buy higher yielding corporate, mbs, abs, interest rate swaps, bond futures, bank loans, private debt, and more. An index fund will diversify away risk but get the yield. An active fund will do same but attempt to find the greatest value among these choices based on market conditions.

Also bond fund intermediate duration can be achieved with intermediate dur bonds OR short and long bonds. The fund duration is an average of the underlying bond durations, so what you have may vary over time for active funds (it is the managers job to pick among maturities as well as above sectors, issuers etc). For index bond fund there will be a somewhat more predictable series of maturities, but can vary based on what maturities are available in the universe of sectors in the index and issued by issuers.

If you only want Treasuries, without searching for value in bonds, then maybe a bond ladder is fine. If you want more potential return then a manager can find value in the marketplace and balance the risk/reward for you.

[StillGoing]

The point I was making, that whether you use a DIY rolling bond ladder or a bond fund of a similar duration is largely immaterial (a few basis points in terms of returns).

cheers
StillGoing

For the period you are rolling your bond ladder, yes there is little difference between that an a bond fund.

What about when you need to fund expenses though? I don't believe your work has examined that. I was looking at viewtopic.php?t=438577, but did you do it someplace else?

For example, results are presented using bond funds of different durations in the Simba data set for backtesting in the thread at viewtopic.php?p=7463791 (towards the end of that thread there is also some testing for duration matching using two funds, with no significant improvements over holding a single fund). Over a retirement period of 30 years, sometimes the shortest durations (T-bills) did best and sometimes the longest durations (over 10 years) did best depending on whether yields are generally rising or falling, respectively. STT (maturities 1 to 5 years) and TBM (maturities 3 to 10 years) did best overall since they were neither worst nor (usually) best performing.

It makes sense that holding duration will perform poorly when rates are generally rising, as you would suffer NAV loss in order to fund your expenses. And it makes sense that holding duration will do well when rates are generally constant falling as you will benefit from a term premium.

Doesn't holding a ladder of individual bonds let you do both - avoiding NAV loss while maximizing the term premium?

I found ladder construction and maintenance to being a bit of a hassle especially since I didn't start 30 years in advance. In any case, I think using a bond fund for convenience and then switching to zero coupon bonds in advance (fund duration matching zero bond maturity) of your spending needs to perhaps be an ideal solution. It's too much work for me to test it historically over different interest rate environments. I don't have access to good data.

Any by the way StillGoing, your work on using funds with different durations to try and match your spending horizon was really good.

Agreed, that the modelling over in the other thread is currently limited to a single range of maturities (0 to 5 years). I've run the simulations when there are withdrawals and, not surprisingly, this changes the results - although I was initially surprised by how the results changed - hopefully, I'll get the results posted later today or tomorrow.

Are zero coupon treasury bonds available to US retail customers? I was told in another thread that they weren't (although this might only have been for TIPS not nominals) which made duration matching to a delayed income annuity somewhat more problematic than I'd initially thought (I'm in the UK, so some things are not easy to find out). While TIPS had very low coupons (0.125%), the unstripped bond was pretty close to a zero coupon, so this may have been the reason.

Thank you for the kind words.

cheers
StillGoing

[typical.investor]

Are zero coupon treasury bonds available to US retail customers? I was told in another thread that they weren't (although this might only have been for TIPS not nominals) which made duration matching to a delayed income annuity somewhat more problematic than I'd initially thought (I'm in the UK, so some things are not easy to find out). While TIPS had very low coupons (0.125%), the unstripped bond was pretty close to a zero coupon, so this may have been the reason.

Thank you for the kind words.

cheers
StillGoing

U.S. Treasury Zeros are prominently displayed on the fixed income/bond page at both Schwab and Fidelity, so yes they are readily available. I don't see them for TIPs though.

[erma]

So you will be selling low and buying at a higher price. A treasury zero or an individual bond with the same 10 year duration will be selling at the same price as your fund with a 10 year duration. That is what you want.

Is it guaranteed that two bonds (or one bond and one bond fund) have the same price if the have the same duration? I can find counter examples using the formulas but I might be plugging values of YTM, PV, CF which are never achieved in a real market scenario.

If so, and if I am correctly understanding your point, you are saying that it is sufficient to match the duration (and price) of your bond fund with that of an individual bond since they (possibly) lost the same NAV amount after (possible) rise of interest rates in the past.
If not, i.e. choosing an individual bond with same duration of my fund but different price (if even possible), will expose me if its price is higher.

Finally, after I buy the individual bond (do not consider the 0% treasure for now), I wait for those 13.7 years (the time lapse I originally planned to wait before cashing out) but my bond NAV might be negatively affected by interest rate rises of the past. What is then the whole point of selling the fund and having an individual bond?
is it the fact that in the worst case I have to wait a little bit more time after the planned duration (I wait the bond to mature)?

[jeffyscott]

I don't think that a single bond that matches the duration of the fund is necessarily going to match the performance of the fund. The yield curve does not change uniformly across all maturities.

If you want to transition from a fund to individual bonds, the closer you match the fund holdings the closer the performance should match. Taking an idealized scenario of a fund with 10% of assets in treasury zeros maturing in each of the next 10 years, buying exactly that will match the fund because you own the same bonds. Holding just a 5 year zero may not have the same performance.

In reality you have to approximate things, of course. If I were doing this, I'd look to buy a set of bonds that match the fund characteristics and also match my desired cash flow. Let's say that I own intermediate treasury index, that holds 3-10 year treasuries and has duration 5. If I decide that I want to spend that money over the next 10 years, I would buy an approximately equal amount of treasuries maturing in each of the next 10 years. So I start with the same duration, but now it declines each year.

Alternatively, for the above scenario, you could gradually move money to a treasury money market to reduce the duration of your holdings each year.

[typical.investor]

So you will be selling low and buying at a higher price. A treasury zero or an individual bond with the same 10 year duration will be selling at the same price as your fund with a 10 year duration. That is what you want.

Is it guaranteed that two bonds (or one bond and one bond fund) have the same price if the have the same duration? I can find counter examples using the formulas but I might be plugging values of YTM, PV, CF which are never achieved in a real market scenario.

No, I wouldn't expect bond prices to always be the same for the same duration.

For a treasury zero, duration equals maturity and there is no coupon. For bonds paying a coupon, the higher the coupon, the shorter the duration (holding maturity constant). So holding duration constant, I would expect different prices depending on the coupon.

If so, and if I am correctly understanding your point, you are saying that it is sufficient to match the duration (and price) of your bond fund with that of an individual bond since they (possibly) lost the same NAV amount after (possible) rise of interest rates in the past.

Yes

Finally, after I buy the individual bond (do not consider the 0% treasure for now), I wait for those 13.7 years (the time lapse I originally planned to wait before cashing out) but my bond NAV might be negatively affected by interest rate rises of the past. What is then the whole point of selling the fund and having an individual bond?
is it the fact that in the worst case I have to wait a little bit more time after the planned duration (I wait the bond to mature)?

The point is that while yes, your bond NAV may have been negatively affect by interest rates in the past, the NAV will recover in its duration. Thus, the individual bond you bought will recover as maturity approaches and it can be spend with no NAV loss.

[erma]

The point is that while yes, your bond NAV may have been negatively affect by interest rates in the past, the NAV will recover in its duration. Thus, the individual bond you bought will recover as maturity approaches and it can be spend with no NAV loss.

Right, thanks.

[typical.investor]

I don't think that a single bond that matches the duration of the fund is necessarily going to match the performance of the fund. The yield curve does not change uniformly across all maturities.

If you want to transition from a fund to individual bonds, the closer you match the fund holdings the closer the performance should match. Taking an idealized scenario of a fund with 10% of assets in treasury zeros maturing in each of the next 10 years, buying exactly that will match the fund because you own the same bonds. Holding just a 5 year zero may not have the same performance.

I absolutely wouldn't expect a treasury zero, or even a basket of treasury zeros used in the way you describe to match the (future) performance of a fund.

The fund will continue to reinvest into new bonds at new rates, buying and selling to maintain a relatively constant duration. Its performance will reflect the NAV fluctuations. The treasury zero or individual bond will mature and take NAV fluctuations out of the equation. So yes, the performance very likely will diverge and that is what you want for safe spending - to take the NAV fluctuation out of it.

If you want to transition from a fund to individual bonds, the closer you match the fund holdings the closer the performance should match. Taking an idealized scenario of a fund with 10% of assets in treasury zeros maturing in each of the next 10 years, buying exactly that will match the fund because you own the same bonds. Holding just a 5 year zero may not have the same performance.

I disagree that that would be the way to transition from a fund with duration to being able to spend without NAV risk.

Your proposal simply does not eliminate NAV risk. First, I will quibble that the duration would be actually 5.5 years and not 5, but that is just a quibble.

If the target is for spending at the duration mark (5.5 years), what will you do with the first rung that matures in a year? It is obvious that you carry reinvestment risk. And what will you do with the rung that matures in 10 years. It is obvious that you still carry rate (NAV fluctuation) risk.

Yes, you have more closely replicated the returns of the fund but your averaging reinvest risk and rate risk may or may not work. Sure, if rates fall and your realize reinvestment risk it will probably be offset by rate risk because lower rates will boost the NAV of the longer dated rungs that you would have to sell at 5.5 years to fund your spending. The actual outcome would depend on the rate curve as you say even with your attempt to resolve it.

So choosing a 5.5 year zero to replace your fund will eliminate reinvestment and rate (NAV fluctuation) risk, and it will not match the performance of a fund or ladder which retains those risks.

If your spending is in 5.5 years, the fact that shorter bonds have a relatively high yield doesn't help you. You would still have to take reinvest risk to try and capture that. Neither would a situation where longer bonds have relatively higher yields. You would have to take rate risk to try and capture that.

From the standpoint of spending certainty, a 5.5 year (or individual bond with the same duration) zero would be your best choice.

[jeffyscott]

Sure, if you want to spend all the money in year 5.5, buy the single bond that matures then. You're matching your cash flow plans and the duration of the fund.

However, if I want to spend a portion of the money in each of the next 10 years, not sure how I'm gonna do that in years 6-10 after the 5.5 year has matured? I'd want to buy bonds maturing in each year, duration will initially be the same. It diverges after because that's now what I want. I'd like to do this by spending the proceeds of each bond that the fund holds as it matures, in order do so I buy the same bonds and let them mature.

[typical.investor]

Sure, if you want to spend all the money in year 5.5, buy the single bond that matures then. You're matching your cash flow plans and the duration of the fund.

However, if I want to spend a portion of the money in each of the next 10 years, not sure how I'm gonna do that in years 6-10 after the 5.5 year has matured? I'd want to buy bonds maturing in each year, duration will initially be the same. It diverges after because that's now what I want. I'd like to do this by spending the proceeds of each bond that the fund holds as it matures, in order do so I buy the same bonds and let them mature.

If you want to spend in years 6-10, just buy a 5.5 year zero 5.5 years before the spending for the amount of spending you anticipate. So maybe buy a 5.5 year now with 20% of amount the fund, and in subsequent years do similar (25% next year, 33% the year after, then 50% then 100%) [1/years of spending remaining obviously].

Or if you have $100k in BND today (duration 6 years), and you want $20k in spending 6 years from now, you could sell $16,090 of BND and buy US Treasury STRIP 0% 08/15/2030 912833XY1. It's price is 80.45200, so it only takes $16,090 now to get $20k on 08/15/2030. Of course, that $4000 discount will be taxed along the way as if you were receiving dividends so that need to be kept in mind.

[jeffyscott]

Sure, if you want to spend all the money in year 5.5, buy the single bond that matures then. You're matching your cash flow plans and the duration of the fund.

However, if I want to spend a portion of the money in each of the next 10 years, not sure how I'm gonna do that in years 6-10 after the 5.5 year has matured? I'd want to buy bonds maturing in each year, duration will initially be the same. It diverges after because that's now what I want. I'd like to do this by spending the proceeds of each bond that the fund holds as it matures, in order do so I buy the same bonds and let them mature.

If you want to spend in years 6-10, just buy a 5.5 year zero 5.5 years before the spending for the amount of spending you anticipate. So maybe buy a 5.5 year now with 20% of amount the fund, and in subsequent years do similar (25% next year, 33% the year after, then 50% then 100%) [1/years of spending remaining obviously].

Or if you have $100k in BND today (duration 6 years), and you want $20k in spending 6 years from now, you could sell $16,090 of BND and buy US Treasury STRIP 0% 08/15/2030 912833XY1. It's price is 80.45200, so it only takes $16,090 now to get $20k on 08/15/2030. Of course, that $4000 discount will be taxed along the way as if you were receiving dividends so that need to be kept in mind.

That seems like a more complex and difficult to understand way to get off the perpetually rolling ladder that is the fund (if that's what it is intended to do).

I am not sure why you think it is a problem to simply sell 100% of the hypothetical fund that holds 1-10 year treasuries and buy those same treasuries that mature over the next 10 years, if my plan is to spend about 10% of the money in each of those 10 years?

In doing this I, initially, have not made a change to my current holdings. I keep the same duration, since it's the same bonds as the fund held. But since I now hold the individual bonds, I can spend the proceeds of the maturing bonds each year.

[typical.investor]

Sure, if you want to spend all the money in year 5.5, buy the single bond that matures then. You're matching your cash flow plans and the duration of the fund.

However, if I want to spend a portion of the money in each of the next 10 years, not sure how I'm gonna do that in years 6-10 after the 5.5 year has matured? I'd want to buy bonds maturing in each year, duration will initially be the same. It diverges after because that's now what I want. I'd like to do this by spending the proceeds of each bond that the fund holds as it matures, in order do so I buy the same bonds and let them mature.

If you want to spend in years 6-10, just buy a 5.5 year zero 5.5 years before the spending for the amount of spending you anticipate. So maybe buy a 5.5 year now with 20% of amount the fund, and in subsequent years do similar (25% next year, 33% the year after, then 50% then 100%) [1/years of spending remaining obviously].

Or if you have $100k in BND today (duration 6 years), and you want $20k in spending 6 years from now, you could sell $16,090 of BND and buy US Treasury STRIP 0% 08/15/2030 912833XY1. It's price is 80.45200, so it only takes $16,090 now to get $20k on 08/15/2030. Of course, that $4000 discount will be taxed along the way as if you were receiving dividends so that need to be kept in mind.

That seems like a more complex and difficult to understand way to get off the perpetually rolling ladder that is the fund (if that's what it is intended to do).

I am not sure why you think it is a problem to simply sell 100% of the hypothetical fund that holds 1-10 year treasuries and buy those same treasuries that mature over the next 10 years, if my plan is to spend about 10% of the money in each of those 10 years?

OK, I am all for simplicity and if your way works, then I am all for it.

However, I don't know of any such hypothetical fund. I mean if I look at something like iShares 7-10 Year Treasury Bond ETF, and needed spending in 7-10 years. I wouldn't know how to do it. Do I buy 25% each of the 7, 8, 9 and 10 year maturity?

As an approximation, perhaps you can. However, the duration for that fund is only 7.25 yrs. So to me, the spending in year 9 and 10 don't exactly match. Which may or may not matter in real life and probably only matters in the extreme (super low or high rates). Still, it holds 36% of 5-7 year maturities and 74% of 7-10 year. So it's not readily clear to me what to do.

How about Vanguard Int Term Treasury VSIGX which follows the Bloomberg US Treasury 3-10 Year Bond Index? It holds 4% 1-3 year, 54% 3-5 year, 12% 5-7 year and 30% 7-10 year. What do you buy?

For me it's easier to say it has a 5 year duration and buy zero coupons such that the weighted average is 5 years (if you wanted to separate the fund and do it all at once) or buy a zero coupon five years in advance of the spending and do the same the next year.

[jeffyscott]

How about Vanguard Int Term Treasury VSIGX which follows the Bloomberg US Treasury 3-10 Year Bond Index? It holds 4% 1-3 year, 54% 3-5 year, 12% 5-7 year and 30% 7-10 year. What do you buy?

For me it's easier to say it has a 5 year duration and buy zero coupons such that the weighted average is 5 years (if you wanted to separate the fund and do it all at once) or buy a zero coupon five years in advance of the spending and do the same the next year.

Sure, I understand that way too. What I don't (or maybe I can say "didn't" now) understand is how you go from that to doing your own ladder or terminating the ladder over a time period that differs from 5 years?

So, if I own VSIGX and now I don't want to own it any more, what I would do would depend on what I plan to do with the money.

If I want to spend an approximately equal portion of it in each year for the next 10, then I buy 1-10 year treasuries. Since I would put an approximately equal amount in each year, that should (initially) match the duration closely enough to satisfy me, even though the fund only owns 3 year and longer. Then I just let the bonds mature to unroll the ladder. (Since these are coupon bonds the duration of an equal amount 1-10, should be somewhat less than 5.5 years)

If instead, I now wanted to do my own rolling ladder with no plans to spend, I'd likely do the same and buy 1-10 year but then buy a new 10 year as each matures. That should maintain a similar duration to the fund.

Now if instead, I wanted to do my own 3-10 year, I would think buy 4-10 year. But that seems to present a problem, since I would probably want to have an equal amount in each year, so the duration would be longer than the fund. I guess I would look for a way to initially match the fund duration but then gradually transition to equal rungs. Perhaps starting with a portion in a money market and then gradually deploy that into the ladder with each purchase of a new 10 year would be one way to make that happen.

If I trade the fund for a 5 year zero coupon. I would not know how to transition to any of those. I can certainly sell and spend the zero coupon over years 1-5 or let it mature and spend it all in year 5.

If I wanted to move to my own 1-10 ladder, my first thought would be to sell about 10% per year and buy a 10 year, but I can only do that for 5 years. So maybe it's sell 20% after a year and buy 5-10 year, then sell 20% after another year and buy 4-10 year, etc.?

If I wanted to spend over 10 years, I guess it would be similar. Sell some portion each year, spend part of that and reinvest the rest in something that matures later than that initial bond. I see now that that's what you were probably laying out with "...just buy a 5.5 year zero 5.5 years before the spending...".

[daviddem]

Understanding bonds and bond ETFs is a post compiling my favorite bond educational resources.

I believe that after carefully reading and re-reading all of this material, one should get a good grasp of how it all works.

The key is to distinguish what you use bonds for:

• as part of a long-term "never sell" portfolio, for example for retirement, in which case the timeless intermediate-term government bond ETF is appropriate (unless one wants to speculate on short or medium term interest rate movements - not the Bogleheads way)

• to match a known liability at a known future time, in which case one should use either:

1. a fixed maturity bond ETF such as the iBond ETFs
2. zero coupon bonds maturing when the liability is due
3. a combination of bond ETFs of different average maturities regularly rebalanced so that the average maturity of the combination matches the deadline for the liability
4. a non-rolling bond ladder of individual bonds