Bonds funds vs rolling bond ladders: How much difference is there?

[jabberwockOG]

I give you permission to keep using the phrase.

I interpreted Tom T's reaction not so much to the caveman phrase, but to you using it to say rolling bond ladders are so very easy in comparison to a bond fund.

I probably needed a smiley face in there. I think bond ladders are fine, but I don't think my wife, if she had to do it, would appreciate it.

The "so easy" remark depends on the person. My electrician actually changed an outlet without turning off the power; he said "I've done this a thousand times." Me, I'm trying to make sure I don't electrocute myself.

I don’t think my wife will continue my ladder after I am gone either … but she knows how! More likely she will have the Edward Jones guy who has been eyeing her during pickleball manage her portfolio. lol.

Same here.

[LotsaGray]

Flat yield curve
No change in yield: Fund and ladder are identical
Step decrease in yield: Ladder performs better
Step decrease in yield: Fund performs better

There are two "Step decrease in yield" cases above. I assume the first is supposed to be "Step increase in yield"?

Well spotted, thank you. The second is supposed to be 'increase' (I've corrected the original)

When has a completely flat curve existed for any sustained period of time. Important question because that is the only time funds and individual bonds will be the same.

To do a complete comparison of bonds you need to consider what rolls. Only principle. Principle and all dividends. How are dividends rolled? What about unwinding.

Reality is funds and bonds are different. One has a fixed maturity or at least they try. The other maturity is always changing. Do that alone makes any comparison intractable for a definitive analysis.

[LotsaGray]

Bond fund
1) New money, (if any), maturing bonds, and coupons are distributed across all maturities such the fund always holds an equal number of bonds at each maturity.
2) Withdrawals (if any), are taken such that the fund is left with an equal number of bonds at each maturity.

cheers
StillGoing

It appears you assume the bond fund holds its bonds to maturity, but that is not the normal case. I think most sell bonds when they get to the lower end of the fund's maturity range. If there is a normal, positively sloped yield curve, the fund will earn a roll yield that the ladder held to maturity will miss out on. Ladder holders might also set a low maturity limit, but then they incur trading costs.

Is the roll yield enough to compensate for missing out on the last year's coupons on say a 5-year Treasury? What if the curve steepens or flattens over the period under consideration? Then, to make the analysis fair, one has to match durations. A 1-5 year fund is longer duration than a five year ladder held to maturity.

This is complicated enough to make funds attractive. I say that with most of my bonds in a non-rolling TIPS ladder. Non-rolling (with reliable real return) works for my own ever-shortening duration on this planet.

Yes, this is what I've assumed so far. There are some indices that do hold until maturity (e.g., some FTSE ones) and, as you say, many that don't.

For the rising yield curve I've used so far (i.e., yield of 1.2% and 1.4% at semesters 1 and 2), with a 3% coupon, the price is 101.58 at 1 year, while a year's worth of coupons is 3, so having a lower limit of 1 year would reduce the the amount of money per bond. Obviously, a steeper gradient (and higher coupon) would increase the benefit of having a lower limit.

cheers
StillGoing

Not many. Nearly all do not. There are only a few fixed term funds that do. They rest are constantly selling bonds getting to short snd buying longer ones to maintain only small changes to maturity

[CuriousGeorgeTx]

I think the difference is mainly psychological for me. In my mind the ladder is more transparent with what is happening with the value of each rung of my ladder as interest rates change and I play the mind game I am not going to sell until maturity so the fluctuations don’t matter. Buying a single bond helped me understand what was happening in my bond fund (when I used to own bond funds) on a grander scale.

This is the way I look at it as well. I'm retired and have a fixed income ladder (CDs, Treasuries (including TIPS)) for Residual Living Expenses (what pension, SS and earnings on taxable doesn't cover). I don't worry about market fluctuations because I know that when the rung matures, I will get the face value. I go through an exercise every year as a rung matures in which I decide whether to add another rung to the end of the ladder (by selling equities) or not. If the market is "good", I extend, at least one year; more if there were previous "bad" years when I didn't extend. Ladder currently goes through 2032.

While I'm quite happy with my system, if I could convince myself that a bond fund could be a satisfactory replacement for the RLE ladder, I would like that because I'm not sure DW would want to mess with the ladder should I predecease her. But in my head, it seems like there is a lot of risk that I would end up selling the bond fund at a loss when I needed it most.

It's not just the mathematics that have been so well discussed in this thread. It's the psychology as well. I have borrowed the term "moat" for my ladder from a BHer, and with it, I sleep well and feel free to spend freely.

[StillGoing]

Taking fixed withdrawals

Up until now, no new money has been added or withdrawals taken. In this post, we will look at the effect of withdrawals. While there are a number of approaches that could be adopted, in this test a constant withdrawal of $400 per year (4% of the original portfolio) is taken in two tranches ($200 per semester, starting with the beginning of Semester 1). This amount is higher than the income derived from the coupons, so the amount in the ladder/fund will gradually reduce. The amount chosen is sufficiently small to ensure that no ladder rung becomes exhausted before the end of the simulation (i.e., after 14 semesters).

Example case

The NAV for the ladder and fund as a function of time are given in the following table together with the annualised return (including the effect of the withdrawal) and the difference in returns for a constant non-inverted yield curve (y=1+0.2s).

Code: Select all

	Semester
	0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	AR (%)	Diff (bp)
Rolling	10000	9948	9895	9840	9784	9728	9670	9611	9552	9491	9431	9370	9308	9245	9180	-1.21	
Fund	10000	9948	9895	9841	9786	9730	9674	9617	9559	9500	9440	9380	9318	9256	9192	-1.20	-1.9

The fund does slightly better by 1.9 bp. This can be compared with the no withdrawal case (see summary table in previous post) where the ladder did better than the fund by 4.4 bp. In other words, taking withdrawals can affect which of the ladder or fund does best.

The explanation for this behaviour comes back to the fundamental operational difference between the fund and the ladder, where the ladder takes the maturing bond and coupons and reinvests them in the highest maturity bond, while the fund (in the simplified form used here) invests the proceeds across all maturities such that the number of bonds at each maturity is equal. When withdrawals exceed the coupons, for the fund this effectively means selling bonds across all maturities to buy the one with the highest maturity and the number of bonds held will reduce (when withdrawals exactly equal the coupons, then the number of bonds held will remain constant). For the non-inverted yield curve, this means the bonds at maturities between the two extremes (1 to 9 semesters) are sold at higher than par and therefore the fund does better than the ladder.

The effect of changing the annual withdrawal amount on the difference in annual returns (in bp) after 7 years (+ve means ladder does better and -ve means fund does better) for the y=1+0.2s yield curve is given in the following table

Code: Select all

	Withdrawal amount ($)
0	100	200	300	400	500
4.4	3.1	1.7	0.1	-1.9	-4.1

For withdrawals at or below the value of the coupons (just under $150 per semester in this case) the ladder performs best (note that although not shown, this also includes accumulation, e.g., for a $400 per year accumulation, the ladder has a return of 8 bp higher than the fund), while ones withdrawals exceed the coupons, then the fund does better.

Summary of results

In the following table, the results for various yield curves and steps are summarised where an annual withdrawal of $400 has been made. Note that the annual return quoted is after withdrawals.

Code: Select all

		Ann ret (%)
Scenario	Rolling	Fund	Diff (bp)
Flat
no change (3%)	-1.14	-1.14	 0.0
3% to 2%	-1.71	-1.71	-0.2
3% to 4%	-0.57	-0.57	 0.2
Non-inverted (x+0.2s)
no change (x=1)	-1.21	-1.20	-1.9
x=1 to x=0	-1.78	-1.75	-2.9
x=1 to x=2	-0.64	-0.64	 0.0
Inverted (3-0.2s)
no change (x=3)	-3.36	-3.41	 6.2
x=3 to x=2	-3.92	-3.98	 6.5
x=3 to x=4	-2.79	-2.84	 5.9

For the flat yield curve, there is still little difference between the fund and the ladder once withdrawals are taken.

For the non-inverted yield curve, whereas with no withdrawals the ladder performed best (see previous post), once withdrawals exceeded the coupons, the fund performed best, except where there was a step increase in yield. As explained above, this is because once withdrawals exceed the coupons, bond are sold rather than bought, so the inverted yield curve which leaves bonds at maturities from 1 to 9 with prices higher than par, benefits the fund more than the ladder.

For the inverted yield curve, the opposite outcome occurs – whereas with no withdrawals, the fund performed best, once withdrawals are taken and bonds must then be sold, the prices of the bonds with maturities 1 to 9 are below par and this results in the ladder performing better than the fund.

I think that will be it as far as results go - just need to make some sense of it all now!

cheers
StillGoing

[StillGoing]

Flat yield curve
No change in yield: Fund and ladder are identical
Step decrease in yield: Ladder performs better
Step decrease in yield: Fund performs better

There are two "Step decrease in yield" cases above. I assume the first is supposed to be "Step increase in yield"?

Well spotted, thank you. The second is supposed to be 'increase' (I've corrected the original)

When has a completely flat curve existed for any sustained period of time. Important question because that is the only time funds and individual bonds will be the same.

To do a complete comparison of bonds you need to consider what rolls. Only principle. Principle and all dividends. How are dividends rolled? What about unwinding.

Reality is funds and bonds are different. One has a fixed maturity or at least they try. The other maturity is always changing. Do that alone makes any comparison intractable for a definitive analysis.

I'd agree - very rarely, which is why non-inverted and inverted yield curves have also been looked at (with and without a step in time). IMV, it is useful to start with a simple case and understand how that works before moving on to more complex scenarios.

In terms of a ladder, what rolls is entirely up to the person running it. The approach I've adopted here where the coupons and maturing bond are then reinvested in the highest desired maturity is one way of going about it and seems to be close enough to how some people on these boards implement their rolling ladders.

However, for an index, all bonds that fulfil the criteria for inclusion must be included in the fund. Criteria are usually (but not always) defined by type (i.e., nominal or TIPS), maturity range, and a threshold amount in issue. Bonds within the maturity range include those already issued, newly issued bonds, and bonds that enter the maturity range. Where an index has a lower limit to maturity bonds that fall below that will leave the index and must be sold. Where callable bonds exist this further complicates the situation (I extended a collection of the UK government bond, gilt, indices back to the 19th century using historic gilt prices, coupons, and issue weights, see https://papers.ssrn.com/sol3/papers.cfm ... id=4742450 and callable bonds were a distinct pain when prices were hovering around par) since they can move in and out of the index on a daily basis. Currently no UK gilts are callable and I believe that is also the case for US treasuries.

So, I agree with you in one respect - reality is far more complicated than the simple model I have adopted here (although based on the same mathematics). But, interestingly, that simplicity allows what I think is useful insight into the differences in behaviour.

cheers
StillGoing

[StillGoing]

I think the difference is mainly psychological for me. In my mind the ladder is more transparent with what is happening with the value of each rung of my ladder as interest rates change and I play the mind game I am not going to sell until maturity so the fluctuations don’t matter. Buying a single bond helped me understand what was happening in my bond fund (when I used to own bond funds) on a grander scale.

This is the way I look at it as well. I'm retired and have a fixed income ladder (CDs, Treasuries (including TIPS)) for Residual Living Expenses (what pension, SS and earnings on taxable doesn't cover). I don't worry about market fluctuations because I know that when the rung matures, I will get the face value. I go through an exercise every year as a rung matures in which I decide whether to add another rung to the end of the ladder (by selling equities) or not. If the market is "good", I extend, at least one year; more if there were previous "bad" years when I didn't extend. Ladder currently goes through 2032.

While I'm quite happy with my system, if I could convince myself that a bond fund could be a satisfactory replacement for the RLE ladder, I would like that because I'm not sure DW would want to mess with the ladder should I predecease her. But in my head, it seems like there is a lot of risk that I would end up selling the bond fund at a loss when I needed it most.

It's not just the mathematics that have been so well discussed in this thread. It's the psychology as well. I have borrowed the term "moat" for my ladder from a BHer, and with it, I sleep well and feel free to spend freely.

To be a pedant, while this is a great solution (and one discussed in Zwecher's book on Retirement Portfolios), this is not a rolling ladder - it is a non-rolling (or collapsable) ladder that is extendable from an external source (i.e., equities).

cheers
StillGoing

[StillGoing]

Bond fund
1) New money, (if any), maturing bonds, and coupons are distributed across all maturities such the fund always holds an equal number of bonds at each maturity.
2) Withdrawals (if any), are taken such that the fund is left with an equal number of bonds at each maturity.

cheers
StillGoing

It appears you assume the bond fund holds its bonds to maturity, but that is not the normal case. I think most sell bonds when they get to the lower end of the fund's maturity range. If there is a normal, positively sloped yield curve, the fund will earn a roll yield that the ladder held to maturity will miss out on. Ladder holders might also set a low maturity limit, but then they incur trading costs.

Is the roll yield enough to compensate for missing out on the last year's coupons on say a 5-year Treasury? What if the curve steepens or flattens over the period under consideration? Then, to make the analysis fair, one has to match durations. A 1-5 year fund is longer duration than a five year ladder held to maturity.

This is complicated enough to make funds attractive. I say that with most of my bonds in a non-rolling TIPS ladder. Non-rolling (with reliable real return) works for my own ever-shortening duration on this planet.

Yes, this is what I've assumed so far. There are some indices that do hold until maturity (e.g., some FTSE ones) and, as you say, many that don't.

For the rising yield curve I've used so far (i.e., yield of 1.2% and 1.4% at semesters 1 and 2), with a 3% coupon, the price is 101.58 at 1 year, while a year's worth of coupons is 3, so having a lower limit of 1 year would reduce the the amount of money per bond. Obviously, a steeper gradient (and higher coupon) would increase the benefit of having a lower limit.

cheers
StillGoing

Not many. Nearly all do not. There are only a few fixed term funds that do. They rest are constantly selling bonds getting to short snd buying longer ones to maintain only small changes to maturity

Sorry, I should have been more precise - I was thinking about the indices that either end at 1 year or 0 years (e.g., there are treasury funds with maturity ranges of 1 to 3 years and 1 to 5 years, but some go from 0 to 5 years - in the UK, for these sorts of funds the FTSE indices typically go from 0 and the MSCI indices go from 1 year). Of course, a fund with a maturity range of 3 to 10 years will never allow any bonds to mature.

cheers
StillGoing

[typical.investor]

Taking fixed withdrawals


The explanation for this behaviour comes back to the fundamental operational difference between the fund and the ladder, where the ladder takes the maturing bond and coupons and reinvests them in the highest maturity bond, while the fund (in the simplified form used here) invests the proceeds across all maturities such that the number of bonds at each maturity is equal.

Won't the fund in the form you used be shortening its duration as time goes by? It isn't maintaining a constant duration like a normal fund would, is it?

[jeffyscott]

I think that will be it as far as results go - just need to make some sense of it all now!

If I haven't missed anything, all of the differences were very small (about the same or less than than a bond index fund ER) and they could go either way.

I found a similar difference, where the intermediate treasury fund resulted in about 8 bp per year less, in my 10 year example that I posted about above. And note that this was under the assumption of a perfectly constructed ladder, where every penny from the maturing bonds was needed for some other purpose exactly when that money became available.

I did some more fiddling with that and found that doing the two fund thing reversed the outcome. If I used intermediate treasury fund and treasury money market, starting with 90% VSIGX and 10% VUSXX and moving in 10% annual increments to 100% VUSXX the funds provided 3 bp per year more than the ladder.

I also tried substituting a total treasury fund (GOVT), this increased the shortfall to about 12 bp. But if I moved 10% per year to VUSXX, the shortfall was reversed just as it was with the intermediate fund. In this case it was a 2 bp difference in favor of the funds.

I think all of these results confirm that there's little difference between a ladder and a fund when used at least somewhat appropriately.

It may be possible at times to construct a ladder of CDs and get a significantly better after-tax yield that way. Since there are no CD funds, that's the only way to gain that benefit.

Additionally, as someone noted above, to the extent you want to hold corporate bonds a fund is likely the only way to go. FWIW, substituting total bond for VSIGX provided about 2.4% per year more than the ladder of treasuries.

[Doc]

Taking fixed withdrawals


The explanation for this behaviour comes back to the fundamental operational difference between the fund and the ladder, where the ladder takes the maturing bond and coupons and reinvests them in the highest maturity bond, while the fund (in the simplified form used here) invests the proceeds across all maturities such that the number of bonds at each maturity is equal.

Won't the fund in the form you used be shortening its duration as time goes by? It isn't maintaining a constant duration like a normal fund would, is it?

Indeed the rolling ladder may shorten it duration SOME OF THE TIME. One factor is how many rungs you have in your ladder. If you setup up your ladder with 52 week rungs your ladder duration doesn't shorten in you set it up to have a rung at each original duration - typically 4 weeks apart. It also has a longer duration than the fund which is buying all maturing rungs. Some of those rungs are not the longest maturity. A 10-3 fund is buying 5, 7 and 10 year bonds while the ladder is only buying 10's. Also this means a ladder has a longer overall duration than the fund because is it only buying the longest duration note while the fund is buying all the duration new issues not just he longest.

[typical.investor]

EDIT ... nevermind.

I misunderstood how the bond fund was reinvesting.

[Doc]

2) I would definately expect a fund that is shortening duration to outperform a ladder with withdrawals over a period of rates rising. However, the question I really have is how much does a rolling ladder help with withdrawals in a period of rising rates relative to a constant duration ladder.

So I don't accept the conclusion that funds outperform ladders in rising rates while making withdrawals as valid. The fund has a shorter average maturity and will also have a shorter duration than the rolling ladder the way this was set up. Of course shorter duration does better in a period of rising rates.

The outperformance of a ladder as compared to a fund is not a matter of duration. When you build a rolling ladder you adjust the original ladder so that it has the basically the same average maturity as the fund that you are trying to compare. So if you want to mimic a 3-10 fund you might use a 3-7 ladder. The advantage of the ladder is that you get greater return from selling the individual rungs at maturity. Since each "rung" in the ladder is longer than the average "rung" in the fund you get some benefit from greater capital gains as well as the higher coupon from the longer average bond in the ladder. In a taxable account you can achieve a little more from the lower cap gains rate.

In any case with the current inverted yield curve a ladder is not a good idea at this time.

[rebellovw]

Bond ladders are going to have a constant decreasing duration which in times of interest decreasing will hurt performance a small amount but of course exact same durations should produce same results mathematically.

Over all I use a 5 year treasury bond ladder - a rung matures and I reinvest it, or rebalance it to equities, or spend it. So easy a caveman can do it. No psychological concern about interest rate changes effects on my bond portfolio.

Please stop saying this. It's not easy, and requires you (or a spouse) to make a decision every single time. "Should I spend it? No, there's no need for the money right now" (until six months later, there is an unexpected need, and now you have to sell before maturity.)

You know what is so simple that a caveman could do it? Owning a bond fund and never making any decisions on what to do with it.

100% agreed - thanks Tom!

[StillGoing]

I think that will be it as far as results go - just need to make some sense of it all now!

If I haven't missed anything, all of the differences were very small (about the same or less than than a bond index fund ER) and they could go either way.

I found a similar difference, where the intermediate treasury fund resulted in about 8 bp per year less, in my 10 year example that I posted about above. And note that this was under the assumption of a perfectly constructed ladder, where every penny from the maturing bonds was needed for some other purpose exactly when that money became available.

I did some more fiddling with that and found that doing the two fund thing reversed the outcome. If I used intermediate treasury fund and treasury money market, starting with 90% VSIGX and 10% VUSXX and moving in 10% annual increments to 100% VUSXX the funds provided 3 bp per year more than the ladder.

I also tried substituting a total treasury fund (GOVT), this increased the shortfall to about 12 bp. But if I moved 10% per year to VUSXX, the shortfall was reversed just as it was with the intermediate fund. In this case it was a 2 bp difference in favor of the funds.

I think all of these results confirm that there's little difference between a ladder and a fund when used at least somewhat appropriately.

It may be possible at times to construct a ladder of CDs and get a significantly better after-tax yield that way. Since there are no CD funds, that's the only way to gain that benefit.

Additionally, as someone noted above, to the extent you want to hold corporate bonds a fund is likely the only way to go. FWIW, substituting total bond for VSIGX provided about 2.4% per year more than the ladder of treasuries.

Your conclusion is about where I have ended up too. There are clearly circumstances where the ladder will outperform the fund (e.g., non-inverted yield curve during accumulation) and other occasions where the ladder will underperform the fund (e.g., non-inverted yield curve during decumulation). However, what would be best over a 30 year period is not clear to me, although your modelling for the last decade or so (if I've understood correctly what you've done) gives some idea - although the conditions over that decade (steadily falling yields followed by a period in which they rose rapidly is an interesting one and possibly explains why the addition of the money market fund improved matters).

cheers
StillGoing

[StillGoing]

Limitations of the results

While I think the results I’ve presented here are useful, there are some limitations to what has been done and how it has been done.

Bond fund operation
There are a number of aspects of how the bond fund has been simulated that are considerably simplified compared to reality.
1) Newly issued bonds. In the model it is assumed that the only newly issued bond is the one at the highest maturity (5 years in the example). In reality, treasuries are issued with maturities of 2, 3, 5, 7, 10, 20 and 30 years. This means that a larger proportion of fund would be at par when coupons and maturing bonds are reinvested.
2) Bonds entering the maturity range of the fund. In the model, it is assumed that only the bonds issued at the highest maturity of the fund exist. However, in reality, bonds that were previously outside the maturity range of the fund, can become eligible when their term to maturity falls sufficiently (what are sometimes known as ‘shorteners’).
3) Bonds are issued at a greater frequency than 6 months

All of these considerations mean that the number of bonds held in the fund would be larger and have a wider range of coupons than in the model and, consequently, the behaviour with changing yields would be somewhat more complex. For example, there are currently (4 September 2024) 212 treasuries in the maturity range of 0 to 5 years (although whether they are eligible for inclusion in an index depends on whether they meet the amount in issue threshold – older bonds tend to have lower amounts in issue) with coupons ranging from 0.25% to 7.625%. The yield curve over that range of maturities is currently inverted with asked yields to maturity (not the necessarily the same as par yield) of around 5.0% at maturities less than 3 months to about 3.6% at 5 years (I note that after 5 years, yields start increasing again although only by a few 10s of bp).

How much difference this additional complexity would make to the returns is difficult to know, although since it is the shape of the yield curve and whether the prices of the bonds are above or below par that determines the whether the fund does better than the ladder or not, perhaps not too much.

Historical yield curves
One advantage of using relatively static yield curves in the model means that the effect of individual properties of the yield curve (e.g., the gradient of the curve or a single step change in time) can be investigated and insight gained. Of course, one obvious question is then how would would a ladder have done compared to holding a broadly equivalent fund over a historical period. One way of doing this is to use historical yield curves (which, on a daily basis go back to the early 1960s) to model the rolling ladder and compare this with the outcome of existing indices (e.g., data exists for the Barclays US Treasury 1-5 yr index back to 1975). The basis of the tools to do this already exist (i.e., the bond fund simulator described at viewtopic.php?t=179425 for the ladder, although it would need modification to account for accumulation or decumulation) and the Simba spreadsheet for the fund/index returns, albeit only on an annual basis). If I understand correctly, a similar approach has been adopted upthread by jeffyscott viewtopic.php?p=8022353#p8022353 albeit over a relatively small number of years (2013 onwards?)

Maturity range
Only a single range of maturities has been modelled. I chose 0 to 5 years since it backtested fairly well historically (e.g., see viewtopic.php?p=7463791) and, with only 10 maturities (11 including 0 years) was fairly manageable to implement as a spreadsheet. Of course, using only a single range of maturity limits the scope of the results that have been presented. However, including longer maturities, and hence higher durations, will accentuate the differences between the ladder and fund, particularly when yields are changing. In other words, it is likely that the differences of a few basis points found here would be larger.

cheers
StillGoing

[StillGoing]

Complexity

Leaving aside whether a rolling ladder or bond fund is better than the other in terms of performance, another consideration is that of complexity. In this post, I'll make an attempt to compare the steps required for each approach.

Accumulation

Fund: Every time some new money is added, some additional units of the bond fund will need to be purchased. Leaving rebalancing aside (assuming equities are held as well as bonds), this could be largely automated, although manual purchases would be relatively simple.

Rolling ladder: Every time new money is added, that together with any cash arising from coupons or maturing bonds will need to used to purchase the bond with the highest maturity in the desired range. Assuming the frequency at which new money is made available corresponds to the schedule of auctions of the relevant bond, then this is a relatively simple manual purchase (but may be difficult to automate).

Decumulation

Fund: The amount required to be withdrawn from the bond fund (which will depend on the withdrawal strategy in use and the relative performance of equities and bonds) will be done by selling the appropriate number of units of the bond fund. Depending on the withdrawal strategy adopted, automating this process may be difficult, so a manual sale is likely.

Rolling ladder: As with the fund, the amount to be withdrawn from the ladder will depend on the withdrawal strategy in place. If the amount of cash from maturing bonds and coupons is greater than the amount to be withdrawn then the surplus would need to be used to purchase a bond at the highest maturity. If the cash from maturing bonds and coupons is below the amount to be withdrawn, then either bonds from additional rungs (starting from the lowest maturity?) must be sold make up the shortfall or a reduced withdrawal taken. The former potentially means not selling at par (although, depending on the exact maturity, coupon, and yields, the price may not be too far from par).

Rebalancing

If rebalancing from equities to bonds, then the process is the same as that of accumulation. If rebalancing from bonds to equities, the process is the same as that of withdrawal.

Overall

It is clear that during accumulation the complexity of operating a rolling ladder is fairly similar to that of using a bond fund save that it is more difficult to automate the process. During withdrawals the number of steps is the same (one purchase or one sale with the ladder compared to one sale with the fund). One additional element of complexity, is that during accumulation and withdrawal, where the amount matured together with coupons is larger than the amount withdrawn, the new bond with the highest maturity in the required range must be identified before purchase.

cheers
StillGoing

[prioritarian]

During withdrawals the number of steps is the same

This isn't true because a withdrawal in a ladder will often require the sale of one or more bonds and reinvestment of residual unneeded funds. Using the residual funds (and other funds) to re-purchase and/rebalance a bond ladder could be difficult given the non-continuous availability of bonds at specific denominations on the second/primary markets. (Some brokers offer fractional bond trading which would obviate this issue but this is very rare at this time.)

A bond fund ladder allows the withdrawal of funds down to a fraction of a dollar and obviates this entire issue.

[gavinsiu]

Accumulating and decumulating seems a bit more complicated with rolling bond funds. As you say, if you add new money, you can buy the highest maturity, resulting in some uneven rungs at the higher end or buy whatever rung is the most attractive? The same thing happens during decumulation. There is the talk of being able to withdraw the lowest rung, but that would make the ladder shorter or lopsided. May be that can be fix later with the next stock/bond rebalance?

When you sell or buy from a bond, are you essentially buying and selling slice of the entire ladder equivalent?

[StillGoing]

During withdrawals the number of steps is the same

This isn't true because a withdrawal in a ladder will often require the sale of one or more bonds and reinvestment of residual unneeded funds. Using the residual funds (and other funds) to re-purchase and/rebalance a bond ladder could be difficult given the non-continuous availability of bonds at specific denominations on the second/primary markets. (Some brokers offer fractional bond trading which would obviate this issue but this is very rare at this time.)

A bond fund ladder allows the withdrawal of funds down to a fraction of a dollar and obviates this entire issue.

Assuming there is enough cash from maturing bonds and coupons, then sales should be rare during withdrawal until the amount withdrawn becomes a significant fraction of the NAV (either towards the end of retirement or when stocks have crashed and most of a withdrawal has fallen on the bond side of things). My understanding, from a perusal of past auctions (e.g. at https://www.treasurydirect.gov/auctions/auction-query/), is that some maturities (e.g., 5-years) have been auctioned very frequently. Looking on https://www.wsj.com/market-data/bonds/treasuries, there appears to be at least one treasury maturing each month in 2029 (i.e., 5 years away) all of which were traded yesterday. However, for a longer ladder the available bonds are sparser (e.g. only 3 bonds maturing in 2034 on the wsj list). I also note that, as you say, the model assumes fractional bonds, and integer bonds would mean that sometimes cash would be left over (by way of comparison, in the UK, gilts are denominated in GBP, £, and holdable down to the the nearest GBX, i.e., penny). Practically, the residual cash could be stored in a money market fund until enough is accumulated to add to the ladder (I think at least one forum member operating a ladder does this), although this does add an extra step. I note many bond funds will keep some cash to act as a cashflow buffer and reduce their transaction charges).

cheers
StillGoing

[StillGoing]

Accumulating and decumulating seems a bit more complicated with rolling bond funds. As you say, if you add new money, you can buy the highest maturity, resulting in some uneven rungs at the higher end or buy whatever rung is the most attractive? The same thing happens during decumulation. There is the talk of being able to withdraw the lowest rung, but that would make the ladder shorter or lopsided. May be that can be fix later with the next stock/bond rebalance?

When you sell or buy from a bond, are you essentially buying and selling slice of the entire ladder equivalent?

Yes, the rungs of the ladder hold uneven amounts - during accumulation the number of bonds held at the highest maturities will always be greater, but they will gradually shuffle downwards in maturity

In the following table, the number of bonds at each maturity during accumulation is given while adding $1000 per year to a ladder that started at $10000 (each row is a successive semester). This is with a rising yield curve (y=1+0.2s).

Code: Select all

Maturity (semester)
0	1	2	3	4	5	6	7	8	9	10
0.00	9.91	9.84	9.80	9.77	9.76	9.77	9.80	9.85	9.92	10.00
0.00	9.84	9.80	9.77	9.76	9.77	9.80	9.85	9.92	10.00	16.39
0.00	9.80	9.77	9.76	9.77	9.80	9.85	9.92	10.00	16.39	16.42
0.00	9.77	9.76	9.77	9.80	9.85	9.92	10.00	16.39	16.42	16.47
0.00	9.76	9.77	9.80	9.85	9.92	10.00	16.39	16.42	16.47	16.54
0.00	9.77	9.80	9.85	9.92	10.00	16.39	16.42	16.47	16.54	16.64
0.00	9.80	9.85	9.92	10.00	16.39	16.42	16.47	16.54	16.64	16.75

You can see the '16.39' bonds gradually working its way down in maturity.

Provided the withdrawal is smaller than the total of maturing bonds and coupons, then no rung will go to zero. However, if the withdrawal is larger than the maturing bonds and coupons, then yes, there will be no cash to reinvest at the highest maturity. The latter would happen either towards the end of retirement if the overall portfolio (i.e., including equities) is getting small compared to withdrawals or if a significant rebalancing is required from bonds to stocks (e.g., after a market crash). Effectively, the rolling ladder converts to a collapsing ladder towards the end of its life. There are several ways this could be dealt with after rebalancing, but it would be interesting to hear from those actually running rolling ladders what is in their IPS.

cheers
StillGoing

[StillGoing]

A few final thoughts

As others have already said it appears that the difference in performance between a rolling ladder and a bond index fund of the same maturity range is relatively small (e.g., compared to the standard deviation of returns), although the difference is likely to be larger for a larger maturity range than the 0 to 5 years used in the examples here. The difference in returns depends on the cash flow in or out of the ladder/fund (i.e., whether in accumulation or decumulation), the gradient in the yield curve, and changes in the yields (only parallel changes have been looked at), and therefore under some circumstances a ladder would be better and in others a fund. This dependence arises because of the differences in how the ladder and fund buy or sell bonds. In the ladder, bonds are usually bought and sold at par (assuming the lowest rung is at term), while this is not always the case for the fund. Therefore, if the price of the bonds between the extremities in maturity are higher than par, the fund will do better than the ladder if selling (e.g, during withdrawal) and worse when buying. The opposite is true when the price of those bonds is below par.

The, admittedly limited, results indicate that a non-inverted yield curve (which is the most commonly occurring) tends to favour the ladder during accumulation and the fund during decumulation (see summary tables in posts viewtopic.php?p=8024458#p8024458 and viewtopic.php?p=8023630#p8023630).

Since the argument for (and against) rolling ladders crops up on the bogleheads forum from time to time, the motivation for doing this work was to see if some definitive argument could be made either way. The results presented here tentatively suggest that there may be some relatively small gains to be found by adopting a ladder during accumulation and a fund during decumulation. However, I think a more substantial analysis is required to confirm that suggestion.

cheers
StillGoing

[Parkinglotracer]

A few final thoughts

As others have already said it appears that the difference in performance between a rolling ladder and a bond index fund of the same maturity range is relatively small (e.g., compared to the standard deviation of returns), although the difference is likely to be larger for a larger maturity range than the 0 to 5 years used in the examples here. The difference in returns depends on the cash flow in or out of the ladder/fund (i.e., whether in accumulation or decumulation), the gradient in the yield curve, and changes in the yields (only parallel changes have been looked at), and therefore under some circumstances a ladder would be better and in others a fund. This dependence arises because of the differences in how the ladder and fund buy or sell bonds. In the ladder, bonds are usually bought and sold at par (assuming the lowest rung is at term), while this is not always the case for the fund. Therefore, if the price of the bonds between the extremities in maturity are higher than par, the fund will do better than the ladder if selling (e.g, during withdrawal) and worse when buying. The opposite is true when the price of those bonds is below par.

The, admittedly limited, results indicate that a non-inverted yield curve (which is the most commonly occurring) tends to favour the ladder during accumulation and the fund during decumulation (see summary tables in posts viewtopic.php?p=8024458#p8024458 and viewtopic.php?p=8023630#p8023630).

Since the argument for (and against) rolling ladders crops up on the bogleheads forum from time to time, the motivation for doing this work was to see if some definitive argument could be made either way. The results presented here tentatively suggest that there may be some relatively small gains to be found by adopting a ladder during accumulation and a fund during decumulation. However, I think a more substantial analysis is required to confirm that suggestion.

cheers
StillGoing

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

[gavinsiu]

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I agree that if you want more transparency and control in a bond portfolio, the ladder is the way to go. However, I feel that a number (percentage unknown) adopt the ladder solely to avoid losses because of the feature of redeeming the lowest rung.

[rkhusky]

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I agree that if you want more transparency and control in a bond portfolio, the ladder is the way to go. However, I feel that a number (percentage unknown) adopt the ladder solely to avoid losses because of the feature of redeeming the lowest rung.

Ease of use makes a bond fund a winner for me. Never have to stress about when a bond (or a CD) is going to mature.
Perhaps feeling in control is why a lot of people buy individual stocks. They don’t have to own the “bad” stocks that are in the indexes.

[Tom_T]

However, I feel that a number (percentage unknown) adopt the ladder solely to avoid losses because of the feature of redeeming the lowest rung.

I think so. How much angst has been expressed on the Forum over Total Bond Market losing 13% in 2022 - a year that saw a rise from historically low interest rates? Has anyone noticed that it was up 5.7% last year, and is up 3.05% YTD?

[prioritarian]

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I think a total treasury fund (e.g. GOVT) or a duration-matched bond fund ladder (e.g. some combination of VGSH, VGIT, VGLT) is just as "transparent" as holding a bond ladder.

[StillGoing]

A few final thoughts

As others have already said it appears that the difference in performance between a rolling ladder and a bond index fund of the same maturity range is relatively small (e.g., compared to the standard deviation of returns), although the difference is likely to be larger for a larger maturity range than the 0 to 5 years used in the examples here. The difference in returns depends on the cash flow in or out of the ladder/fund (i.e., whether in accumulation or decumulation), the gradient in the yield curve, and changes in the yields (only parallel changes have been looked at), and therefore under some circumstances a ladder would be better and in others a fund. This dependence arises because of the differences in how the ladder and fund buy or sell bonds. In the ladder, bonds are usually bought and sold at par (assuming the lowest rung is at term), while this is not always the case for the fund. Therefore, if the price of the bonds between the extremities in maturity are higher than par, the fund will do better than the ladder if selling (e.g, during withdrawal) and worse when buying. The opposite is true when the price of those bonds is below par.

The, admittedly limited, results indicate that a non-inverted yield curve (which is the most commonly occurring) tends to favour the ladder during accumulation and the fund during decumulation (see summary tables in posts viewtopic.php?p=8024458#p8024458 and viewtopic.php?p=8023630#p8023630).

Since the argument for (and against) rolling ladders crops up on the bogleheads forum from time to time, the motivation for doing this work was to see if some definitive argument could be made either way. The results presented here tentatively suggest that there may be some relatively small gains to be found by adopting a ladder during accumulation and a fund during decumulation. However, I think a more substantial analysis is required to confirm that suggestion.

cheers
StillGoing

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I agree that there are other, perhaps more subjective, considerations - I've tried to concentrate on quantifiable factors in this thread (even the complexity is, to some extent, quantifiable).

Taking your point about transparency, I think this depends. Bond indices with a fixed maturity range (e.g., the Barclay's 1 to 5 year treasury index) are well defined and, except in extreme cases, changes in the bonds issued or their weights will not make a huge difference to the average maturity or weighted duration. Of course, the duration will change with yields. However, something like total bond market BND is entirely dependent on the underlying issues. While I note that BND is described as an intermediate term bond fund at https://investor.vanguard.com/investmen ... rofile/bnd, in reality the maturity is dependent on the weighted of maturities issued by the treasury and AAA corporates. For example in the unlikely event that the treasury decides to only issue 20 year bonds, then the weighted maturity (and duration) of BND would gradually increase over time. More realistically, the graph on page 23 of https://home.treasury.gov/system/files/ ... Q22024.pdf shows the weighted average maturity of treasuries since 1980. This has varied from about 4 years (early 1980s) to a recent peak of about 6 years and so has been 'intermediate' for at least the last 40 years. It is interesting to note that UK debt management has been far more variable with the average Macaulay duration of the UK debt (corresponding to that of the 'all stocks' index which contains all UK gilts) ranging from 25 years (in the 1930s) to a minimum of about 5 years in the 1980s and about 12 years in 2020 (Figure 6 in my paper at https://papers.ssrn.com/sol3/papers.cfm ... id=4742450 shows this variation for a variety of UK indices).

cheers
StillGoing

[StillGoing]

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I agree that if you want more transparency and control in a bond portfolio, the ladder is the way to go. However, I feel that a number (percentage unknown) adopt the ladder solely to avoid losses because of the feature of redeeming the lowest rung.

I've seen this argument about avoiding the losses too. What the (limited) analysis presented here indicates, is that any difference in performance between the ladder and the fund is likely to be small when considered over a period of years. While there is certainty about the maturing bond in a rolling ladder (it will mature at par), there is uncertainty as to what yield it will be reinvested at and hence future coupons (which, in the long term, are the main source of growth in the ladder or fund).

For those wanting to be certain about income in retirement, a rolling ladder is not the optimum tool since a non-rolling (collapsing) ladder provides more certainty.

cheers
StillGoing

[StillGoing]

I think the owner’s experiences are factors to consider too. I think there is more transparency and control in a bond ladder.

I agree that if you want more transparency and control in a bond portfolio, the ladder is the way to go. However, I feel that a number (percentage unknown) adopt the ladder solely to avoid losses because of the feature of redeeming the lowest rung.

Ease of use makes a bond fund a winner for me. Never have to stress about when a bond (or a CD) is going to mature.
Perhaps feeling in control is why a lot of people buy individual stocks. They don’t have to own the “bad” stocks that are in the indexes.

I do not understand your point about stress over when a bond (or CD) are to mature... the maturity date is exactly known in advance. However, stress about the NAV of bond funds during recent interest rate rises has been visible even on these boards.

Leaving aside corporate bonds, there are no 'bad' or 'good' treasuries only ones that have the desired properties (e.g. maturity, coupon, etc.) and their response to changes in yield mathematically certain. To a large extent, the properties of bond funds, at least expressed through their average maturity or weight duration, are also broadly known (e.g., a larger duration leads to a larger standard deviation in returns). For example, provided the relevant range of maturities is covered, the inclusion of all 116 bonds in the vanguard short term treasury fund (VFIRX) does not provide diversification beyond including perhaps only 12 of them covering maturities at intervals of 2 months (there would be a bit of tracking error). Diversification in stocks is a completely different matter and not relevant when talking about treasuries.

cheers
StillGoing

[rkhusky]

I do not understand your point about stress over when a bond (or CD) are to mature... the maturity date is exactly known in advance. However, stress about the NAV of bond funds during recent interest rate rises has been visible even on these boards.

I suppose if you have your bonds and CD’s on auto-roll and never need to withdraw, there is no stress. But if you have expenses coming due, especially unexpected ones, such that you need cash, then you need to look up when your bonds or CD’s are maturing, and if one is not maturing soon enough, calculating the cost of withdrawing before maturity. And if the cost is high, look for ways to avoid withdrawing before maturity, such as getting a short term loan from family or friend.

In addition, if you are not auto-rolling your bonds and CD’s, you have to keep track of all the maturities so that you can make a decision on how to reinvest the maturing bond or CD. And, if you are auto-rolling your bonds and CD’s, you might wonder if there are better options available, and so you still might want to keep track of current rates compared to what your auto-rolled bonds and CD’s will be paying.

If I never need to withdraw from a bond fund, then there is no stress about the NAV either, the same as for auto-rolled bonds and CD’s. And if you are auto-rolling without paying attention, you are ceding control to the issuers of the bonds and CD’s, leaving you with as much control as using a bond fund.

[FellsGuy]

OP respectfully I've glazed (due to my low intellect probably) over you've made so many posts. Are you asking a question or telling all of us something? I am interested in the topic but have lost the plot.
As a reset can you help me at least understand what your conclusion is on the topic? My take is you believe you're one step ahead of the mortals at Fidelity and Vanguard that run the bond funds but I don't know if that can be replicated by the rest of us or why we should exert the level of effort you do on this? or stated simply what is actionable?

[Tom_T]

OP respectfully I've glazed (due to my low intellect probably) over you've made so many posts. Are you asking a question or telling all of us something? I am interested in the topic but have lost the plot.
As a reset can you help me at least understand what your conclusion is on the topic? My take is you believe you're one step ahead of the mortals at Fidelity and Vanguard that run the bond funds but I don't know if that can be replicated by the rest of us or why we should exert the level of effort you do on this? or stated simply what is actionable?

My personal opinion is that there is no way to know which method will be better in the future. No theory is going to match real-life situations, and nobody can predict how the yield curve (all parts of it) will change in the future. You decide which risks you want to take with your fixed income.

[jeffyscott]

As a reset can you help me at least understand what your conclusion is on the topic?

Really The OP's conclusion seems quite clear and requires reading less than one sentence under a bolded heading a few posts back.

A few final thoughts

As others have already said it appears that the difference in performance between a rolling ladder and a bond index fund of the same maturity range is relatively small...

[FellsGuy]

As a reset can you help me at least understand what your conclusion is on the topic?

Really The OP's conclusion seems quite clear and requires reading less than one sentence under a bolded heading a few posts back.

A few final thoughts

As others have already said it appears that the difference in performance between a rolling ladder and a bond index fund of the same maturity range is relatively small...

Around the 12th post by OP on his own thread I confess I started to lose the trail but thank you very much for pointing it out No criticism meant I'm interested but when the jargon-meter redlines I'm lost. I'm very simple on the bonds Ibonds, tbills, FXNAX, SGOV buy them forget them or have them on autopilot so yes "really" haha

[gavinsiu]

I agree with jeffyscott. The difference between a rolling bond ladder and bond fund is essentially single digit basis points. The rest are just people (including myself) commenting and justifying their positions.

[Doc]

I agree with jeffyscott. The difference between a rolling bond ladder and bond fund is essentially single digit basis points. The rest are just people (including myself) commenting and justifying their positions.

A bond fund and a rolling ladder have two different sets of holdings. The bond fund buys bonds with different initial maturities. Say fives, sevens and tens for a 3-10 fund. So when the five it purchased reaches three years remaining it is sold and it buys a new five. The 3-10 ladder buys only the bond with the longest maturity. So when one of it's positions reaches the lower limit (three years) it buys another TEN not another five.

This results in the two of them having different weighted average maturities. Which one of them has the highest return depends on the shape of the yield curve and how it changes over the time period. Seven years in this example.

With a "normal" rising rate yield curve the ladder yields more because it has a higher average maturity. Whether it's high enough to incur the added work involved and or the decision as to what to sell if you need to withdraw money is another matter. Given today's inverted yield curve the fund "wins" because it's buying some "high" yielding fives instead of a lower yielding ten.

(In a taxable account there may also be some tax benefits to the ladder because of cap gains vs. interest tradeoffs.)

[rkhusky]

Many of the intermediate bond funds seem to have maturities from 1 - 20 years.

[StillGoing]

OP respectfully I've glazed (due to my low intellect probably) over you've made so many posts. Are you asking a question or telling all of us something? I am interested in the topic but have lost the plot.
As a reset can you help me at least understand what your conclusion is on the topic? My take is you believe you're one step ahead of the mortals at Fidelity and Vanguard that run the bond funds but I don't know if that can be replicated by the rest of us or why we should exert the level of effort you do on this? or stated simply what is actionable?

A discussion of the merits (or otherwise) of a bond ladder compared to a bond fund comes up on these boards periodically with, as we see in this thread, strong opinions on either side. The modelling I've done here was an attempt to actually quantify the difference between them - i.e., an attempt to actually use evidence rather than opinion (so I have no 'beliefs' in this area). As it turns out the evidence suggests is that the difference between a fund and a rolling ladder appears to be small. FWIW, my opinion is therefore that holding either is largely a matter of preference and not one of outperformance.

For full disclosure, our portfolio contains a collapsing inflation linked gilt ladder (UK equivalent to TIPS) to provide inflation protected income during a period before our state pensions (social security in US terms) are paid, a one year rolling CD 'ladder' (two one year accounts maturing 6 months apart timed to match our portfolio withdrawals), and two bond index funds (one short and one intermediate duration). So I'm a fairly conventional boglehead in that respect.

cheers
StillGoing

[FellsGuy]

OP respectfully I've glazed (due to my low intellect probably) over you've made so many posts. Are you asking a question or telling all of us something? I am interested in the topic but have lost the plot.
As a reset can you help me at least understand what your conclusion is on the topic? My take is you believe you're one step ahead of the mortals at Fidelity and Vanguard that run the bond funds but I don't know if that can be replicated by the rest of us or why we should exert the level of effort you do on this? or stated simply what is actionable?

A discussion of the merits (or otherwise) of a bond ladder compared to a bond fund comes up on these boards periodically with, as we see in this thread, strong opinions on either side. The modelling I've done here was an attempt to actually quantify the difference between them - i.e., an attempt to actually use evidence rather than opinion (so I have no 'beliefs' in this area). As it turns out the evidence suggests is that the difference between a fund and a rolling ladder appears to be small. FWIW, my opinion is therefore that holding either is largely a matter of preference and not one of outperformance.

For full disclosure, our portfolio contains a collapsing inflation linked gilt ladder (UK equivalent to TIPS) to provide inflation protected income during a period before our state pensions (social security in US terms) are paid, a one year rolling CD 'ladder' (two one year accounts maturing 6 months apart timed to match our portfolio withdrawals), and two bond index funds (one short and one intermediate duration). So I'm a fairly conventional boglehead in that respect.

cheers
StillGoing

Thank you SG I get it now especially with the context…

[dbr]

OP respectfully I've glazed (due to my low intellect probably) over you've made so many posts. Are you asking a question or telling all of us something? I am interested in the topic but have lost the plot.
As a reset can you help me at least understand what your conclusion is on the topic? My take is you believe you're one step ahead of the mortals at Fidelity and Vanguard that run the bond funds but I don't know if that can be replicated by the rest of us or why we should exert the level of effort you do on this? or stated simply what is actionable?

A discussion of the merits (or otherwise) of a bond ladder compared to a bond fund comes up on these boards periodically with, as we see in this thread, strong opinions on either side. The modelling I've done here was an attempt to actually quantify the difference between them - i.e., an attempt to actually use evidence rather than opinion (so I have no 'beliefs' in this area). As it turns out the evidence suggests is that the difference between a fund and a rolling ladder appears to be small. FWIW, my opinion is therefore that holding either is largely a matter of preference and not one of outperformance.

For full disclosure, our portfolio contains a collapsing inflation linked gilt ladder (UK equivalent to TIPS) to provide inflation protected income during a period before our state pensions (social security in US terms) are paid, a one year rolling CD 'ladder' (two one year accounts maturing 6 months apart timed to match our portfolio withdrawals), and two bond index funds (one short and one intermediate duration). So I'm a fairly conventional boglehead in that respect.

cheers
StillGoing

Your analysis is a helpful contribution with respect to return of the two choices.

I have the impression many investors strongly preferring a rolling ladder are not interested in whether the return is better or not but are interested in the fact that they never sell a bond for anything except redemption at maturity of the promised face value known in advance. This sentiment is strongly driven recently by the fact that bond fund values fell a lot when interest rates went up a lot but the face value of individual bonds for redemption at maturity did not change.

I don't know where to put that criterion as a measure of performance in the usual terms of interest to investors.

[rkhusky]

I have the impression many investors strongly preferring a rolling ladder are not interested in whether the return is better or not but are interested in the fact that they never sell a bond for anything except redemption at maturity of the promised face value known in advance. This sentiment is strongly driven recently by the fact that bond fund values fell a lot when interest rates went up a lot but the face value of individual bonds for redemption at maturity did not change.

I don't know where to put that criterion as a measure of performance in the usual terms of interest to investors.

It’s a matter of personal preference, similar to the percentage to hold in bonds vs stocks. But it’s good to recognize the tradeoffs one makes in these decisions - expected return vs volatility vs determinability vs ease of use.