Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

This is not a new idea or evaluation but I nonetheless thought this backtest was interesting.

https://www.portfoliovisualizer.com/bac ... tion5_3=40

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

My issue with back testing like this is it's highly subjective on the start date...

I randomly selected a 2003 and 2009 start date, the first were much closer returns between all 3, the latter had the best returns from total bond.

I have no idea what the future will look like...
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

I agree but it is not just about return, also downside protection. Even if you start in 2003 or 2009, the portfolio with treasuries had superior risk measures and Sharpe ratio.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Using your link, I also tested monthly DCA which shows a similar result.

How is it possible that there was no positive premium for the corporate bonds in total bond for that long? A period with no mass corporate bankruptcies and plenty of monetary and fiscal support? It's almost like the concept of risk premiums is just something made up by someone trying to model irrational behavior, but who has a mental block making it impossible for them to believe that it is in fact irrational??

OTOH I find it quite a hard sell to place an entire fixed income allocation in a single issuer, no matter how much screeching about the impossibility of sovereign default.

Of course the performance of TIPS since inception - 2019 may be an even bigger question for the risk premium hypothesis.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

If there actually were a risk of default by the US treasury, I think it is unlikely that US corporate bonds or US mortgages would save you. Those are priced by a risk premium on top of the yield of treasuries.

One theory is that there is an asymmetry of information with corporate bonds. When a company issues corporate bonds, they have more information about the company than the investors who buy them, and also can use that information to decide when to issue etc. Another theory is that liquidity issues with corporate bonds sometimes get in the way of market efficiency. Another is that a bond index has to sell a bond if it falls slightly below investment grade, then buy it back if the bond recovers back to investment grade. This constitutes selling low and buying back higher.

I hold corporate bonds, but as the largest class of assets held in a stable value fund that I use. I don’t think it is a mistake to hold a total bond index, but it is not the holy grail either, and likely is not the optimal mix of bond subclasses.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Northern Flicker wrote: Sat May 14, 2022 12:18 am If there actually were a risk of default by the US treasury, I think it is unlikely that US corporate bonds or US mortgages would save you. Those are priced by a risk premium on top of the yield of treasuries.
This is a classic case of just what I was talking about above. You are trying to force your model onto reality. Securities are priced by supply and demand for them. Models may help further understanding of how things work under "normal" circumstances similar to when the model was created, but they are not the final arbiter of prices, real trades are. After all, corporates and mortgages didn't go up in smoke during previous or other sovereign defaults.
One theory is that there is an asymmetry of information with corporate bonds. When a company issues corporate bonds, they have more information about the company than the investors who buy them, and also can use that information to decide when to issue etc. Another theory is that liquidity issues with corporate bonds sometimes get in the way of market efficiency. Another is that a bond index has to sell a bond if it falls slightly below investment grade, then buy it back if the bond recovers back to investment grade. This constitutes selling low and buying back higher.
All of these are valid points to some degree, yet most active corporate active bond funds underperform. Even Vanguard's have not been consistent outperformers.

Another possible reason is natural demand by insurers and such.

One might also wonder to what degree a bond market can even operate as a market with the presence of one large player with great liquidity.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Most active bond funds underperform because of costs. Models of a nature of what I referred to are how institutional traders who drive market movements establish value of a bond, which guides their trading that establishes supply and demand.

It is very difficult to know why a market behaves as it does. My issue with total bond is that I give up control over the allocation weights to bond subclasses. Low cost is no longer as much of a motivation to hold a total bond index when treasury index funds are available for 3-4 bp from Fidelity, Vanguard, and probably iShares.
Last edited by Northern Flicker on Sat May 14, 2022 11:05 pm, edited 3 times in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

Northern Flicker wrote: Sat May 14, 2022 12:18 am but it is not the holy grail either, and likely is not the optimal mix of bond subclasses.
Has anyone said that a total bond is a "holy grail" or the "optimal bond" fund?

Admittedly, I'm not a bond expert. But I don't think I've ever seen someone claim total bond as being either...

Instead, what I've seen is lots of debate on bonds. And usually the total bond gets picked apart. But much as in your data, I look and shrug. I don't see a massive difference. Sure, under certain conditions others might do better (or worse), or have different (maybe better) risk characteristics. But the total bond fund seems to be "good enough".

Which reminds me of the qoute from Bogle. I can't remember if it was specific to the "3 fund" model, or more likely the general idea of a total stock market index fund. I likely have the quote wrong, but it was something like "it may not be the best portfolio, but it's better then an infinite number of alternatives".

Having said that, my fixed income is split across a mix of total bond, long term treasuries, state & national municipal bonds, and I & EE Savings Bonds. Oddly, I have far more complexity in my "fixed income" portion than my stocks. But in the grand scheme of things, I doubt I'll see a notable difference versus had I simply stayed with total bond. For those that want to keep things simple, total bond is a great option.
Tom_T
Posts: 2545
Joined: Wed Aug 29, 2007 2:33 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Tom_T »

SnowBog wrote: Sat May 14, 2022 2:22 am
Northern Flicker wrote: Sat May 14, 2022 12:18 am but it is not the holy grail either, and likely is not the optimal mix of bond subclasses.
Has anyone said that a total bond is a "holy grail" or the "optimal bond" fund?

Admittedly, I'm not a bond expert. But I don't think I've ever seen someone claim total bond as being either...
Oh, anyone who asks about switching from Total Bond to some other fixed-income alternative gets plenty of criticism around here. "What do you know that bond managers don't know?" "That's market timing." "Do you think you can predict the future?" And so on.
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

Tom_T wrote: Sat May 14, 2022 6:23 am
SnowBog wrote: Sat May 14, 2022 2:22 am
Northern Flicker wrote: Sat May 14, 2022 12:18 am but it is not the holy grail either, and likely is not the optimal mix of bond subclasses.
Has anyone said that a total bond is a "holy grail" or the "optimal bond" fund?

Admittedly, I'm not a bond expert. But I don't think I've ever seen someone claim total bond as being either...
Oh, anyone who asks about switching from Total Bond to some other fixed-income alternative gets plenty of criticism around here. "What do you know that bond managers don't know?" "That's market timing." "Do you think you can predict the future?" And so on.
Sure... Some people get - for lack of a better term - "dogmatic" about certain things.

But in one thread recently (I think unrelated to bonds), an interesting observation was shared. Often times the "pushback" that we see isn't so much about something being a "bad" idea - it's just that in the grand scheme of things - it's likely to have minimal impact. So it's more like "you could, but why would you..."

And I guess I think of alternatives to total bond the same way. I'm not convinced they will make a significant difference.

Even the original link only showed a 0.17% better CAGR (which was actually a lower CAGR if you pick different start dates) and a 0.04 better Sharpe ratio. I doubt either of those is going to have a material impact on someone's retirement...

Granted - some of this comes down to personality. Some people like to squeeze every ounce of performance they can out of something. I won't fault them. And I'm probably part way there myself, as noted above, my bonds are split across multiple funds.

But for new investors, those looking to simplify their portfolio, and for those who would be tempted to keep "tweaking" their portfolio chasing a better return, I think a total bond fund will work great for them. Keep it simple, stay the course.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

The max drawdown of -29.45% with treasuries vs. -32.17% with total bond is what I find most significant, not the difference in return. A 3-fund portfolio with total bond is no simpler than a 3-fund portfolio with a treasury bond index fund. In a financial crisis bigger and deeper than what we’ve seen in the last 25 years, treasuries are likely to provide tangibly better downside protection than total bond.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
alluringreality
Posts: 693
Joined: Tue Nov 12, 2019 10:59 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by alluringreality »

Northern Flicker wrote: Sat May 14, 2022 2:35 pm treasuries are likely to provide tangibly better downside protection than total bond.
David Swensen took a similar position in Unconventional Success when comparing treasuries to corporates and mortgage-backed securities.
30% Savings Bonds, 45% US Indexes, 25% Ex-US Indexes
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

Northern Flicker wrote: Sat May 14, 2022 2:35 pm The max drawdown of -29.45% with treasuries vs. -32.17% with total bond is what I find most significant, not the difference in return. A 3-fund portfolio with total bond is no simpler than a 3-fund portfolio with a treasury bond index fund. In a financial crisis bigger and deeper than what we’ve seen in the last 25 years, treasuries are likely to provide tangibly better downside protection than total bond.
But again, highly dependent on start date. Using my random pick of 2009 from before, the max downdraw for treasuries (portfolio 1) was -12.27% - or 0.24% more than total bond (portfolio 3) which was -12.03%.

I'd feel differently if the results held up consistently. But when I see these back tests, and I randomly pick a few other years to compare the results, and I see different results - it makes me question how much was "luck" (so to speak) based on the start date - versus a "real difference" between the assets. I honestly don't know...
Angst
Posts: 2908
Joined: Sat Jun 09, 2007 11:31 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Angst »

Well, instead of those bond funds, if you plug in Vanguard's Interm-Term Bond Index VBIIX (aka BIV), its CAGR at 7.41% exceeds them all, albeit with higher volatility than the others. Personally, I'm still fairly wedded to the notion that Treasuries complement equity the best when it matters most.
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

Angst wrote: Sat May 14, 2022 3:25 pm Well, instead of those bond funds, if you plug in Vanguard's Interm-Term Bond Index VBIIX (aka BIV), its CAGR at 7.41% exceeds them all, albeit with higher volatility than the others. Personally, I'm still fairly wedded to the notion that Treasuries complement equity best when it matters most.
FWIW with the original start date, and my randomly selected 2003 and 2009, VBIIX had a consistently better CAGR but also a consistently higher (worse) max downdraw compared against the original portfolios # 1 + 3.

Personally, i personally wouldn't try to tell someone they need to pick one over the other - not for these minor differences.

But I would tell people that have already invested in one - what's the point in changing? As it's likely not a one time change for them... When they see the next "best" thing, they'll likely change again. And again... That much "tweaking" and performance chasing is potentially hazardous to their long term goals.
User avatar
Doc
Posts: 10304
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Doc »

Northern Flicker wrote: Fri May 13, 2022 11:53 pm This is not a new idea or evaluation but I nonetheless thought this backtest was interesting.

https://www.portfoliovisualizer.com/bac ... tion5_3=40

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
Why do so many posters try to make their "which bond" decisions by looking only at fixed income?

My portfolio and I believe the vast majority of Bogleheads, have both fixed income and equities in their portfolio. I think how Treasuries, Total Bond and GNMAs interact with those equities is far more important in determining which to use then how much better or worse one fixed income is compared to another fixed income type.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Yes. Overall portfolio diversification matters more than bond portfolio diversification.

I have come to view a portfolio as treasuries and everything else, where the treasuries are for safety and liquidity, and everything else is to drive a positive portfolio real return with reasonable overall portfolio risk.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
KlangFool
Posts: 24992
Joined: Sat Oct 11, 2008 12:35 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by KlangFool »

Northern Flicker wrote: Fri May 13, 2022 11:53 pm

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
Northern Flicker,

In order to be optimal, we need the ability to predict the future. I choose "good enough" since I cannot predict the future.

KlangFool
40% VWENX | 12.5% VFWAX/VTIAX | 11.5% VTSAX | 16% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 40% Wellington 40% 3-funds 20% Mini-Larry
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

KlangFool wrote: Sat May 14, 2022 5:29 pm
Northern Flicker wrote: Fri May 13, 2022 11:53 pm

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
Northern Flicker,

In order to be optimal, we need the ability to predict the future. I choose "good enough" since I cannot predict the future.

KlangFool
That is not correct. You do not have to predict the outcome of an event to develop an optimal strategy stochastically. There are probabilistic games with provably optimal strategies for instance.

Of course, we will not have a provably optimal investment strategy, but I think there is reasonable evidence that treasuries provide better downside protection than total bond without giving up measurable expected return.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
SnowBog
Posts: 3083
Joined: Fri Dec 21, 2018 11:21 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by SnowBog »

Northern Flicker wrote: Sat May 14, 2022 5:38 pm
KlangFool wrote: Sat May 14, 2022 5:29 pm
Northern Flicker wrote: Fri May 13, 2022 11:53 pm

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
Northern Flicker,

In order to be optimal, we need the ability to predict the future. I choose "good enough" since I cannot predict the future.

KlangFool
That is not correct. You do not have to predict the outcome of an event to develop an optimal strategy stochastically. There are probabilistic games with provably optimal strategies for instance.

Of course, we will not have a provably optimal investment strategy, but I think there is reasonable evidence that treasuries provide better downside protection than total bond without giving up measurable expected return.
Learned a new word today... stochastically :beer

While I see no compelling reason for myself - or others - holding Total Bond funds to switch...

I see no reason why you'd be harmed in switching (at least assuming this is a one-time strategic change - and not something you'll keep tweaking as you see different random results with the data moving forward).

I think both could equally be said to fall into the "good enough" category. I don't think either - especially on their own - is going to contribute a meaningful difference over the long haul - the difference is likely in the "noise".
KlangFool
Posts: 24992
Joined: Sat Oct 11, 2008 12:35 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by KlangFool »

Northern Flicker wrote: Sat May 14, 2022 5:38 pm
KlangFool wrote: Sat May 14, 2022 5:29 pm
Northern Flicker wrote: Fri May 13, 2022 11:53 pm

The gaps between treasuries and total bond with respect to return and risk measures over 26 years now seems to be enough to seriously question if a total bond index fund is optimal. For grins, I also included the much maligned GNMAs in the test.
Northern Flicker,

In order to be optimal, we need the ability to predict the future. I choose "good enough" since I cannot predict the future.

KlangFool
That is not correct. You do not have to predict the outcome of an event to develop an optimal strategy stochastically. There are probabilistic games with provably optimal strategies for instance.

Of course, we will not have a provably optimal investment strategy, but I think there is reasonable evidence that treasuries provide better downside protection than total bond without giving up measurable expected return.
Northern Flicker,

A) You are looking at the fixed income return. I am looking at the portfolio return.

B) We do not get something for nothing. With better downside protection, We lose higher interest rate. Will the tradeoff work out to be better in the future? My crystal ball is hazy.

C) For absolute safety, I keep cash and physical gold and silver. I do not need my bond to do that.

KlangFool
40% VWENX | 12.5% VFWAX/VTIAX | 11.5% VTSAX | 16% VBTLX | 10% VSIAX/VTMSX/VSMAX | 10% VSIGX| 40% Wellington 40% 3-funds 20% Mini-Larry
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

KlangFool wrote: Sat May 14, 2022 6:05 pm
A) You are looking at the fixed income return. I am looking at the portfolio return.

B) We do not get something for nothing. With better downside protection, We lose higher interest rate. Will the tradeoff work out to be better in the future? My crystal ball is hazy.
I think the opposite— using the yield difference between total bond and treasuries is looking at bond yield in isolation instead of looking at overall portfolio risk and return, which is what I was looking at. It is known from modern portfolio theory that when assets are not fully correlated, portfolio risk is not additive from the risk of underlying components.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
phoroner
Posts: 6
Joined: Thu Jul 09, 2020 11:58 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by phoroner »

My take is first agreement that it is the behavior of the total portfolio that should be considered, and the stock to fixed-income ratio is primary driver of risk. For most portfolios, as long as high-quality investment-grade bonds are used and costs are low, it shouldn’t make a big difference. For someone with a high stock allocation, Treasurys alone may make the most sense, but for lower stock allocations a Total Bond fund may make the most sense. For me, because I tilt my equities towards small and value, which are expected to do especially poorly during bad times, I only use Treasurys for my bond holdings in the expectation that will do best at that time. If I want to adjust risk and return I will change my stock-to-bond ratio. If I only held market capitalization weighted stock funds at my current AA, I would probable use a TBM fund and stick to a 3-fund portfolio. Interesting to hear how others approach this.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

SnowBog wrote: Sat May 14, 2022 2:59 pm
Northern Flicker wrote: Sat May 14, 2022 2:35 pm The max drawdown of -29.45% with treasuries vs. -32.17% with total bond is what I find most significant, not the difference in return. A 3-fund portfolio with total bond is no simpler than a 3-fund portfolio with a treasury bond index fund. In a financial crisis bigger and deeper than what we’ve seen in the last 25 years, treasuries are likely to provide tangibly better downside protection than total bond.
But again, highly dependent on start date. Using my random pick of 2009 from before, the max downdraw for treasuries (portfolio 1) was -12.27% - or 0.24% more than total bond (portfolio 3) which was -12.03%.
Max drawdown is a different thing from return with respect to bias in a backtest. Sure, relative return may vary for different start dates. But when you shorten the test to start to 2009 and look at max drawdown, you are excluding the two deepest bear markets since the Great Depression from the set of drawdowns to select the max from.

The evidence from the 2007-2009 global financial crisis and Great Depression is that corporate bonds have significantly more downside tail risk than treasuries. The fact that portfolio return is unlikely to be much different basically says that the extra tail risk protection from treasuries most likely is a bit of a free lunch in a 60/40 portfolio.
Last edited by Northern Flicker on Sun May 15, 2022 1:39 am, edited 1 time in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Northern Flicker wrote: Sat May 14, 2022 2:18 am My issue with total bond is that I give up control over the allocation weights to bond subclasses.
Indeed. One might also decide to use a different strategy in the one than the other, like indexing treasuries and using an active fund for other areas.

I think my biggest issue with total bond is the forced panic selling of bonds that drop into junk space. One might consider IUSB as an alternative which includes junk, but IIUC bonds of an entity undergoing bankruptcy might not be liquid enough to be held in this fund either and may need to be sold?
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Another point is that the market assumes that the govt will bail out FNMA and FMAC and does not seem to discount in an additional credit risk premium relative to GNMAs that have a contractual backing of the US Treasury for credit failure. The prospectus for FNMAs and FMACs explicitly states that they are not backed by the full faith and credit of the US Treasury.

In the direst of crises with courts and Congress sorting out a mess and deciding what will and won’t be made whole, this would be an important distinction. Practical credit risk of FNMAs, FMACs, and GNMAs may be the same, but credit tail risk is not equivalent. The total bond index is about 17% FNMAs and FMACs, and about 8% GNMAs.

Despite all of this, I don’t consider it to be an error to hold a total bond index fund. And many investors will do well to have a single, low cost high quality bond fund. With FUAMX at a cost of 3 bp/yr and VGIT at 4 bp/yr, a single fund, a low cost, intermediate treasury portfolio is also available, and worthy of consideration for a single fund bond portfolio. It also has advantages if TIPS are also to be held because it avoids concern about the moderately high historical correlation of TIPS and corporate bonds.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Any thoughts on why have GNMAs not dropped as much as total bond or treasuries (slightly) this year? I thought GNMAs were supposed to be even more problematic during a time of rising rates, but there is also a component to the steepness of the move, right??

Over its short life so far, IUSB has had a higher return and sharpe ratio and lower max drawdown than BND in spite of including the maligned junk bonds.

https://www.portfoliovisualizer.com/bac ... ion2_2=100
MrJedi
Posts: 2036
Joined: Wed May 06, 2020 11:42 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by MrJedi »

I'm in the camp that the primary purpose of bond is safety / risk management, so Treasury only is what I choose. But I have no problem with total bond and would choose it without Treasury only options, like what is commonly seen in qualified plans (including my own 401k).
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

000 wrote: Sun May 15, 2022 2:19 am Any thoughts on why have GNMAs not dropped as much as total bond or treasuries (slightly) this year? I thought GNMAs were supposed to be even more problematic during a time of rising rates, but there is also a component to the steepness of the move, right??
My guess is that active management decisions to reduce extension risk and/or duration were in play for funds like the Vanguard GNMA fund, but the GNMA index also seems to have overperformed:

https://www.portfoliovisualizer.com/bac ... ion3_3=100

The convexity issue with MBS often is described deterministically, even by some well respected professionals. The convexity of an MBS is a random entity. MBS are amortizing bonds. Each monthly payment includes some principal prepayment at par. If rates rise so that the bond is under par, the monthly principal repayment leads to a faster turnover and reinvestment of bond principal than would occur for a treasury where the principal is repaid at maturity. Home sales and foreclosures also lead to a GNMA being pre-paid in full at par even if rates have risen.

So there is a tug of war between principal repayment at par and the negative convexity associated with extension risk of an MBS. If rates rise at a sufficiently moderated pace, the principal repayment at par will be the larger effect of the two. When rates rise sharply enough, i.e. with enough velocity, duration extension will be the dominant effect of the two.

Seasoned GNMAs, with mortgages well along to completed repayment, obviously have less convexity risk— they have less remaining term to extend into relative to a newly issued one, and principal repayments are occurring at a faster velocity. This provides an active management opportunity to adjust the extension risk of a GNMA portfolio.

Prepayment risk also is typically modeled for active management. Properties of individual GNMAs and macroeconomic factors may be used to estimate the expected prepayments of individual GNMAs.

Prepayment risk and extension risk trade off to a significant extent, so that optimizing for one can or often will increase the exposure of a portfolio to the other.

GNMAs generally show up prominently in the results of portfolio optimizations, so I consider their more significant risks to be more in the category of tail risk.

https://www.portfoliovisualizer.com/opt ... ints=false

But conditional value-at-risk (tail risk) models may favor GNMAs even more:

https://www.portfoliovisualizer.com/opt ... ints=false

That of course may not use data from an inflationary period in the tail risk calculation.

The late Dave Swensen, formerly the CIO of the Yale endowment and author of a couple of very good books on investing, considered the optionality and extension of MBS to be sufficiently complex to model that mis-pricing risk becomes an issue. He felt this also applied to modeling the call risk of corporate bonds, and considered that individual investors should just avoid them all and stick to treasuries and TIPS. We don’t have to agree with his position on mis-pricing risk to nonetheless consider that his recommendation of limiting one’s bonds to treasuries and TIPS leads to the design of excellent bond portfolios while sidestepping the complexities of MBS and credit.
Last edited by Northern Flicker on Sun May 15, 2022 2:47 pm, edited 1 time in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
User avatar
Doc
Posts: 10304
Joined: Sat Feb 24, 2007 1:10 pm
Location: Two left turns from Larry

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Doc »

Northern Flicker wrote: Sun May 15, 2022 2:03 pm The problem is that the convexity issue with MBS often is described deterministically, even by some well respected professionals.
I'm not sure what that means.

But: "... because the expected maturity of an MBS depends to a great degree on the current level of interest rate relative to the level at the time the mortgages backing the MBS were ordinated, an MBS can exhibit negative convexity - duration actually shortens when rates fall." A respected professional by the name of Larry Swedroe. "The Only Guide to a Winning Bond Strategy You'll Ever Need: The Way Smart Money Preserves Wealth Today", Swedroe and Hempen, 2006

We can do all the back testing we want but the pros are doing the same thing and they are better at it than we are.

We can make a general observation however. If bond type "A" has a higher yield than type "B" than "A" must have more risk of some type.
That's all we novices need to know.

Also we should not be comparing bond types without looking at the rest of our portfolio. What I belive gets lost in these "which bond" threads is what happens when equities are tanking. It doesn't take a lot of calculation to know that Treasuries do better than MBS or corporates in that situation. And since we can't predict with any certainty what stocks are going to do we should have a fixed income plan that covers us no matter what. The answer is to have all three bonds types held separately so you make a choice of which to sell when the time is ripe. If you like the idea of TBM fine. Just break it up into its components so at least one of those components will be a better rebalancer when the need arises.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Doc wrote: Sun May 15, 2022 2:46 pm
Northern Flicker wrote: Sun May 15, 2022 2:03 pm The problem is that the convexity issue with MBS often is described deterministically, even by some well respected professionals.
I'm not sure what that means.
That’s why the remainder of the post explained what it means.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
BigJohn
Posts: 2236
Joined: Wed Apr 02, 2014 11:27 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by BigJohn »

All arguments about the validity of back testing aside, it’s unclear to me that max drawdown is the best measuring stick right now. With high inflation and the uncertainty on when/if the Fed will get it under control, the purchasing power loss even with no drawdown seems a bigger worry. Why not use TIPS instead of IT treasuries or GNMAS and eliminate that worry?
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Sure, TIPS and I bonds can play an important role. A retirement saver would bias more to nominal bonds than a retiree. Job loss during disinflation is a bigger risk for a retirement saver than inflation, especially if wages keep up with the inflation rate. Obviously job loss during inflation would be the worst of both.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
BigJohn
Posts: 2236
Joined: Wed Apr 02, 2014 11:27 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by BigJohn »

Certainly not impossible or unheard of but there are not many negative numbers (ie deflation) in this table. Even when it happens, other than the Great Depression, those numbers are pretty small compared to the inflation numbers. Risk that needs to be managed seems skewed to unexpected high inflation from my perspective. Although I agree, that risk is lower when you are young and employed.

https://www.usinflationcalculator.com/ ... ion-rates/
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
Drew31
Posts: 350
Joined: Mon Oct 22, 2012 1:18 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Drew31 »

Northern Flicker wrote: Sun May 15, 2022 3:05 pm Sure, TIPS and I bonds can play an important role. A retirement saver would bias more to nominal bonds than a retiree. Job loss during disinflation is a bigger risk for a retirement saver than inflation, especially if wages keep up with the inflation rate. Obviously job loss during inflation would be the worst of both.
So let's a assume a retirement saver disregards and goes 100% TIPS or iBonds. So the risk here is a disflationary event w/out a job loss where a nominal bond would provide a greater return vs. TIPS? Just trying to understand this as where my head has started to turn is why wouldn't I use 100% TIPS to lock in my known real return. (I do get why a retiree w/out income this is more important, but why a retirement saver couldn't benefit from this as well)
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

BigJohn wrote: Sun May 15, 2022 3:32 pm Certainly not impossible or unheard of but there are not many negative numbers (ie deflation) in this table. Even when it happens, other than the Great Depression, those numbers are pretty small compared to the inflation numbers. Risk that needs to be managed seems skewed to unexpected high inflation from my perspective. Although I agree, that risk is lower when you are young and employed.

https://www.usinflationcalculator.com/ ... ion-rates/
You don’t have to have deflation for TIPS to sell off sharply. Disinflation (slowing inflation) and elevated deflation risk is sufficient.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

Drew31 wrote: Sun May 15, 2022 4:06 pm
Northern Flicker wrote: Sun May 15, 2022 3:05 pm Sure, TIPS and I bonds can play an important role. A retirement saver would bias more to nominal bonds than a retiree. Job loss during disinflation is a bigger risk for a retirement saver than inflation, especially if wages keep up with the inflation rate. Obviously job loss during inflation would be the worst of both.
So let's a assume a retirement saver disregards and goes 100% TIPS or iBonds. So the risk here is a disflationary event w/out a job loss where a nominal bond would provide a greater return vs. TIPS? Just trying to understand this as where my head has started to turn is why wouldn't I use 100% TIPS to lock in my known real return. (I do get why a retiree w/out income this is more important, but why a retirement saver couldn't benefit from this as well)
I bonds have no deflation risk. In fact, deflation increases their real return.

Here is a comparison of a 60/40 portfolio with treasuries, TIPS, and a mix of the two. The use of TIPS provided significantly less downside protection to a retirement saver who experienced job loss at the depth of the deepest downturn in the backtest.

https://www.portfoliovisualizer.com/bac ... tion3_3=20

Regardless of what happened to prices, a deeper drawdown still results less assets to rely on to pay the prices. TIPS have a higher equity correlation, which makes sense if equities also protect against inflation in their own way.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Broken Man 1999
Posts: 7170
Joined: Wed Apr 08, 2015 11:31 am
Location: West coast of Florida, near Champa Bay !

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Broken Man 1999 »

I prefer Treasuries, myself. We hold Short-Term, Intermediate-Term, and Long-Term at the present.

We still have a bit of Vanguard Total Bond Index in our Transamerica VA, I had been using it to rebalance into equities, though I haven't needed to rebalance for a while. We are down to around 9% of bond portfolio with Total Bond Index, all other bonds are treasuries and I-Bonds.

Over time I have been increasing our equities from 50% equities/50% bonds in 2015 to 55% equities/45% bonds in 2020, aiming for a final AA of 60% equities/40% bonds in 2025. In 2025 DW and I will be 72, ten years down the retirement road from 2015, when I added my lump-sum pension to my TIRA and we started distributions from our TIRAs.

Given our rising equity %, I thought it made sense to use treasuries for the most part.

Broken Man 1999
“If I cannot drink Bourbon and smoke cigars in Heaven then I shall not go." - Mark Twain
BigJohn
Posts: 2236
Joined: Wed Apr 02, 2014 11:27 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by BigJohn »

Northern Flicker wrote: Sun May 15, 2022 5:06 pm
BigJohn wrote: Sun May 15, 2022 3:32 pm Certainly not impossible or unheard of but there are not many negative numbers (ie deflation) in this table. Even when it happens, other than the Great Depression, those numbers are pretty small compared to the inflation numbers. Risk that needs to be managed seems skewed to unexpected high inflation from my perspective. Although I agree, that risk is lower when you are young and employed.

https://www.usinflationcalculator.com/ ... ion-rates/
You don’t have to have deflation for TIPS to sell off sharply. Disinflation (slowing inflation) and elevated deflation risk is sufficient.
Agree with this but only as a short term liquidity issue. Beyond that that are inextricably linked to nominal treasuries of the same duration by the break even inflation rate but without the risk of high unexpected inflation that nominals have. However, everyone has different ways of weighing these risks so many will choose nominals.
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
User avatar
Taylor Larimore
Advisory Board
Posts: 31271
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Taylor Larimore »

Bogleheads:

The figures below show that the Total Bond Market Index did just fine since 1976.

Past rate of inflation (CPU) and fund returns in The Three-Fund Portfolio:

YEAR--INFLATION--BOND INDEX--S&P 500 INDEX------MSCI EAFE INDEX
1976-------4.9%--------15.6%------------23.8%--------------------3.6%
1977-------6.7-----------3.0-------------(-7.2)-------------------17.5
1978-------9.0-----------1.4---------------6.6--------------------33.1
1979------13.3-----------1.9--------------18.4-------------------10.9 (Highest Annual Inflation Rate)
1980------12.5-----------2.7--------------32.4-------------------25.4
1981-------8.9-----------6.3-------------(-4.9)------------------(-2.5)
1982-------3.8----------32.6--------------21.6------------------(-0.3) (Highest Bond Index Return)
1983-------3.8-----------8.4--------------22.6-------------------24.8
1984-------3.9----------15.2---------------6.3--------------------3.5
1985-------3.8----------22.1--------------31.7-------------------51.4
1986-------1.1----------15.2--------------18.7-------------------65.8 (Vanguard Total Bond Market Inception )
1987-------4.4-----------2.8----------------5.2-------------------24.6
1988-------4.4-----------7.9---------------16.6-------------------27.8
1989-------4.6----------14.5---------------31.7------------------11.4
1990-------6.1-----------8.9---------------(-3.1)---------------(-22.8)
1991-------3.1----------16.0---------------30.5------------------12.4
1992-------2.9-----------7.4-----------------7.6----------------(-11.9) (Vanguard Total Stock Market Inception)
1993-------2.7-----------9.7----------------10.1------------------32.6
1994-------2.7---------(-2.9)----------------1.3--------------------7.6 (Lowest Bond Index Return)
1995-------2.5----------18.5---------------37.6-------------------11.8 (Highest S&P Index Return)
1996-------3.3-----------3.6----------------23.0--------------------7.2 (Vanguard Total International Stock Market Inception)
1997-------1.7-----------9.7----------------33.4--------------------2.6
1998-------1.6-----------8.7----------------28.6-------------------19.1
1999-------2.7---------(-0.8)---------------21.0-------------------28.3
2000-------3.4----------11.6---------------(-9.1)----------------(-15.8)
2001-------1.6-----------8.4--------------(-11.9)----------------(-19.8)
2002-------2.4----------10.3-------------(-22.1)----------------(-15.3)
2003-------1.9-----------4.1----------------28.7-------------------40.4
2004-------3.3-----------4.3----------------10.9-------------------20.9
2005-------3.4-----------2.4-----------------4.9-------------------15.8
2006-------2.5-----------4.3----------------15.8------------------26.8
2007-------4.1-----------7.0-----------------5.5------------------11.6
2008-------0.1-----------5.2--------------(-37.0)---------------(-43.1) (Lowest U.S. and International Stock Returns)
2009-------2.7-----------5.9----------------26.5------------------32.5
2010-------1.5-----------6.5----------------15.1-------------------8.2
2011-------3.0-----------7.7-----------------2.1----------------(-11.7)
2012-------1.7-----------4.3----------------16.0------------------17.9
2013-------1.5---------(-2.0)---------------32.4------------------23.3
2014-------1.6-----------6.0----------------13.7-----------------(-4.5)
2015-------0.7-----------0.5-----------------1.4-----------------(-0.4)
2016-------2.1-----------2.6----------------12.0-------------------1.5
2017-------2.1-----------3.5----------------21.8------------------25.6
2018-------2.5---------(-0.1)--------------(-4.4)---------------(-13.4)
2019-------2.3-----------8.7----------------31.5------------------22.7
2020-------1.4-----------7.7----------------18.4------------------11.3
2021-------7.0---------(-1.7)---------------25.7-------------------8.6

Sources: Vanguard, U.S. Labor Department (CPI-U), Standard & Poors, Bloomberg Barclays Aggregate Bond Index, and DFTurner

Past performance does not guarantee future performance.

Best wishes.
Taylor
Jack Bogle's Words of Wisdom: "There may be better investment strategies than owning just three broad-based index funds but the number of strategies that are worse is infinite."
"Simplicity is the master key to financial success." -- Jack Bogle
User avatar
nisiprius
Advisory Board
Posts: 46543
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by nisiprius »

Sure. Now plug in PTTRX (PIMCO Total Return) and the portfolio using PTTRX does better than the one with VFIIX. In Sharpe ratio as well as return.

There's probably a bond fund out there that would give higher return and Sharpe ratio than PTTRX?

Maybe Total Bond was not "optimal" over that time period. I don't care. I don't do "optimal," I do "good enough." I picked something reasonable, its performance has been decent, and I'm sticking with it. It's not as if the other bond funds haven't sagged year-to-date, too.

Source

Image
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Northern Flicker wrote: Sun May 15, 2022 2:03 pm GNMAs generally show up prominently in the results of portfolio optimizations, so I consider their more significant risks to be more in the category of tail risk.

https://www.portfoliovisualizer.com/opt ... ints=false
That is interesting.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

BigJohn wrote: Sun May 15, 2022 6:09 pm
Northern Flicker wrote: Sun May 15, 2022 5:06 pm
BigJohn wrote: Sun May 15, 2022 3:32 pm Certainly not impossible or unheard of but there are not many negative numbers (ie deflation) in this table. Even when it happens, other than the Great Depression, those numbers are pretty small compared to the inflation numbers. Risk that needs to be managed seems skewed to unexpected high inflation from my perspective. Although I agree, that risk is lower when you are young and employed.

https://www.usinflationcalculator.com/ ... ion-rates/
You don’t have to have deflation for TIPS to sell off sharply. Disinflation (slowing inflation) and elevated deflation risk is sufficient.
Agree with this but only as a short term liquidity issue. Beyond that that are inextricably linked to nominal treasuries of the same duration by the break even inflation rate but without the risk of high unexpected inflation that nominals have. However, everyone has different ways of weighing these risks so many will choose nominals.
I don’t think it is just short-term liquidity issues that are in play. There is no bound in either direction on the TIPS breakeven inflation rate, so there is no bound on the gaps that can develop between TIPS and nominal treasuries either way. In any case, the thread was about nominal bonds. Treasuries are also the better nominal bonds to combine with a TIPS portfolio due to various correlations. TIPS are better aligned with a retiree’s liabilities. Treasuries are better at diversifying equity risk.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

nisiprius wrote: Sun May 15, 2022 6:51 pm Sure. Now plug in PTTRX (PIMCO Total Return) and the portfolio using PTTRX does better than the one with VFIIX. In Sharpe ratio as well as return.
The max drawdown was higher with PTTRX. My main point was not to recommend GNMAs but to point out that using treasuries to diversify equity generally has resulted in lower max drawdowns than with total bond (or with GNMAs for that matter) without giving up any measurable expected return relative to total bond. This effect in the past has been a result of correlations lowering volatility and max drawdown, not return characteristics.

There are two ways to look at indexing. One is to say there is an imperative to hold every asset class at market cap weight. Another is to view indexing as a cost-effective and efficient way to hold the asset classes one chooses to hold. I subscribe more to the latter view.

For the record, we do hold corporate bonds and commercial mortgages through a stable value fund, so have not used a treasury-only portfolio. Our current asset allocation calls for about equal allotments of intermediate treasuries and a stable value fund for fixed income assets. When the stable value fund was a higher percentage of our fixed income, it was combined with GNMAs and short-term TIPS.
Last edited by Northern Flicker on Mon May 16, 2022 2:31 am, edited 1 time in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
BigJohn
Posts: 2236
Joined: Wed Apr 02, 2014 11:27 pm

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by BigJohn »

Northern Flicker wrote: Sun May 15, 2022 7:55 pm
BigJohn wrote: Sun May 15, 2022 6:09 pm
Northern Flicker wrote: Sun May 15, 2022 5:06 pm
BigJohn wrote: Sun May 15, 2022 3:32 pm Certainly not impossible or unheard of but there are not many negative numbers (ie deflation) in this table. Even when it happens, other than the Great Depression, those numbers are pretty small compared to the inflation numbers. Risk that needs to be managed seems skewed to unexpected high inflation from my perspective. Although I agree, that risk is lower when you are young and employed.

https://www.usinflationcalculator.com/ ... ion-rates/
You don’t have to have deflation for TIPS to sell off sharply. Disinflation (slowing inflation) and elevated deflation risk is sufficient.
Agree with this but only as a short term liquidity issue. Beyond that that are inextricably linked to nominal treasuries of the same duration by the break even inflation rate but without the risk of high unexpected inflation that nominals have. However, everyone has different ways of weighing these risks so many will choose nominals.
I don’t think it is just short-term liquidity issues that are in play. There is no bound in either direction on the TIPS breakeven inflation rate, so there is no bound on the gaps that can develop between TIPS and nominal treasuries either way. In any case, the thread was about nominal bonds. Treasuries are also the better nominal bonds to combine with a TIPS portfolio due to various correlations. TIPS are better aligned with a retiree’s liabilities. Treasuries are better at diversifying equity risk.
My apologies, didn’t mean to derail your discussion by bringing in TIPS so we’ll just agree to disagree and leave it at that.
"The greatest enemy of a good plan is the dream of a perfect plan" - Carl Von Clausewitz
McQ
Posts: 379
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by McQ »

This post supplies additional historical data to address questions of the form, Why was the incremental performance of total bond so low in Northern Flicker’s analysis? Why didn’t the inclusion of corporate bonds, with their risk premium, add more return relative to Treasuries? Several iterations of such questions, asked of the OP, can be found upthread.

Short answer: because the corporate bond premium has been smaller than many investors have been wont to believe.

Caveat: the data that follow concern top-grade bonds, rated Aaa or Aa, with scattered A. The premium may have been/would be larger for bonds which straddle the investment grade boundary (i.e., Baa & Ba)—if the data were available, which alas they are not, at least, not far enough back, for Ba.

Why do investors have inflated expectations for the corporate bond premium? Because the corporate bond data in the Stocks Bonds Bills & Inflation yearbook are problematic before 1973, and just plain screwy before 1946 (i.e., show smallish drawdowns during the early 1930s).

I assembled an alternative long corporate bond index to correct these issues. Details are in this paper: “When Do Corporate Bond Investors Earn a Premium for Bearing Risk? A Test Spanning the Great Depression of the 1930s,” at https://papers.ssrn.com/sol3/papers.cfm ... id=3740190.

I assessed the earned premium (not the yield premium, but rather, total return) from 1909 to 2019, comparing long corporate bonds to long Treasuries/governments. After 1973, my new data is spliced to the SBBI series, which I perceive to be no longer problematic because it matches well with corresponding Lehman Brothers indexes (results of a cross check against same are in the paper).

Here is the chart:

Image

An easy summary: corporate bonds generally earn a small premium of one or two dozen basis points over corresponding Treasuries. But when the brown smelly stuff hits the whirling blades, that premium vanishes—even if measured over the trailing 100 years.

More generally, there is a corporate bond premium if you accept averaging over many decades. But a more accurate description would be that the corporate bond premium cycles through regimes, sometimes running positive for a while, sometime negative.

For this chart I created ten-year rolls, starting with the ten years from 1909 to 1919. The annualized return on corporate bonds is either greater or less than the realized return on government bonds, premium or deficit, as shown by up or down bars.

Image

Sometimes corporate bonds outperform government bonds, sometimes not (only long bonds are considered in these tests—a bond was removed once its maturity dropped below 15 years).

Fun factoid to motivate you to download the paper: in the 1920s, the bottom of the investment grade set was Ba, not Baa bonds. The familiar category of “investment grade bonds” was created in the mid-1930s by New Deal bank regulation. Regulators drew a line at Baa, and told banks they’d be put on the hot seat if they owned anything lower. The NAIC and predecessors laid down the same rule for insurance companies.

Second factoid: outside the Depression, and before the late 1960s, xx% of large bonds traded on the NYSE were investment grade (where x is a number like ‘8’). Within the same time boundaries the proportion of corporate bonds rated Aaa, Aa, or A was a yy%, where y is a number like ‘7’ or ‘8’, depending on decade.

Compare the rating distribution for a Vanguard corporate bond fund today, and you will see how much has changed.
They that read the footnotes, they shall be saved; but they that pass over the appendices, they shall wander forever.
000
Posts: 7631
Joined: Thu Jul 23, 2020 12:04 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by 000 »

Very interesting, McQ. I wonder if there was a higher premium in previous centuries. From reading, I have gotten the general understanding that corporate bonds were how many serious long term investors made money pre-1900 when stocks were viewed more negatively. But perhaps that understanding is tainted by the survivorship bias of those who did not suffer serious defaults on their bonds.
Topic Author
Northern Flicker
Posts: 10434
Joined: Fri Apr 10, 2015 12:29 am

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by Northern Flicker »

I still consider the lower correlation of treasuries with equities relative to the correlation of corporate bonds with equities to be a key issue.

Here is a comparison of treasuries and corporate bonds. Risk was rewarded for the corporate bond fund during the period if volatility is your risk measure (though max drawdown may still be a different story):

https://www.portfoliovisualizer.com/bac ... ion2_2=100

But if we package the same two bond funds with a U.S. equity index fund each in a 60/40 portfolio, risk was not rewarded for the portfolio using the corporate bond fund by either measure over the same period:

https://www.portfoliovisualizer.com/bac ... tion3_2=40

I also am skeptical of indexing for MBS portfolios. Two virtually identical MBS at issue can have very different outcomes based on the timing of individual home sales or other prepayments for houses financed with mortgages in the securities. This creates challenges for tracking an index. Imagine the treasury issuing two 5-year notes in the same auction and them having a different return. Or imagine that a company does a bond issue, and selectively calls in some of the bonds in the issue, but not other identical bonds that were part of the issue.

The actively managed VFIJX has generally beaten the GNMA index by 15-20 bp/yr (though not in 2022 so far) and the index ETF GNMA has generally trailed the GNMA index it tracks by 15-20 bp/yr.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
McQ
Posts: 379
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by McQ »

Northern Flicker wrote: Mon May 16, 2022 1:03 am I still consider the lower correlation of treasuries with equities relative to the correlation of corporate bonds with equities to be a key issue.
...
To your point about correlations, Northern Flicker, here is a paper by Kizer et al. that backs up your concerns: https://papers.ssrn.com/sol3/papers.cfm ... id=3147005
They that read the footnotes, they shall be saved; but they that pass over the appendices, they shall wander forever.
McQ
Posts: 379
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Total Bond vs IT Treasuries vs GNMAs in a 3-fund portfolio

Post by McQ »

000 wrote: Sun May 15, 2022 11:28 pm Very interesting, McQ. I wonder if there was a higher premium in previous centuries. From reading, I have gotten the general understanding that corporate bonds were how many serious long term investors made money pre-1900 when stocks were viewed more negatively. But perhaps that understanding is tainted by the survivorship bias of those who did not suffer serious defaults on their bonds.
Yes and no, 000. The US bond market was a very different place before 1918, but not in precisely the way you suggested. It wasn’t that return-maximizing investors bought bonds instead of stocks; bonds were bought then, as now, for a stable income at minimum risk.

The difference was the near-complete absence of Treasury securities. That market had gone moribund after the Civil War refinancing was completed circa 1879. Not much trading, not much liquidity, not much face value available to trade, and unattractive rates (Treasuries were locked up in bank vaults to gain circulation privileges). It’s not really possible to compute a meaningful yield spread in this era.

There were municipal bonds then, but these had the same liquidity problems, etc. Railroad bonds were available in the tens of millions, were blue chip securities, paid a market rate of interest, and were essentially, the only fixed income game in town until Liberty bonds debuted in 1917, after which the modern Treasury market gradually began to take shape (an evolution not completed until the 1970s, BTW).

PS re defaults: except for the 1890s, and later the 1930s, defaults on blue chip railroad bonds were rare, and there was generally a work out where new securities equal to the par value eventually got paid out to the bondholders, albeit with a gap in income and some loss of security. "Blue chip" is the constraint here: if these were carpet bagger Southern railroad bonds created out of thin air through legislative bribery after the Civil War, these mostly went bust after 1873. And if it was the Erie railroad, you could be in for a rough ride, but no one would call "the Scarlet Woman of Wall Street" a blue-chip security ...
They that read the footnotes, they shall be saved; but they that pass over the appendices, they shall wander forever.
Post Reply