Can you do better than BND, Part 2: Test across bear and bull markets

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

In the first phase of this thread, I compared the performance of the Total Bond fund, as a component of the 60/40 portfolio, to various alternative bond funds. I used mutual fund data only, from 1992 to 2023: viewtopic.php?t=424247

The answer to the thread title was “yes, you could have done better than BND.”

Ah, but an objection was raised, as follows:
“Excepting 2022, your entire sample period falls within one of the great bond bull markets of all time. Would the test outcomes, particularly the strong showing by long Treasuries, come out the same in a period where interest rates were rising rather than falling?”

Good question.

This follow-up thread attempts an answer.

Primer on bond market history

If you’ve read Sydney Homer’s History of Interest Rates, you’ll be aware that the bond market tends to proceed through long waves of bull and bear moves. That’s another way of saying that interest rates, when they go up, tend to rise (irregularly) for a long, long time; and when interest rates fall, they can drift down and down for decades.

The only such wave that most investors today have personally experienced (Taylor is an exception) is the recently concluded bull move that ran from 1982 to 2020.

You’d have to be Taylor’s age at least, or as well read as Bill Bernstein, to remember the preceding great bond bear market of 1946 – 1981, as rates moved up, slowly at first, from 2% to 15%.

And only historians like me have a good sense of the preceding bond bull market, which ran from 1921 to 1945, or the bear market before that, which ran from 1900 to 1920, or the great bull move that began after the Civil War and ran to 1900.

Ideally, we would test BND against alternatives across all five cycles, but alas, the data aren’t there.

What we can do is construct a matched test between the 1946-1981 bear market and the following 36 years through 2017, covering most of the 1982 – 2020 bull market. Two adjacent 36-year periods, one where rates rose and rose, and another where they fell and fell.

Data

1. Mutual fund data from 1992 to 2017 from Simba (VOO, BND, VGIT, VGLT and their predecessor mutual funds)
2. Earlier data as follows:

a) stocks: S&P 500 index from the SBBI

b) Total bond: Bloomberg Aggregate for 1973 – 1991, and then my newly constructed total bond index from 1946 through 1972*

*See paper for a description. It combines corporate and government, short, medium and long bonds: https://papers.ssrn.com/sol3/papers.cfm ... id=3947293

c) intermediate Treasuries: the 5-year series from the SBBI

d) long Treasuries: the approximately 20-year maturity Treasury series from the SBBI.

e) T-bills: the 90 day (3 month) series from FRED.

Test

I look first at the performance of a 60/40 portfolio using Total Bond, versus the alternatives in various logical mixes, such as 60% stocks, 30% long Treasuries, and 10% T-bills (this 60/30/10 mix was one of the best performers in the earlier post-1992 thread). This example is a logical alternative because it reflects the old Bogle/Vanguard nostrum that “investors should hold a mix of stocks, bonds, and cash consistent with their risk tolerance.” I’ll also look at just holding 40% in long Treasuries—a level of duration risk that most here on the forum would eschew (=Greek for “run screaming in the other direction”).

And of course, there is the intermediate Treasury case, offering approximately the same protection against duration risk as BND, but with the safety of Treasuries, but minus the spread from owning riskier credits.

Note carefully: outside the TIPS ladder/liability matching case, very few people here on the forum, or in the larger world, expect to hold only bonds in their portfolio. Most of us, especially those over 50, will hold a mix of stocks and bonds (and who knows what else.)

The question that titles the thread, then, is to be understood as, “Can you find a different fixed income asset, other than BND, that better supports the risk asset portion of the portfolio (=stocks)?”

And: has that alternative fixed income asset offered superior performance over bear as well as bull markets?

Just as an aside, stocks performed well in both sub-periods, returning 10.3% during the bond bear market through 1981 and 11.8% during the bond bull market through 2017. The 60/40 portfolio won’t have as high a return as that all stock portfolio but may be much less volatile.

In short: this test is in the spirit of Markowitz, where a portfolio must be evaluated in terms of its volatility (standard deviation) as well as its return, and where what matters is portfolio outcomes, not the characteristics of the individual assets making up the portfolio.

Bond bear market performance

The annually rebalanced 60/40 portfolio, stocks and total bond, returned 7.76% compounded, turning $10,000 into $147,400, at a volatility “cost” of an 11.14% standard deviation.

A 100% allocation to stocks would have compounded at 10.23%, at a cost (SD) of 17.4%.

The 60/40 portfolio with total bond can be termed a success: you got 76% of the return of a 100% stock portfolio, for 65% as much volatility.

But can you do better?

Yes, an intermediate (borderline short) Treasury investment was an even better foil to stocks: $161,200 in wealth, with compounded return of 8.03% and volatility cost 90 bp lower than the BND alternative, at an SD of 10.24%.

As you may recall, IT also beat BND in the post-1992 results; it will again best BND in the post-1981 results shared below; and it beat BND in my theoretical analysis (Markowitzian diversification Part 2:viewtopic.php?t=398529).

It is important to understand why. The bonds in Total Bond, many of them long bonds, and many of these corporate bonds, will generally offer a higher yield than a 5-year Treasury. From the standpoint of maximizing wealth It would seem logical to own Total Bond instead of IT.

But to succumb to such a narrow view of how to maximize utility, to reach for those tempting corporate bond yields, is to make the fundamental Markowitzian error: looking at individual assets in isolation rather than in terms of what they contribute to the portfolio. In this 1946 – 1981 era Total Bond was correlated .36 with stocks; 5-year Treasuries were correlated at negative -.21. That’s huge.

You got more Markowitzian diversification with IT, with diversification defined here narrowly as more zig against stock zag, thus cancelling out more volatility, and thus producing a higher compounded return on the 60/40 mix that includes IT rather than BND.

Both have an intermediate duration, but one is less correlated with stocks and a better diversifier. That’s the first takeaway.

Other alternatives

Consider next a 60/30/10 mix, stocks/long Treasuries/T-bills. The big advantage here is to always have cash on hand for unexpected withdrawals. Many affluent Bogleheads, and professional couples more generally, will accumulate 401(k) balances of $1 million, $2 million or more by the time they reach their late 50s. Having $100,000 or $200,000 in cash makes for a nice security blanket indeed. Talk about ballast against stock risk!

But oh that duration risk from LT … and in a long-running bear market to boot.

You may be surprised to learn how well the 60/30/10 mix fared: a small shortfall in return relative to 60/40 with BND (7.55% vs. 7.76%), but a noticeably lower volatility cost (10.41% vs. 11.14%).

Even the 60/40 LT mix had somewhat lower volatility than the BND mix (10.55% vs 11.14%), at the cost of a more noticeable drop in return, to 7.32%.

The explanation again comes back to correlation: negative -.07 for LT versus +.36 for BND.

Maybe … duration is not the key metric when you are MIXING bonds with stocks. Maybe … long Treasuries can deliver such a large amount of zig relative to stocks’ zag as to be a better diversifier than the superficially attractive Total Bond option, with its safe-seeming intermediate duration, and the sweetener of the corporate yield spread.

But to accept high duration bonds as a diversifier to stocks you must accept a Markowitzian definition of diversification: “a reduction in portfolio volatility such that portfolio compounded return increases, relative to holding the two assets separately without rebalancing.”

If instead you demand that every component of your portfolio other than stocks always have a positive return--especially your "safe" assets--then you will underperform the tough-minded Markowitzian investor. Don't think that's a thing?--agony when your bonds go down over the short-term?--then you need to read more widely on the forum. Start about April 2022 ...

Efficient Frontiers

I’ve focused on the 60/40 mix in both threads thus far, but now it is time to relax that constraint. With annual returns of each asset in the spreadsheet, it is easy enough to look at all blends between 100% stocks / 0% bonds and 0% stocks and 100% bonds.

As before, I begin with the total bond mix.

Image

The dotted line is a linear interpolation of returns. To the extent the blue line bulges up and to the left, some diversification benefit has been gained from holding a fixed weight allocation and rebalancing back to the target weights each year. Total bond, given its correlation far under 1.0, does provide some benefit.

Next, let’s take a look at blends with intermediate Treasuries.

Image

As expected from the 60/40 point estimate, the orange IT return-risk line is uniformly above and to the left of the total bond line. I’ve put a red square around the 60/40 mixes. The advantage of using intermediate Treasuries in place of BND grows as you move left to consider more conservative portfolios. If your risk tolerance puts you at the Wellesley or Vanguard Retirement Income level (e.g., 35% - 30% stocks), you did WAY better with intermediate Treasuries than with BND: dozens of bp more in return, with over 120 bp less in risk.

At 30/70 stocks/IT you got 60% of the stock return for 30% of the stock risk. That’s diversification!

In fact, in this era, it was per se irrational for the conservative IT owner not to add an allocation of at least 20% to stocks—note the bulge out to the left, such that a with-stock allocation produces more return, for no greater risk, than a 100% IT bond allocation.

In short: do not be misled by the fool’s gold of the corporate spread. It’s correlation that matters. Returns come from owning stock, risk reduction comes from bonds, and ceteris paribus, the lower the correlation of those bonds with the stock portfolio, the greater the impact with respect to risk reduction. Go, intermediate Treasuries!

Now here is the 60/30/10 mix with long Treasuries and T-bills.

Image

Surprisingly not bad, might be a fair judgment on the long Treasury mix. Returns are somewhat lower—this is the Great Bond Bear Market of 1946-1981 after all—but the risk reduction is great enough to be visible on the chart as a leftward shift relative to the total bond mix.

You can obtain the liquidity of T-bills, without much of a hit to return, if you are willing also to take on the duration risk of long Treasuries. And you didn’t do that much worse than the omnibus total bond option, where you cannot reach in and just liquidate the short end of the holdings should an emergency arise.

Action takeaways

These are deferred until after the bull case is reviewed.

Next up: bull market performance of BND and the alternatives
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
gavinsiu
Posts: 4693
Joined: Sun Nov 14, 2021 11:42 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by gavinsiu »

Thanks for the analysis and the conclusion are too different than part 1. I have some questions.
  • Based on what you are saying, you are saying Treasury are better than Total Bond probably due to its lower correlation and probably because corporate bonds don't bring additional return in a total portfolio sense?
  • Does the correlation play a factor in bond asset selection and how much difference before it matters. For example, if you find something that has a very similar return to Treasury but lower correlation, the that might be even better?
  • You are saying that a combination of 60/30/10 Stock/Long Treasury/cash is even better than 60/40 stock/IT? By the way, does LT means treasury bond that are 10 years or does it mean 30 years?
  • What would happen if you drop the duration to short bond?
  • Will you ever do anything with TIPS in a future analysis?
Thanks again for the education.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

gavinsiu wrote: Mon Mar 25, 2024 11:01 pm Thanks for the analysis and the conclusion are too different than part 1. I have some questions.
  • Based on what you are saying, you are saying Treasury are better than Total Bond probably due to its lower correlation and probably because corporate bonds don't bring additional return in a total portfolio sense?
  • Does the correlation play a factor in bond asset selection and how much difference before it matters. For example, if you find something that has a very similar return to Treasury but lower correlation, the that might be even better?
  • You are saying that a combination of 60/30/10 Stock/Long Treasury/cash is even better than 60/40 stock/IT? By the way, does LT means treasury bond that are 10 years or does it mean 30 years?
  • What would happen if you drop the duration to short bond?
  • Will you ever do anything with TIPS in a future analysis?
Thanks again for the education.
Re your questions by number:
1. Mostly lower correlation, which swamps the realized yield spread on corporates.

2. Ceteris paribus, the lower the correlation, the greater the diversification benefit. But we are talking basis points. The corporate bonds in total bond COULD have had a high enough incremental return to overcome the higher correlation; empirically, they did not.

3. Nope, note that the orange IT line is higher, here, as expected in a bond bear market where the longer the bond, the harder the fall. LT is the 20-year series in the Stocks, Bonds, Bills & Inflation yearbook.

4. Other analyses not shown found a still shorter Treasury not to do any better than IT: too little return, no lower correlation.

5. Will be touched on a the very end, but overall "not in this thread."
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Gecko10x
Posts: 399
Joined: Thu Jun 06, 2013 8:10 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Gecko10x »

This is very informative, thank you.

It has me questioning my LTT allocation, wondering if it would be better glide down to ITT. Of course after already enduring the last couple of years of drawdowns, navigating the timing of adjusting down the duration would be tricky without locking in large losses.
Tib
Posts: 320
Joined: Tue Oct 27, 2020 9:57 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Tib »

So, intermediate Treasuries look to be a better stock complement than BND in a rebalanced portfolio. And even during bond bear-markets, McQ finds no significant advantage for short-term over intermediate-term Treasuries. All good to know, especially for those with a long investment horizon.

For a retiree focused on the next 10 years and worried mainly about inflation, this article by Rekenthaler makes a case for the preferability of cash to intermediate Treasuries. The results presented don't take into account the lower correlation of intermediate Treasuries with stocks. But during an inflationary era, perhaps that wouldn't be a significant advantage for the Treasuries. Those wanting to rebalance during 2022 into their depressed stocks would have been glad to be doing so from cash rather than from also-depressed Treasuries.

https://www.morningstar.com/alternative ... -revisited
User avatar
grabiner
Advisory Board
Posts: 35439
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by grabiner »

I would expect that the poor diversification from long-term Treasuries is caused by the different risk profiles of investors. Long-term bonds are attractive to insurance companies and pension funds, which have long-term fixed-dollar future liabilities and are thus not sensitive to inflation risk; this increases demand for those bonds, and thus reduces the risk premium for other investors.
Wiki David Grabiner
Nahtanoj
Posts: 131
Joined: Tue Apr 04, 2017 7:01 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Nahtanoj »

This nice analysis seems consistent with David Swensen’s recommendation to generally use Treasuries rather than corporate or mortgage backed bonds as a fundamental building block in a portfolio. Swensen emphasizes that — during “bad times” — non-callable, full faith and credit Treasuries perform better than mortgage backed securities and corporate bonds, which have prepayment and equity-type risks that Treasuries do not. (Those risks of mortgage backed and corporate bonds tend to materialize precisely during “bad times” for the rest of the portfolio, when you especially want your bonds to be doing well.)

Swensen also recommends using TIPS as a separate asset class alongside nominal Treasuries, because the performance of the two types of securities is driven by some fundamentally different factors (unexpected inflation vs. unexpected disinflation or deflation). Unfortunately, the TIPS performance data only go back to the late 1990s, so it’s hard to do the full analysis using historical data!
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Gecko10x wrote: Tue Mar 26, 2024 5:23 pm This is very informative, thank you.

It has me questioning my LTT allocation, wondering if it would be better glide down to ITT. Of course after already enduring the last couple of years of drawdowns, navigating the timing of adjusting down the duration would be tricky without locking in large losses.
Stay tuned--I'll be adding additional perspective on long Treasuries.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Nahtanoj wrote: Tue Mar 26, 2024 9:22 pm This nice analysis seems consistent with David Swensen’s recommendation to generally use Treasuries rather than corporate or mortgage backed bonds as a fundamental building block in a portfolio. Swensen emphasizes that — during “bad times” — non-callable, full faith and credit Treasuries perform better than mortgage backed securities and corporate bonds, which have prepayment and equity-type risks that Treasuries do not. (Those risks of mortgage backed and corporate bonds tend to materialize precisely during “bad times” for the rest of the portfolio, when you especially want your bonds to be doing well.)

Swensen also recommends using TIPS as a separate asset class alongside nominal Treasuries, because the performance of the two types of securities is driven by some fundamentally different factors (unexpected inflation vs. unexpected disinflation or deflation). Unfortunately, the TIPS performance data only go back to the late 1990s, so it’s hard to do the full analysis using historical data!
Very true, and I won't be able to do much with TIPS in this historical analysis.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Now to the bull market case ...

To review the history, in late 1981 the Great Bond Bear Market that began in 1946 reached its nadir. A furious rally began, taking the long Treasury up just over 40% in 1982. Double digit returns on long Treasuries recurred over the next decade, in part because yields had reached almost 15%. Add 5% annual capital appreciation on top of a starting yield of 15, 12, 9, even 6% and … you get a heckuva of a total return.

The Great Bond Bull Market had got underway.

Here is the Markowitz chart for the mix with Total Bond 1982 - 2017:

Image

Returns during this bull market for bonds were of course much higher. Where a 100% bond allocation 1946 – 1981 earned 3% and change, from 1982 through 2017 that same all-bond portfolio, in total bond, earned just shy of 8%.

The correlation for total bond with stocks dropped to .23 from .36; that, plus a larger standard deviation—the secret sauce of Markowitzian diversification, low correlations combined with substantial volatility—produces a more noticeable bulge up and to the left of the dashed linear interpolation line.

At 60/40, the portfolio with total bond captured 90% of the stock return for two-thirds the stock risk. Not too shabby.

Now, speeding up a bit, here is the Markowitz chart comparing intermediate Treasuries and also the 60/30/10 mix with long Treasuries.

Image

Taking the IT orange line first, it continues to be more efficient than the total bond line, shifted up and to the left, more return for less risk, EXCEPT at stock allocations less than 10%. That’s right: a 100% allocation to IT gives less return for more risk than a 100% allocation to total bond. That’s the magic of Markowitzian diversification: the inferior-on-its-own IT, because it is less correlated with stocks, produces a better portfolio outcome.

Examining next the gray line, the long Treasury/T-bill mix, we see again the importance of evaluating investment outcomes at the portfolio level rather than asset by asset. Evaluated on its own, the 100% LT/T-bill mix is dominated by total bond: 300 bp of extra risk for 50 bp of extra return. Likewise, the return on a 100% T-bill portfolio would have been just under 4% annualized, far below any of the portfolios shown on the chart.

But when a 60/30/10 mix is formed, magic results. For higher stock allocations of 50, 60, and 70%, the gray line is noticeably above both blue and orange lines, at a modest cost in risk relative to the IT line, and actually less risk than the total bond line.

For grins, I decided to assemble the bear and the bull market performance lines onto one chart.

Image

Interesting. During bear market moves, the return-risk line lengthens out and portfolio returns are depressed. That makes perfect sense: one portion of the portfolio is suffering, because bonds are in a bear market, causing the overall portfolio return to suffer.

By contrast, in a bond bull market, the entire performance line gets shifted up and also contracts, since bond returns in a bond bull market come closer to approximating stock returns.

What does not change is the fact of a diversification benefit: in both cases all the bond mixes are shifted up and to the left of a linear interpolation of returns. The bulge is actually greater in the bull market case, because standard deviation is higher when returns are higher, and, with correlation held constant at a low level, greater volatility fuels a greater degree of Markowitzian diversification.

What also does not change across bear and bull markets is the superiority of intermediate Treasuries as a diversifier, especially for allocations to stock of less than 50%. Unsurprisingly, that superiority is greater during a bond bear market.

The most surprising element in the comparison is the performance of the long Treasury/T-bill blend. Ah, maybe not so surprising—“bear” and “bull” apply in spades to long duration instruments.

The bull market case shows a beautiful nose cone for the LT mix. Under these conditions, it would be per se irrational for an investor in a 3:1 LT/T-bill mix not to add at least 50% stock: they get 200 bp extra return for no additional risk. Now that’s diversification! For the bond native, that is ... at 50-50 the native stock investor has sacrificed 130 bp in return for 670 bp in risk reduction. Good enough?

Alas, that isn’t what happens during a bond bear market.

Exploit the bull or protect against the bear, greed or fear, the eternal choice facing investors.

Don’t want to play that game? Then own enough intermediate Treasuries to meet your risk tolerance.

Preliminary conclusions

In theory, a total bond market index must be the most efficient bond portfolio, offering the best available combination of risk and return. I call that MPTOS, Modern Portfolio Theory On Steroids (front syllable rhymes with that of “empty”).

BND, as currently implemented, is NOT total—it excludes all bonds rated below Baa. More generally, it is not clear to me that individual bonds are fungible in the way that exchange-listed common stocks may well be fungible. A Treasury isn’t the same thing as a corporate bond, even an AAA-rated one.

Or maybe, the idea that a “total” [anything] fund, one that actual investors can buy in the world, must necessarily be most efficient, remains, to use a phrase from the philosopher Karl Popper, a bold conjecture. Not settled truth.

The efficiency argument for BND fails. What’s left? Okay, BND covers the total maturity range. Ah, what’s the theory there?

With these handicaps visible, BND reduces to a bet on the corporate bond spread versus the necessarily higher correlation of riskier bonds with the risk asset (stocks), when evaluated as a complement to stocks within a portfolio.

Empirically, across a matched bear-bull market pair, a 72-year span, that bet did not work out.

Sticking with an intermediate duration is good; but best to do that using only Treasuries.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
cosmos
Posts: 246
Joined: Thu Oct 16, 2008 4:11 pm
Location: Third rock from the Sun

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by cosmos »

McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
It's 106 miles to Chicago, we've got a full tank of gas, half a pack of cigarettes, it's dark... and we're wearing sunglasses. Hit it.
gips
Posts: 1781
Joined: Mon May 13, 2013 5:42 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by gips »

cosmos wrote: Tue Mar 26, 2024 11:50 pm McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
Do you feel using an IT fund instead of bnd in an 80-20 portfolio significantly impacts your risk?
bluestone
Posts: 4
Joined: Fri Sep 16, 2022 7:32 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by bluestone »

Would you then recommend VGIT over BND? Or what would you choose instead?
muffins14
Posts: 5584
Joined: Wed Oct 26, 2016 4:14 am
Location: New York

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by muffins14 »

Would you also show the LT-only lines rather than 3:1 long-term : bills for the bull case?
Crom laughs at your Four Winds
rkhusky
Posts: 17992
Joined: Thu Aug 18, 2011 8:09 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by rkhusky »

One thing that is difficult to separate out in these historical analyses is the effects from costs, technology and participation. In earlier years, it was more expensive and cumbersome to invest and to analyze choices, and not many people participated. Today, it is super simple and inexpensive to invest in a wide variety of products, and to pull up performance charts and investing analyses, and most everyone is at least dabbling in investing.

So, it’s not clear if one should expect the next bull or bear bond market to look like these past ones. Is the narrowing of Treasury vs Total Bond performance due to bear vs bull or to the different investing eras?

Edit: a couple possible follow-ups. Compare the first half of the recent bull with the last half. Compare with a pure corporate index.
Last edited by rkhusky on Wed Mar 27, 2024 8:58 am, edited 1 time in total.
goaties
Posts: 544
Joined: Fri Jan 29, 2010 3:15 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by goaties »

I've long questioned the validity of an index fund for bonds. The bond universe is vast, varied, and they keep inventing new flavors all the time (does BND have MYGAs?). Conversely, indexing makes sense for the relatively limited universe of stocks.

Settling on one type of bond, intermediate Treasuries, does make a lot of sense. It is the quintessential bond, if you will. Maybe that's why it's so successful as a diversifier?
User avatar
BrooklynInvest
Posts: 1211
Joined: Sun Jul 28, 2013 9:23 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by BrooklynInvest »

This is really interesting - thank you.

As we go from what happened to what we should do, how do you think the short end of the curve is impacted by eventual falling rates and a return to a normal-shaped curve? In theory S-D benefits less from that drop, but I'm no expert on what's already priced into the market.

With hindsight I should have definitely added S-D to my BND before 2022 - most of the yield/less of the decline, but past performance and all that....?
User avatar
Indyhou
Posts: 103
Joined: Tue Nov 01, 2016 12:15 pm
Location: Texas

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Indyhou »

Very interesting thread. Thanks for posting.

One question/comment- These numbers seem to be nominal. Have you looked at what the real returns would look like?


The mean would shift down by the average inflation rate over the period, but the variance/standard deviation would be different.

This would also give a better understanding of how introducing TIPS/TIPS ladder into the mix- even though we don't have very much historical data on them, it would give prospective if current TIPS rates would have improved the portfolio had they been available.
User avatar
cosmos
Posts: 246
Joined: Thu Oct 16, 2008 4:11 pm
Location: Third rock from the Sun

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by cosmos »

gips wrote: Wed Mar 27, 2024 2:24 am
cosmos wrote: Tue Mar 26, 2024 11:50 pm McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
Do you feel using an IT fund instead of bnd in an 80-20 portfolio significantly impacts your risk?
Impacts it how? As in greater risk?? no, just the opposite. the roller coaster ride that BND can have is not what I want for the ballast portion of the portfolio. I also just do not want corporate or reit swaps or anything else in there. My risk is in the us and international stock index funds.

I may switch to ST treasuries or likely just a TIPS ladder once I hit my 60s.
It's 106 miles to Chicago, we've got a full tank of gas, half a pack of cigarettes, it's dark... and we're wearing sunglasses. Hit it.
User avatar
anagram
Posts: 1778
Joined: Fri Aug 04, 2023 1:03 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by anagram »

BrooklynInvest wrote: Wed Mar 27, 2024 8:41 am This is really interesting - thank you.

As we go from what happened to what we should do, how do you think the short end of the curve is impacted by eventual falling rates and a return to a normal-shaped curve? In theory S-D benefits less from that drop, but I'm no expert on what's already priced into the market.

With hindsight I should have definitely added S-D to my BND before 2022 - most of the yield/less of the decline, but past performance and all that....?
What is S-D? Another abbreviation that does not seem to be used on BH after searching.
User avatar
BrooklynInvest
Posts: 1211
Joined: Sun Jul 28, 2013 9:23 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by BrooklynInvest »

anagram wrote: Wed Mar 27, 2024 2:52 pm
BrooklynInvest wrote: Wed Mar 27, 2024 8:41 am This is really interesting - thank you.

As we go from what happened to what we should do, how do you think the short end of the curve is impacted by eventual falling rates and a return to a normal-shaped curve? In theory S-D benefits less from that drop, but I'm no expert on what's already priced into the market.

With hindsight I should have definitely added S-D to my BND before 2022 - most of the yield/less of the decline, but past performance and all that....?
What is S-D? Another abbreviation that does not seem to be used on BH after searching.
Short duration.
jbowman
Posts: 222
Joined: Sat Jan 09, 2021 4:34 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by jbowman »

Thanks for the interesting and insightful analysis. Will definitely be considering the implications when I next review my portfolio.
Kidmoe
Posts: 5
Joined: Sun Mar 08, 2020 7:57 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Kidmoe »

Be curious to see how a 60/40 mix with TBills did especially in the 47-81 time frame.
HomeStretch
Posts: 11621
Joined: Thu Dec 27, 2018 2:06 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by HomeStretch »

Thank you for the analysis and conclusions.

If IT are preferable to BND, what does that mean for an investor in early retirement who holds a BH 3-fund 60/40 portfolio using BND? With interest rates expected to decline, should that investor exchange BND for IT today?
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

cosmos wrote: Tue Mar 26, 2024 11:50 pm McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
Thanks for those kind words. The sentiment you express against BND is fairly widely shared on the forum. My goal in the analysis was to go beyond sentiment and to show that objectively, under a Markowitz rule, total bond has been an inferior diversifier to intermediate Treasuries, across bear and bull market cases.

Or, as I put it: “do not pursue the fool’s gold of the corporate yield spread.”

Also, a reminder: after 1992 the “intermediate Treasury” is a fund holding bonds from 3 – 10 years of maturity, consistent with the modern understanding of “intermediate.” Before 1992, and thus throughout the bear market case, it is a 5-year Treasury held for one year and then rolled into a new 5-year Treasury, per the SBBI; thus, rather less duration than the post-1992 fund.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

bluestone wrote: Wed Mar 27, 2024 6:19 am Would you then recommend VGIT over BND? Or what would you choose instead?
Yep. VGIT substituting for BND in a good Boglehead's implementation of the 3-fund portfolio.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

muffins14 wrote: Wed Mar 27, 2024 6:56 am Would you also show the LT-only lines rather than 3:1 long-term : bills for the bull case?
Here you are:

Image

Removing the tempering influence of the 10% allocation to T-bills has the expected effect: in the bear market case, long Treasuries alone are inferior to total bond. That’s the outcome that makes forum members hesitant to invest in bonds with too much duration.

In the bull market case, there is again an accentuation, with the nosecone so marked that a bond-only investor would be per se irrational if they didn’t allocate 75% to stocks: 200 bp extra return for no more risk. Alternatively, a 40% allocation to stocks would have cut risk 300 bp and raised return 150 bp, rather than holding just long Treasuries. That’s the “free lunch” of diversification along with two drink tickets and a hotel room thrown in to boot.

More generally, when we speak of “bear” and “bull” markets in bonds, the cleanest purest case will always be the longest duration instrument available: in this data, a 20-year Treasury (regular issuance of 30-year bonds did not occur until the mid-1970s).

I almost hesitate to show the bull case for LT only; I’m not sure even my children will live long enough to see those halcyon results recur.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Indyhou wrote: Wed Mar 27, 2024 10:49 am Very interesting thread. Thanks for posting.

One question/comment- These numbers seem to be nominal. Have you looked at what the real returns would look like?


The mean would shift down by the average inflation rate over the period, but the variance/standard deviation would be different.

This would also give a better understanding of how introducing TIPS/TIPS ladder into the mix- even though we don't have very much historical data on them, it would give prospective if current TIPS rates would have improved the portfolio had they been available.
That’s an excellent suggestion, Indyhou. It will take me a bit to implement, but now you’ve got me curious.

I’ll go on record with my predictions, and then I’ll compare these to the facts once available.

1. Inflation averaged 4.7% in the bear case and 2.7% in the bull case, so all curves will move down the chart, the bear set to a greater degree; in fact, some of these bear curves will stretch into negative territory once inflation-adjusted (because inflation is often what makes a bond bear market a bear market).

2. Standard deviations may not shrink much; if inflation were a constant, subtracting it from the mean wouldn’t change SD at all. Inflation is NOT a constant, so, may be more or less correlated with one series or another, producing difficult-to-intuit changes in SD of the series and their combination into portfolios. We’ll see.

3. The rank order among the curves shouldn’t change, at a guess; IT will remain the best diversifier.

But again: we’ll see! Coming in a day or two.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

HomeStretch wrote: Wed Mar 27, 2024 10:41 pm Thank you for the analysis and conclusions.

If IT are preferable to BND, what does that mean for an investor in early retirement who holds a BH 3-fund 60/40 portfolio using BND? With interest rates expected to decline, should that investor exchange BND for IT today?
That one's above my pay grade. Interest rates are expected to decline--maybe; and who knows for how long?

So, a rigorous thinker, fully convinced by the analysis, would immediately swap out BND for VGIT. And not look back.

An ordinary human being? Already committed to BND, and prone to regret on multiple fronts? Eh, as I said, above my pay grade.

I'm not licensed or qualified to give investment advice to individuals ...
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
dkturner
Posts: 1959
Joined: Sun Feb 25, 2007 6:58 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by dkturner »

Kidmoe wrote: Wed Mar 27, 2024 8:30 pm Be curious to see how a 60/40 mix with TBills did especially in the 47-81 time frame.
For the period 1946-1981 a portfolio consisting of 60% S&P 500 and 40% Treasury Bills had an annualized nominal return of 7.8% and an inflation adjusted annualized return of 3.1%.
Valuethinker
Posts: 49129
Joined: Fri May 11, 2007 11:07 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Valuethinker »

McQ wrote: Tue Mar 26, 2024 10:24 pm

Preliminary conclusions

In theory, a total bond market index must be the most efficient bond portfolio, offering the best available combination of risk and return. I call that MPTOS, Modern Portfolio Theory On Steroids (front syllable rhymes with that of “empty”).

BND, as currently implemented, is NOT total—it excludes all bonds rated below Baa. More generally, it is not clear to me that individual bonds are fungible in the way that exchange-listed common stocks may well be fungible. A Treasury isn’t the same thing as a corporate bond, even an AAA-rated one.

Or maybe, the idea that a “total” [anything] fund, one that actual investors can buy in the world, must necessarily be most efficient, remains, to use a phrase from the philosopher Karl Popper, a bold conjecture. Not settled truth.

The efficiency argument for BND fails. What’s left? Okay, BND covers the total maturity range. Ah, what’s the theory there?

With these handicaps visible, BND reduces to a bet on the corporate bond spread versus the necessarily higher correlation of riskier bonds with the risk asset (stocks), when evaluated as a complement to stocks within a portfolio.

Empirically, across a matched bear-bull market pair, a 72-year span, that bet did not work out.

Sticking with an intermediate duration is good; but best to do that using only Treasuries.
Also US Agency bonds have negative convexity arising from mortgage borrowers refinancing or extending their mortgage term v what was priced into the security at issue. In plainer English their duration shortens when interest rates are falling, and increases when they are rising. That's not helpful as an investor.

Corporate bonds often have call risk. I believe BND is technically actively managed and tries to avoid such issues.

Another problem with BND is shared by all IG bonds. When they are downgraded below BBB/ Baa2, the fund manager will sell the bonds because they have only limited ability (some fixed period) to hold onto bonds rated below BBB- ie junk. There's an "air pocket" around such bonds, which other active investors can exploit, due to this institutional limitation.

It's observable that the credit cycle moves with the economy. Thus at the peak of the economic cycle, bond funds tend to buy corporate bonds of lower credit quality. They may not realise they are doing this, because credit rating is a lagging indicator - rating agencies are not quick to downgrade their own clients. Thus the credit quality is pro cyclical with the economy, and that increases the risk for IG bond funds.
GAAP
Posts: 2646
Joined: Fri Apr 08, 2016 12:41 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by GAAP »

McQ wrote: Tue Mar 26, 2024 10:24 pm For grins, I decided to assemble the bear and the bull market performance lines onto one chart.

Image

<snip>

The most surprising element in the comparison is the performance of the long Treasury/T-bill blend. Ah, maybe not so surprising—“bear” and “bull” apply in spades to long duration instruments.
This chart does a great job of showing the volatility differences due to term in a way that is different than typically shown.

I also interpret this chart to show that above an investor-dependent level of stock allocation, it really doesn't matter what is chosen, since the volatility difference are minimal. It also makes it extremely hard to justify bond allocations below at least 20% -- there's a limit to how much "safer" you can get. If you really want zero volatility, get an annuity.

People expecting "zigs" to counteract "zags" will not be happy all of the time regardless of the allocation.
McQ wrote: Tue Mar 26, 2024 10:24 pm Preliminary conclusions

In theory, a total bond market index must be the most efficient bond portfolio, offering the best available combination of risk and return. I call that MPTOS, Modern Portfolio Theory On Steroids (front syllable rhymes with that of “empty”).

BND, as currently implemented, is NOT total—it excludes all bonds rated below Baa. More generally, it is not clear to me that individual bonds are fungible in the way that exchange-listed common stocks may well be fungible. A Treasury isn’t the same thing as a corporate bond, even an AAA-rated one.

Or maybe, the idea that a “total” [anything] fund, one that actual investors can buy in the world, must necessarily be most efficient, remains, to use a phrase from the philosopher Karl Popper, a bold conjecture. Not settled truth.

The efficiency argument for BND fails. What’s left? Okay, BND covers the total maturity range. Ah, what’s the theory there?

With these handicaps visible, BND reduces to a bet on the corporate bond spread versus the necessarily higher correlation of riskier bonds with the risk asset (stocks), when evaluated as a complement to stocks within a portfolio.

Empirically, across a matched bear-bull market pair, a 72-year span, that bet did not work out.

Sticking with an intermediate duration is good; but best to do that using only Treasuries.
Core assumption: Treasuries remain the lowest-risk option for the duration of the investor's investment horizon. Treasuries may be the safest bet today, but investors should be evaluating that periodically going forward.
“Adapt what is useful, reject what is useless, and add what is specifically your own.” ― Bruce Lee
User avatar
Charles Joseph
Posts: 2504
Joined: Tue Apr 05, 2022 10:49 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Charles Joseph »

Thanks McQ. I am 40/45/15 stocks/intermediate-term-treasuries/T-Bills, and I am quite happy with treasuries instead of BND.

Enjoying these threads.
"The big money is not in the buying and selling, but in the waiting." - Charles Munger | Retired June 2023 | 40% US equities/40% US bonds/20% T-Bills.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Kidmoe wrote: Wed Mar 27, 2024 8:30 pm Be curious to see how a 60/40 mix with TBills did especially in the 47-81 time frame.
Thanks for requesting to see a straight T-bill blend; the results were somewhat surprising to me.
Here is the chart with green lines representing stock-T-bill blends. To keep the clutter down, I removed the 60/30/10 LT & T-bill blend.

Image

Taking the dashed green line first—the bear market case—a T-bill blend was in fact more efficient than the intermediate Treasury blend, especially at very low stock allocations below 40%. Looking at the dot at the end of each dashed line, a 100% T-bill portfolio had greater return for less risk than any of the other three bond options.

It’s easy to understand why; in the last three years of the bear case, T-bill returns were in double digits, peaking at 15%. Remember Paul Volcker?

Before getting too excited about what you could get from a default-free, near zero volatility, very short term asset, however, it helps to study the solid green line showing the performance of T-bill blends in the bull case.

Here the performance of T-bill blends absolutely [not family-friendly idiom removed]. Adding even 10% T-bills to an all stock portfolio results in a noticeable diminution of return relative to the other bond blends. By the time you get to 60/40, the T-bill blend underperforms the IT blend by 150 bp in return for no reduction in risk.

Continuing to slide down to the left on the solid green line, by the time you get to a 30/70 blend you have reduced portfolio risk below any of the other blends, but also reduced portfolio return below any of the others. From there it gets worse, with more and more reduction in return for less and less reduction in risk. At 90% T-bills, real return would have been about 2.0%, same as you can lock in on a 30-year TIPS today; at 100% T-bills, only 1.2% real, less than you can get on an ibond.

Stepping back, if you could be certain we were embarked on a new multi-decade bear market in bonds, T-bill blends might be attractive; else, not.

Thanks again for suggesting a straight T-bill analysis. If you study the green lines versus the black lines (long Treasuries) you get a nice tutorial on how portfolio blends using the ultimate in short (long) duration instruments respond during periods of rising and falling interest rates.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
BolderBoy
Posts: 6793
Joined: Wed Apr 07, 2010 12:16 pm
Location: Colorado

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by BolderBoy »

This is really a fabulous thread. Thank you, McQ!
"Never underestimate one's capacity to overestimate one's abilities" - The Dunning-Kruger Effect
Tib
Posts: 320
Joined: Tue Oct 27, 2020 9:57 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Tib »

McQ: "Stepping back, if you could be certain we were embarked on a new multi-decade bear market in bonds, T-bill blends might be attractive; else, not."

Intermediate Treasuries were almost as good as T-bills even in the bear case; they do indeed look to be a good choice. But if I were choosing between T-bills and BND, and if I thought it probable, as I do, that the next 10 years will be bad for bonds, I'd prefer the T-bills. Especially since if the next 10 years turn out to be good for bonds, overall portfolio returns would probably be good enough even with the T-bills.
User avatar
anagram
Posts: 1778
Joined: Fri Aug 04, 2023 1:03 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by anagram »

Tib wrote: Thu Mar 28, 2024 5:57 pm McQ: "Stepping back, if you could be certain we were embarked on a new multi-decade bear market in bonds, T-bill blends might be attractive; else, not."

Intermediate Treasuries were almost as good as T-bills even in the bear case; they do indeed look to be a good choice. But if I were choosing between T-bills and BND, and if I thought it probable, as I do, that the next 10 years will be bad for bonds, I'd prefer the T-bills. Especially since if the next 10 years turn out to be good for bonds, overall portfolio returns would probably be good enough even with the T-bills.
My reading of McQ is that in the bear case Intermediate Treasuries were 5 year Treasuries held for a year and then sold. Vanguard does not have such a fund. VSIGX is a Bloomberg US Treasury 3-10 Year Bond Index and that's a lot different than a 5 year note.

I think that is why Dr Bernstein suggests using VSBSX which is a Bloomberg US Treasury 1–3 Year Bond Index.
User avatar
grabiner
Advisory Board
Posts: 35439
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by grabiner »

How would munis do in a taxable account in this study? The longest test I could do on Portfolio Visualizer gives equal standard deviations for 38% Intermediate-Term Treasury and 62% US stock, or 40% Intermediate-Term Tax-Exempt and 60% US stock. The portfolio with the Treasury bonds has returns 0.73% higher, and given the Treasury returns of 6.38%, the break-even tax rate would be about 29%.
Wiki David Grabiner
Tib
Posts: 320
Joined: Tue Oct 27, 2020 9:57 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Tib »

anagram wrote: Thu Mar 28, 2024 7:21 pm
Tib wrote: Thu Mar 28, 2024 5:57 pm McQ: "Stepping back, if you could be certain we were embarked on a new multi-decade bear market in bonds, T-bill blends might be attractive; else, not."

Intermediate Treasuries were almost as good as T-bills even in the bear case; they do indeed look to be a good choice. But if I were choosing between T-bills and BND, and if I thought it probable, as I do, that the next 10 years will be bad for bonds, I'd prefer the T-bills. Especially since if the next 10 years turn out to be good for bonds, overall portfolio returns would probably be good enough even with the T-bills.
My reading of McQ is that in the bear case Intermediate Treasuries were 5 year Treasuries held for a year and then sold. Vanguard does not have such a fund. VSIGX is a Bloomberg US Treasury 3-10 Year Bond Index and that's a lot different than a 5 year note.

I think that is why Dr Bernstein suggests using VSBSX which is a Bloomberg US Treasury 1–3 Year Bond Index.
Interesting. Note, though, that Vanguard shows their settlement fund, VMFXX, outperforming VSBSX over the last 10 years 1.30% (annualized) to .94%.

https://investor.vanguard.com/tools-cal ... sbsx,vmfxx

I may well just stay put in the settlement fund.
Last edited by Tib on Fri Mar 29, 2024 12:12 pm, edited 1 time in total.
User avatar
Kenkat
Posts: 9610
Joined: Thu Mar 01, 2007 10:18 am
Location: Cincinnati, OH

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Kenkat »

The evidence presented is convincing - IT treasuries have been superior to total bond over a long period of time and under different conditions. Depending on how you want to look at things, the superiority is either huge or it is not huge at all.

Using two competing portfolios: https://www.portfoliovisualizer.com/bac ... UI1jMfaAZl, one 50% IT treasuries / 50% total stock market index and the other 50% total bond / 50% total stock market index, beginning in 1993 with an initial $10,000, rebalanced annually and with no addition contributions, we can make the following observations:
  • The IT treasury portfolio has a final value of $103,209; the Total Bond portfolio $98,622, a difference of $4,587. Almost half the original investment in additional gain. That’s huge.
  • However, from a different perspective - the final value of the IT treasury portfolio is only 4.7% higher than the Total Bond portfolio after over 30 years. It’s not nothin’, but it’s not a life changing difference either. Both investors are probably driving similar cars, living in similar houses, and taking similar vacations.
  • The IT treasury portfolio is better diversified and less volatile. I see a lot that people want that effect from their bonds, the “safe” part of their portfolio. In 2022, the IT treasury portfolio declined 15.02%; the Total Bond portfolio declined 16.43%. That’s a difference of 1.41%. On a million dollar portfolio balance, that’s $14,100. It’s a lot.
  • However, from a different perspective - in 2022, the million dollar IT treasury portfolio would have an ending balance of $849,800; the million dollar Total Bond portfolio balance would have an ending balance of $835,700. IT treasuries are “safer”, but not exactly safe. Neither investor likely walked away feeling their safe bond investment offered the safety they expected. The overall impact to their lifestyle is similar in either scenario.
I am not saying all of this to just be the forum gadfly. The data presented is convincing. My initial reaction upon reading this thread was that maybe I should switch my holding in Total Bond to the IT Treasury fund. Maybe I should do it today.

I will note that I also hold TIPS, short term treasuries, a stable value fund, a money market fund and even a little high yield, so I don’t hold the 3 fund portfolio anyway. But Total Bond is still 7% of my total portfolio and 14% of my fixed income allocation, so it’s not insignificant. It is pretty convincing that IT treasuries are a better portfolio building block than Total Bond. But I also thought through all of my bullet points (above) as well. How much will it actually matter in the end? Will past history continue? Government debt is higher than it historically has been. Corporations seem to be more influential than perhaps they were in the past. So - I just don’t know.

I’m gonna have to think about this one a little more.
gips
Posts: 1781
Joined: Mon May 13, 2013 5:42 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by gips »

cosmos wrote: Wed Mar 27, 2024 11:31 am
gips wrote: Wed Mar 27, 2024 2:24 am
cosmos wrote: Tue Mar 26, 2024 11:50 pm McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
Do you feel using an IT fund instead of bnd in an 80-20 portfolio significantly impacts your risk?
Impacts it how? As in greater risk?? no, just the opposite. the roller coaster ride that BND can have is not what I want for the ballast portion of the portfolio. I also just do not want corporate or reit swaps or anything else in there. My risk is in the us and international stock index funds.

I may switch to ST treasuries or likely just a TIPS ladder once I hit my 60s.
looking at mcQ's graphs, what's the difference in standard deviation for bnd vs IT for an 80-20 portfolio?
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

BolderBoy wrote: Thu Mar 28, 2024 5:14 pm This is really a fabulous thread. Thank you, McQ!
Such kind words are what keep me going :beer
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

Kenkat wrote: Thu Mar 28, 2024 8:36 pm The evidence presented is convincing - IT treasuries have been superior to total bond over a long period of time and under different conditions. Depending on how you want to look at things, the superiority is either huge or it is not huge at all.

Using two competing portfolios: https://www.portfoliovisualizer.com/bac ... UI1jMfaAZl, one 50% IT treasuries / 50% total stock market index and the other 50% total bond / 50% total stock market index, beginning in 1993 with an initial $10,000, rebalanced annually and with no addition contributions, we can make the following observations:
  • The IT treasury portfolio has a final value of $103,209; the Total Bond portfolio $98,622, a difference of $4,587. Almost half the original investment in additional gain. That’s huge.
  • However, from a different perspective - the final value of the IT treasury portfolio is only 4.7% higher than the Total Bond portfolio after over 30 years. It’s not nothin’, but it’s not a life changing difference either. Both investors are probably driving similar cars, living in similar houses, and taking similar vacations.
  • The IT treasury portfolio is better diversified and less volatile. I see a lot that people want that effect from their bonds, the “safe” part of their portfolio. In 2022, the IT treasury portfolio declined 15.02%; the Total Bond portfolio declined 16.43%. That’s a difference of 1.41%. On a million dollar portfolio balance, that’s $14,100. It’s a lot.
  • However, from a different perspective - in 2022, the million dollar IT treasury portfolio would have an ending balance of $849,800; the million dollar Total Bond portfolio balance would have an ending balance of $835,700. IT treasuries are “safer”, but not exactly safe. Neither investor likely walked away feeling their safe bond investment offered the safety they expected. The overall impact to their lifestyle is similar in either scenario.
I am not saying all of this to just be the forum gadfly. The data presented is convincing. My initial reaction upon reading this thread was that maybe I should switch my holding in Total Bond to the IT Treasury fund. Maybe I should do it today.

I will note that I also hold TIPS, short term treasuries, a stable value fund, a money market fund and even a little high yield, so I don’t hold the 3 fund portfolio anyway. But Total Bond is still 7% of my total portfolio and 14% of my fixed income allocation, so it’s not insignificant. It is pretty convincing that IT treasuries are a better portfolio building block than Total Bond. But I also thought through all of my bullet points (above) as well. How much will it actually matter in the end? Will past history continue? Government debt is higher than it historically has been. Corporations seem to be more influential than perhaps they were in the past. So - I just don’t know.

I’m gonna have to think about this one a little more.
Thoughtful and insightful reflections, Kenkat, thanks for contributing these to the thread.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

gips wrote: Thu Mar 28, 2024 8:48 pm
cosmos wrote: Wed Mar 27, 2024 11:31 am
gips wrote: Wed Mar 27, 2024 2:24 am
cosmos wrote: Tue Mar 26, 2024 11:50 pm McQ this is a great article!

I run a 3 fund portfolio (80% stocks/ 20% bonds) 60/20 domestic/intl stocks split and my bond portion is in the intermediate treasury fund. It occured to me a decade ago that ballast meant safety and I want all my risk on the stocks side with no surprises on the safety/ballast side.
Do you feel using an IT fund instead of bnd in an 80-20 portfolio significantly impacts your risk?
Impacts it how? As in greater risk?? no, just the opposite. the roller coaster ride that BND can have is not what I want for the ballast portion of the portfolio. I also just do not want corporate or reit swaps or anything else in there. My risk is in the us and international stock index funds.

I may switch to ST treasuries or likely just a TIPS ladder once I hit my 60s.
looking at McQ's graphs, what's the difference in standard deviation for Bnd vs IT for an 80-20 portfolio?
Easier for me to answer, since I can float my mouse over the charts and see values.

Bear case SD: IT blend 13.8%, BND blend 14.2%
Bull case: IT blend 12.9%, BND 13.2%

In general, as noted above, if stocks dominate the blend, as in 80:20, the difference between the different bond blends reduces to small male rodent generative organs.

It's when the stock proportion drops below 40% that the differences begin to loom larger. Left side of the performance lines, to put it simply.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
jaMichael
Posts: 331
Joined: Sun Mar 12, 2023 12:23 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by jaMichael »

What an interesting discussion! Would the TSP G Fund do as well as (or better than) IT, or should TSPers be moving from G to IT?
Kbg
Posts: 369
Joined: Thu Mar 23, 2017 11:33 am

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Kbg »

Indeed, a fantastic thread! Many thanks McQ.
Parkinglotracer
Posts: 4057
Joined: Fri Dec 20, 2019 2:49 am
Location: Upstate NY

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Parkinglotracer »

goaties wrote: Wed Mar 27, 2024 8:35 am I've long questioned the validity of an index fund for bonds. The bond universe is vast, varied, and they keep inventing new flavors all the time (does BND have MYGAs?). Conversely, indexing makes sense for the relatively limited universe of stocks.

Settling on one type of bond, intermediate Treasuries, does make a lot of sense. It is the quintessential bond, if you will. Maybe that's why it's so successful as a diversifier?
I share the skepticism of the usefulness of index funds for bonds because you remove a bond’s most valuable feature when you pool assets guaranteed to return its value at its maturity with other bonds. It is like buying a winter coat and putting it in a closet with thousands of other people - when you need the warmth (money) and you go and get it, it depends on the weather (interest rates) as to what you have to wear (spend). Why not put it your own closet so you know what is there when you plan to use it?

I was taught by my mentor CFA to take market risk on the equity side of my 65-35 portfolio not on the bond side. Seems to make sense unless corporates are paying much more for the added risk.

At age 63 retired, I use mainly a 5 year treasury ladder for fixed income (small amount of tsp G fund, cash in mm fund) - when a bond matures i move it to mm fund to spend, or rebalance it to equities, or buy another rung. A ladder is a little more work than BND but it relieves the psychological pain of the effects of interest rates on my portfolio - I can sleep well knowing my coat will be there to keep me warm when I plan to get it in the closet,
User avatar
Svensk Anga
Posts: 1647
Joined: Sun Dec 23, 2012 4:16 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by Svensk Anga »

jaMichael wrote: Thu Mar 28, 2024 10:55 pm What an interesting discussion! Would the TSP G Fund do as well as (or better than) IT, or should TSPers be moving from G to IT?
I think it depends on which regime we will be in for the coming years. If it is the bond bear, G, considered in isolation, should do better than IT, since it won't suffer losses due to rising interest rates. OTOH, if it is the bond bull, G should lag, since it won't get the bond price gains from falling rates. G also does not get the price gains from flight to safety moves that occur in a crisis. It should not get as much rebalanced portfolio benefit of weak or negative correlation with stocks.

In 'normal" times, when the yield curve is positively sloped, the IT fund should gain a bit from roll yield. But we are not in normal times at present. G does not get this, AFAIK.

If you want the better investment, you need a good forecast. Maybe those with this option should hedge their bets. I could see using G for money you need in the short term and IT for the longer term portfolio which will be rebalanced.
GAAP
Posts: 2646
Joined: Fri Apr 08, 2016 12:41 pm

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by GAAP »

A graphical way to view the risk/reward profile is to draw a line for the ranges at any given allocation -- for example, at 50/50:
Image
Some conclusions for this particular allocation:
  • Term directly affects the range of risk/reward
  • Term also directly affects the volatility
  • T-Bills basically had very little change in risk/reward
  • IT outperformed BND across both extremes
  • LT vs BND depends upon the conditions at the time
Note that the results would be quite a bit different with some of the bond-heavy allocations.

None of this should be news, but it may help an investor visualize what they can expect for their own allocation.

I did notice that the slope of the lines was not positive for all choices, notably BND -- and it appears to change as the stock allocation increases. LT also does this, starting at higher allocations to stock.

Edit: Change, not necessarily for the worse since it shows a risk reduction in this case.

Apologies to McQ for messing up his chart.
Last edited by GAAP on Fri Mar 29, 2024 3:28 pm, edited 1 time in total.
“Adapt what is useful, reject what is useless, and add what is specifically your own.” ― Bruce Lee
Topic Author
McQ
Posts: 1458
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Can you do better than BND, Part 2: Test across bear and bull markets

Post by McQ »

GAAP wrote: Fri Mar 29, 2024 2:00 pm A graphical way to view the risk/reward profile is to draw a line for the ranges at any given allocation -- for example, at 50/50:
Image
Some conclusions for this particular allocation:
  • Term directly affects the range of risk/reward
  • Term also directly affects the volatility
  • T-Bills basically had very little change in risk/reward
  • IT outperformed BND across both extremes
  • LT vs BND depends upon the conditions at the time
Note that the results would be quite a bit different with some of the bond-heavy allocations.

None of this should be news, but it may help an investor visualize what they can expect for their own allocation.

I did notice that the slope of the lines was not positive for all choices, notably BND -- and it appears to get worse as the stock allocation increases. LT also does this, starting at higher allocations to stock.

Apologies to McQ for messing up his chart.
Compliments to GAAP for adding value to my chart :)
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Post Reply