Global Bond indices: Justification for weights

For the stock part of their portfolio, the average investor should buy a market cap weighted stock ETF.
The weights of the various holdings have a theoretical justification: they are the weights that maximize returns and minimize variance (sort of, it is a bit more complicated than that).

How about the weights of a global, government bonds fund?
What are “optimal” weights? Why going bond-market cap weighted and not e.g. equally weighted?
I was looking at IGLO and I didn’t like that more than 50% is allocated to US bonds.

Thanks for any hint!

As long as it’s hedged, what should matter is the quality and duration (I don’t know if diversifying too much is actually helpful, unlike stock the return of a bond should only return on those 2 things + currency which you do hedge).

1 Like

I think for hedged bond ETF, going global is mainly for credit risk management

From return perspective as @nabalzbhf said your returns would be similar to CHF govt bonds minus hedging costs

1 Like

I’ve been wondering the same and apparently worked under false assumptions so this is an opportunity for me to dig a bit deeper into this. I’m building my knowledge so may work with insufficient data and/or draw the wrong conclusions. Thanks in advance to all those who will participate and help me understand the risks and returns involved better.

Starting with an overview of the Bloomberg fixed income indices: https://assets.bbhub.io/professional/sites/27/Bloomberg-Fixed-Income-Indices-Overview.pdf

More specifically the Bloomberg Global Aggregate index, which is widely used: https://assets.bbhub.io/professional/sites/27/Global-Aggregate-Index.pdf

And with the main question of weighting. For which I’ve found this article from Morningstar which, while it’s not its main topic, allowed me to better understand what Market-Value weighting means for bonds: https://www.morningstar.com/funds/what-makes-market-value-weighting-work-bond-investing

My understanding is that, for fixed income, most funds will have credit rating and liquidity requirements for the securities they include. That’s the first line of protection for us: by using those funds, we trust the credit rating agencies to properly identify the risk of the securities held by the fund and said securities have been reviewed.

The weighting, to me, is a bit harder to understand. We trust market participants to individually evaluate each security and price it properly but since the bond market is a bit more opaque than the stock one, could inefficiencies occur?

The most important factor, for me, is to diversify credit risk. If a risk of default has not been properly evaluated by the credit agencies and the market, I’d want it not to affect my assets too much. How much holdings of a singular creditor is too much to hold is a personal assessment. I’d still check what the weightings in the fund are and reduce exposure to a single creditor by adding another/other fund(s) without that creditor as needed.

That being said, the global aggregate approach seems to me to be secure enough on a broad basis.

As a non-US investor, I wouldn’t feel safe with 50%+ in US securities. Other instruments have less of them, for example AGGS ( iShares Core Global Aggregate Bond UCITS ETF, hedged in CHF) has 28.45% of US Securities: https://www.ishares.com/ch/individual/en/products/295830/ishares-core-global-aggregate-bond-ucits-etf-fund

If that still seemed too concentrated for me, I could see myself overweighting Switzerland and adding a fixed income ETF of Swiss securities to my holdings.

Hedging in funds is based on the difference of short-term interest rates, I think 3-month swaps are common but not completely sure.

The average maturity / effective duration in bond funds such as AGGS is relatively long, and changes in the market expectation of longer term interest rates can have a big effect on the market value of bonds.

These interest rate changes may differ among currencies, of course, which means that also the performance across bond currencies will differ, even if the bonds are hedged and have the same credit rating and duration. This may be welcome diversification but the investor should be aware that hedging doesn’t eliminate all currency differences.

1 Like

That’s what I do. However, in addition to limiting the impact of (hedged) foreign currency bonds, I do this also because the effective duration of AGGS is a bit too long for me. I invest in a SBI AAA-BBB 1-5 fund to lower the sensitivity of my portfolio to interest rate changes.

1 Like

This is just a result of IGLO using the FTSE G7 Gov and EMU Bond Index as a benchmark. The latter is market cap weighted.

Looks like the market consists of more than 50% of such issues in USD.

If you don’t like the weighting, luckily, as a “passive” investor, you can always choose another bond ETF / underlying benchmark … :wink:

1 Like

I’ve been pondering US treasuries and still do not understand fully market weighting.

If someone with a high credit rating (make is the US Treasury) issues a lot of bonds. Would the market be flooded with those and would that increase their share of the market-weighted bond indexes vs a more conservative debtor (say the Swiss Confederacy) with a similarly good, or better, credit rating issuing less securities?

Market participants should correct for it but how would you do it? I don’t really grasp the mechanisms of it.

It should return a bit more than CH government bonds. As those have 0 credit risk premium, due to their safeness.

Well, I am also not a specialist, but it always takes two to make a trade. These treasuries are not issued into vacuum, someone is buying them at the issue, right? On the other hand, this “someone” could be also Fed. :person_shrugging:

To be honest, I think that Market Cap weighted Indexing of Bonds was a terrible idea - but unless you would want to go active, there is probably no alternatively.

In my view - Global Aggregate Bond Indices may serve as an indicator to establish a 4 dimensional distribution consisting of i) Risk Rating, ii) Duration, iii) Currency and iv) Bond Type (Treasury, ABS, Corporate, …). We then best establish a Portfolio that matches this distribution, but optimizes against both Issuer, Jurisdictional and (for corporate) Sector Concentration.

Meaning - I don’t see any reason why a country / company with material debt shall get “its full share” when attracting me as an investor. I rather invest in a way where concentration was managed.

From a Risk/Return point of view - such optimization should not have any effect at all. The only thing that probably changes is the distribution of adverse events in a sense that they was probably a bit higher background noise of defaults but the magnitude/impact of those was muted. Or in other words - less long tail / black swan risk.

The Problem is - how to efficiently invest into bond markets (unless you pick and choose your own bonds)? I just don’t know. ETF are probably not really the right answer here.

1 Like

Treasuries are auctioned off.

They’re bought by “the market” because they’re safe, stable, liquid, reserve currency, etc etc.
The Fed will step in doing regular Open Market Operations to maintain the Fed funds rate within its range or to reinvest maturing securities on its balance sheet. Occasionally they’ll venture into quantitive easing, tightening, or to stabilize the repo market or “the market” in general.

Bond index providers calculate the weightings (of, say, the FTSE G7 Gov Bond Index) based on debt outstanding issued by the G7 countries in their currencies modulo some inclusion criteria (e.g. minimum issue size, maturity range, etc) and they re-balance the weighting regularly (e.g. monthly).

Bond funds or ETFs replicating such an index will buy and sell based on the index composition. Often, they’ll try to “front run” index rebalancing at the time of anticipated “larger” index composition changes (where “large” is usually quite small, but still worth making a buck if you can trade ahead of the exact rebalancing date).

2 Likes

Its a matter of arbitrage potential. If the treasury floods the market with new issuances, there would in theory be insufficient buyers for those. Meaning that the Interest the US Government had to pay would increase. This until additional buyers were found, that were willing to absorb this issuance. This would technically lead to a situation where US Government and USD nominated Canada Government Treasuries (if/as considered equal risk) of equal duration would have a different interest rate. Meaning that arbitrators could now short-sell Canadian Treasuries and invest in US Treasuries. This until the Interest Rate differential among both was equal/zero.

What happened? The US flooded the market with new issuance and the combined market consisting of all Global Treasury Debt absorbed the volume, not just “US Bond Buyers” as such. And as the Market anticipates this to happen, we would not even in the first place end up with an Interest Rate Delta among both. Its kind of a self-fulfilling prophecy.

This clearly only works as long as the mass issuance of the US Governement does not adverse the either stated and/or perceived Credit Default Risk of US Debt.

1 Like

Thanks for your insights.

@Your_Full_Name @TeaGhost

Both your answers lead me to believe that countries/issuers get rewarded for issuing more bonds/increasing their debt provided it doesn’t affect their credit ratings.

In this case, the weight of any given security in that index depends on the total amount of debt it represents available, provided they meet the inclusion criteria of the index.

In that situation, USD denomintated Canadian Government securities being shorted should reduce their price and increase their yield, that is, their issuers have to pay more interests in order to keep attracting buyers.

The US Government securities are bought, which should bring their price up and their yield down: the US government has to pay less interest to service their debt.

It seems to me that there is a threshold effect related to credit rating/inclusion criteria in indexes and that as long as they manage to maintain their inclusion in said indexes and their position within a range of credit rates, a debt issuer can do no wrong. Is that true?

As an additional question, let’s say AAA to BBB- securities are included in an index. The AAA securities, being more secure, have a lower yield. The BBB- securities have a higher yield. Both meet the inclusion criteria for the index. How does the index provider manage the weighting of both securities? Has their credit rating/risk/yield any effect on it? And is that effect that higher yield while meeting the inclusion criteria means more weight given towards them, favoring the less reliable borrowers of the lot (who are still reliable enough to be included)?

Well, as long as it doesn’t slip below investment grade (BBB-) … :wink:

The US Treasury has already been downgraded by the S&P and by Fitch (Moody’s still has them at the highest rating). In practice, the market and bond investors don’t care.

Yes, for those indices purely putting together the index based on “market cap”.

One underlying thought is of course that a country will only issue debt up to a “reasonable” amount, otherwise their rating dips below investment grade at which point they won’t make the index anymore.

The US, given its unique global position, can (and does) stretch things a bit, obviously.

They don’t?

Unless the index provider has an explicit policy for taking into account the credit rating. Previously mentioned FTSE G7 Gov doesn’t, it only has a “minimum” credit rating requirement and the weighting is then done by market cap.

1 Like

I know we all like to lament how high the US government debt is, but “flooding” isn’t quite the right term when the US Treasury issues new debt … in fact, US Treasury auctions are often oversubscribed, i.e. “the market” is soaking up that debt like a sponge!

Bid-to-cover ratios tend to be in the 2-3 range (above 1 is oversubscribed) – the world is thirsty for the US Treasuries …

While there is some base demand by Primary dealers* who are required to bid on U.S. Treasuries the oversubscription is really driven from both domestic and international investors, enabling the U.S. Treasury to secure lower borrowing cost.

If you take a non-Swiss perspective and as an asset allocator you had to put really large amounts of money into bonds while taking into account the respective yields, what government bonds would you buy?


* Financial institutions designated by the Federal Reserve to participate directly in Treasury auctions.

3 Likes

Thanks for your answers.

As a fund manager, managing my career risk, I would buy the bond everybody is buying and then, if it fails suddenly, lament “nobody could have seen it coming, I’ve done what the entire profession does”.

As an individual investor living in a country with a stable currency, I would see the above as an additional risk and would diversify my assets away from market cap for fixed income.

Of course, if I lived in a country with a not so stable currency, I’d have much more faith in the stability of the US dollar (as a Swiss investor, not so much so. Bond money isn’t there to get lost on sudden mood changes in the market, even for some extra returns).

2 Likes

:rofl:

Reminds me of the past century when nobody in charge of the computer purchase department ever got fired for buying IBM, even when it was already clear IBM was becoming a living fossil … :wink:

Slightly off topic: in fact, I am amazed IBM is still around, even with a respectable market cap.

1 Like

Some argue that global aggregate, while having slightly higher returns, are more correlated to stocks than global government.
For a psychological aspect, I was more leaning towards the second kind of bond ETF: government, not aggregate.

FWIW I still think bonds are not like stock, and having a ton of diversification might not be that more helpful.

Quoting Matt Levine:

Intuitively, “bonds are all kind of the same” has to be even more true than “stocks are all kind of the same.” Some stocks are Nvidia; many others are not. But a bond is like “you put in $100 and you get paid $X per year for Y years and your $100 back at the end.” X and Y will vary a bit, but they are known in advance. Your upside is finite and known; your downside is limited. Some bonds are riskier than others — some bonds are issued by safe companies with lots of money and will definitely pay you as promised, while others are issued by unstable companies that might not — but even those risks are pretty standardized with well-known public credit ratings. “One BBB+ bond is as good as another BBB+ bond so just buy whatever BBB+ bond you see first” is not a good investing strategy, and there are lot of credit investors who get paid a lot of money not to do that, but isn’t it kind of plausible? Doesn’t it sound about as right as “stocks are largely linear combinations of factor exposures”?

https://www.bloomberg.com/opinion/articles/2024-10-28/you-want-some-stocks-that-go-down

1 Like