The Best Bond ETFs for a Passive Portfolio

The Bottom Line

We compared more than 25 Bond ETFs listed in US and London to identify the best options for a passive portfolio.

Our choices were primarily based on the 15-year total cost of ownership of the ETFs and their liquidity, although we considered many other factors.

The uncertainty around the risk of an estate tax on US ETFs is hard to resolve and each investor should make that decision for themselves. We recommend some Global and regional US ETFs and UCITS.

We may add recommendations for more geographies gradually but the current recommendations should cover a majority of any passive portfolio.

If you have already read our post on the best equity ETFs, you should at least read the portion on withholding tax below as the treatment for bond ETFs is slightly different.

Table 1: Best ETFs for Bond Exposure

The Research

Who this is for

If you want to invest in a globally diversified passive portfolio at a low cost but are confused by the large number of ETFs listed on various exchanges, our recommendations are for you.

This post focuses on the best Bond ETFs. See here for more on the best equity ETFs.

Our research was largely focused on people living in Singapore but most of the analysis should be relevant for people in other locations as well.

Why you should trust us

We spent 10s of hours researching the differences between ETFs listed on different venues, primarily in US and London, but also in Hong Kong & Singapore.

We looked through the details on 100s of ETF to identify the most promising 27 bond ETFs. We then applied the most comprehensive system of comparison to definitively answer the question of which Bond ETFs work best as a part of a globally diversified portfolio.

How we picked

According to Wikipedia, there were more than 7,600 ETFs listed across the world in 2020. While most of these are not relevant if you’re trying to create a low-cost, globally diversified portfolio, selecting the right ones can still be hard.

This is because often you have multiple options for a single type of exposure. For example, there are more than 80 ETFs that provide a broad-based exposure to US Fixed Income market. How do you choose the best one?

Our approach for shortlisting bond ETFs was similar to what we used for equity ETFs. We summarize it below but if you want to understand our method in detail, you should read that post.

We relied on 3 main factors to make our recommendations: 15-year total cost of ownership (TCO), Total AUM, and Liquidity.

Our recommendations do NOT take into account any risk of the US Estate Tax being applied to US ETFs. You can read more on this here (click and scroll down a bit).

Total AUM or assets under management is the total amount that has been invested in the ETF by investors. This is important because smaller ETFs may not generate enough revenue to be viable for the issuer, and the issuer may eventually close the ETF and return the funds to investors.

Usually this will be a small inconvenience as you can simply reinvest the refunded money into another fund. There is a risk, though, that you need to stay out of the market for a few days while you wait for the refund and are able to reinvest.

To avoid this risk, we preferred ETF that had large AUMs, ideally at least US$1bn and where possible we eliminated ETFs with lower AUMs from our consideration.

Liquidity More liquid ETFs are preferable as they allow you to quickly buy or sell as much as you need to. Less liquid ETFs can require you to buy or sell over multiple hours or even days, depending on the size of your position.

All things equal, we preferred more liquid ETFs, ideally trading more than US$1 million a day or (if the ETF is likely to have a lower weight in a passive portfolio) able to support a at least US$10 million portfolio.

If your portfolio is likely to be larger than this amount, look at the daily liquidity of the recommended ETF. If its in the same ballpark as the size of your likely position, you should consider using other (or multiple) ETF, even if they are not otherwise the best.

15-year Total Cost of Ownership (TCO) Why 15-year TCO? There are a number of different costs associated with owning ETFs that trade in different venues. Some of these costs are one off while others are annual.

So comparing the costs over a single year will be incorrectly penalize a venue where the one off costs are higher but ongoing costs are lower.

Since we are looking for ETFs for a core, long term portfolio, we assessed total cost assuming that you hold the ETF for 15 years and then sell.

We preferred ETFs with the lower total cost.

We have assumed no capital gains tax, which is the case in Singapore. In other countries, the treatment may be different depending on various factors.

Unless your country has a specific treaty with the country where the ETF is listed, the treatment should be identical across all ETFs and should not change the recommendation.

 – All-in Trading Cost We used the all-in trading cost for each market on Interactive Brokers based on our analysis of brokers in Singapore. This includes the broking commission as well as the FX spread charged by IB on both, the buy and the sell transaction.

If you use another broker in Singapore, chances are you will pay an additional 0.60%-0.80% for converting your SGD to USD or GBP when buying and back to SGD when selling.

We found that the all-in trading cost isn’t that important for the 15-year total cost no matter which broker or trading venue you use.

One cost we did not include was the bid-ask spread for different ETFs. We found it difficult to get this information for many ETFs. However we think the impact of bid-ask spreads on the total cost should be small. This is because where the data was available, the spreads ranged from <0.01% to 0.2%.

Since this cost is incurred only twice over 15 years, it does not move the needle on our recommendations.

 – Total Expense Ratio (TER) is the total fees and costs charged to you by the company operating the ETF. This is charged every year and was included 15 times in the calculation of the 15-year total cost.

This is not strictly accurate since the management companies can, and do, change the TER for ETFs over time, but it is the best estimate we have for now.

 – Tracking Difference measures how much an ETF’s performance differs from its benchmark index. The difference may happen because of fees, difference in composition of the ETF and the benchmark index, and many other factors.

TER is usually the single largest contributor to the Tracking Difference although tracking difference can sometimes be lower than the TER or even positive. For calculating the total cost over 15 years we took the more negative of the two for a conservative estimate.

This is a somewhat conservative assumption and may penalize some ETFs unfairly.

We do not think this has impacted our recommendations anywhere but it was an important factor in many cases.

 – Dividend Withholding Tax Impact In our post on equity ETFs we discussed that that Irish UCITS listed in London get a better tax treatment than US ETFs.

Since this tax benefit occurs every year, it is a very significant portion of the all in cost for an investor.

This impact can be much higher for bond ETFs listed in US because the coupons and other payments can sometimes form a much larger part of the overall return compared to equity ETFs.

In case of bond ETFs it is theoretically possible to avoid a fair bit of this tax. That’s because for many bond ETFs, a large portion of their distributions are “qualified” as capital gains or interest payments.

These are not subject to the withholding tax for non-US residents. The distributions are only “qualified” if they are from a US source i.e. the company paying them to the ETF is domiciled in the US. ETFs that focus on non-US bonds will not benefit much from these exemptions e.g. BNDX.

For a number of popular bond ETFs, the proportion of qualified distributions ranged from 70-100%. This should reduce the effective withholding tax on these ETFs to 0-9% instead of 30%.

However you can only benefit from this lower effective tax if your ETF provider and broker are both aware of these regulations and able to implement them correctly. Otherwise you may still pay the full 30%.

Interactive Brokers does not seem to implement these regulations. Dividends from our position in IUSB were taxed at the full 30%, despite the fact that iShares designates nearly 75% of the distributions from IUSB as qualified.

In this case you have the option of filing form 1040-NR with the US IRS and asking for a refund. However this may not be easy. We also wonder if, longer term, this interaction with the US IRS increase the risk of your assets being subject to the estate tax.

Because of this complexity, we largely ignored this benefit for US ETFs in our calculations.

Overall, dividend withholding tax was the single largest cost element over 15 years for most Bond ETFs.

In our calculations, we assumed 0 withholding tax impact for Irish UCITS and an annual cost equal to 30% of the dividend yield for US ETFs.

This overstates the tax disadvantage of US ETFs.

When comparing the Irish UCITS and US ETFs, the tax impact can be split into two parts.

First is that on a like for like basis, an Irish ETF will have a lower dividend yield than a US ETF (assuming both ETFs own some US Bonds). This is because the coupon paid to the ETF by the US company is not taxed for the US ETF and could be taxed at 15% for the Irish UCITS.

Taking the example of US Aggregate Bond Market ETFs (e.g. AGG & IUAA), with a hypothetical dividend yield of 3%, the Irish UCITS (IUAA) will only  receive 2.55% (85% of 3%) while the US ETF (AGG) will receive the entire 3% from the underlying companies.

When the two ETF then distribute these dividends to you, the Irish UCITS does not withhold any tax while the US ETF withholds 30%. So you receive 2.55% with the Irish UCITS and 2.1% with the US ETF.

The true difference in tax cost for the two ETFs is 0.45% per year. However if we simply assume a 30% tax rate for the US ETF and 0 for the Irish UCITS, the calculated impact will be 0.9%.

In the above example we could simply calculate the cost disadvantage of the US ETF by subtracting 2.1% from 2.55%. However many Irish UCITS do not pay out the dividend. Instead they reinvest it.

This is great for investors as it saves you the cost of reinvesting but makes it difficult to calculate the cost differential.

For Bond ETFs, we were also not able to find good information on how Irish UCITS treat qualified income. Because of this, unlike with Equity ETFs, we did not make any adjustments for this effect.

This is not tax advice. You should consult tax accountants and lawyers before investing in these instruments.

Additional Factors

Reinvestment of Dividends US ETFs are required by law to pay out at least 90% of the income they receive from their portfolio companies. As a result, they usually pass on the full interest & dividends.

UCITS do not have any such requirement. Some of them pay dividends while others simply reinvest them back in the fund. Often, there are 2 or more versions of the same UCITS, varying by hedging currency and dividend payment policy.

If you do not need the dividend income from your portfolio for your daily expenses, choosing the accumulating version of UCITS can be more efficient. This is because if you receive the dividend, you need to reinvest it or it will earn close to zero in the bank. If you do reinvest, you must incur trading costs. Cost of automatic reinvestment is included in the TER of the UCITS.

Investors who need the dividend income can either choose the US ETF or need to be sure they are choosing the correct version of the UCITS, which adds to the complexity of investing.

We did not factor this in our recommendations.

Hedging A number of UCITS ETFs come in versions that hedge the returns in different currencies. For e.g. iShares Core Global Aggregate Bond UCITS ETF comes in 8 versions, from unhedged to SGD hedged and NZD hedged.

We preferred the most liquid version, which was the USD or USD hedged version in most cases (but not always).

Data Availability for ETFs is not great, especially for Irish UCITS. We had to collect the data for Irish UCITS from multiple sources, which means that is it hard to be sure that data points for different ETFs are comparable.

For example, the sources almost never specified a time period for data points like dividend yield or tracking difference.

Was the dividend yield quoted the actual yield over the last 12 months or was it the projected yield for next 12 months. Was the tracking difference for last 1 year or for 3 years or 5 years?

We do not know. Data on US ETFs was relatively better and we were able to source most of it from a single source and should be consistent across ETFs.

But still, we need to emphasize again that these TCO estimates are quite rough and we should accept the cost differential as a deciding factor only if it is a large difference.

The Best ETFs

We recommend both UCITS and US ETFs for each major geography, where possible. If the liquidity for the UCITS is adequate for your portfolio, you should choose the UCITS as total cost of ownership over 15 years is typically much lower for these.

This is especially true for groupings with no US exposure e.g. EM. Here the withholding tax really hurts over the long term.

Global

These are the best ETFs if you want an all-in-one solution for the bond allocation in your portfolio. AGGU UCITS tracks the Bloomberg Global Aggregate Bond Index, which provides exposure to more than 10,000 bonds globally. It has about a 40% in US and 9% in China.

If the liquidity of AGGU is not sufficient for your portfolio, you can consider topping up with AGGG which is identical to AGGU, except it not currency hedged for USD and it pays a dividend.

We did not find a good all-in-one option in US. Instead you can replicate a global exposure by combining AGG US, which tracks Bloomberg US Aggregate Bond Index and BNDX US, which tracks Bloomberg Global Aggregate ex-USD Float Adjusted RIC Capped Hedged to USD Index (phew!).

Table 2: Best ETFs for Global Bonds Exposure

USA

If you want just US Bonds exposure in your portfolio, IUAA is the best option among UCITS. If you need an US ETF, choose BND, which tracks Bloomberg U.S. Aggregate Float Adjusted Index or AGG (above) .

Table 3: Best ETFs for US Bonds Exposure

Emerging Markets

If you need a higher yielding bond exposure in your portfolio, an aggregate EM Bond UCITS or ETF may do the job. A passive portfolio should ideally have a fairly small exposure to these though. Both JPEA and EMB provide EM exposure via USD bonds issues by EM Governments. There are some UCITS and ETFs that offer local currency bond exposure (EMDD LN, EMLC US, EBND US etc.) but we felt they might be more appropriate for active portfolios.

Table 4: Best ETFs for Emerging Markets Bonds Exposure

Complete Data

The complete data used in our analysis is available as a Google sheet.

Disclaimer: We have made every effort to ensure the accuracy of all data and analysis included above. However it is possible that some errors made it into our analysis anyway. The quality of data we had to work with was also not great. Before investing in any of the recommended ETF, do verify all the details. If you spot an error, please let us know and we will do our best to fix it.