Matt, you definitely have it right when you suggest that bond ETFs present a new level of complexity when compared to equity ETFs – but I’d still use them.
Here are just some of the problems compilers of bond indices, and the developers of associated ETFs, face.
- Too many bonds – over 3 million of them, said James Rieger of Standard and Poor’s at the recent EDHEC conference in Paris.
- Pricing problems – many of these bonds are the offspring of the credit bubble and are now illiquid. There is great scope for market manipulation and inaccurate pricing.
- Seasoning – in many bond markets, only recently issued bonds remain relatively liquid, whereas older, harder-to-trade issues may dominate the indices.
- Opacity – often you can’t see through bond indices to the underlying securities as easily as you can for an equity index. Some bond indices are compiled by entities that do not make the index composition publicly available.
- Cap-weighting is even more problematic – if you think that weighting by market value presents a problem for equity indices, since this leads to excess holdings in the most overvalued companies, and too little in the most undervalued, just think what happens if you use cap-weighting in bonds. You end up with most of your money invested in the bonds of the most indebted issuers – surely a recipe for disaster.
- Interest rate/duration risk and credit risk get mixed up – typically, these components of an investor’s return are not separated when you buy a fixed income ETF (though, of course, credit ETFs do allow you to trade only the credit component). With the recent rise in the credit risk of government issuers, this problem now affects all kind of bond trackers.
- Too many financials – as investors in European corporate bond ETFs are well aware, the debt of financial companies tends to be overrepresented in indices. This is far from ideal, given that financials are the most highly-leveraged sector of the economy and we’re in a credit crunch.
- Ratings problems – if you select bonds for your index based on credit ratings, you are faced with the problem that the ratings agencies have a terrible record of reflecting problems at the issuer level in a timely fashion.
- Index “lumpiness” – as most bonds have fixed maturities, they “roll down (or up)” the yield curve as they get closer to expiry. This makes bond indices which are structured by maturity bands “lumpy”, there can be a significant drift in the average maturity of the relevant index band as bonds shorten in life and join the next maturity segment, and there will have to be frequent rebalancing of associated ETFs.
- Bonds can have negative convexity, or have other features that affect their maturity, such as call or put options – mortgage-backed securities, for example, suffer from negative convexity, meaning that their maturity/duration rises in rising interest rate environments, and shortens in falling interest rate environments. This presents a major problem if you are trying to define indices by maturity segments.
- Differences in tax treatment – as a UK investor, for example, it makes more sense to invest directly in bonds than to buy them via a fund, since you are not liable to capital gains tax if you go the first route. In Italy, for example, there is a tax advantage in holding money market ETFs when compared to bank deposits. These tax effects can be complex.
Having said all that, I think that bond ETFs are very much a part of the landscape and are likely to remain so, particularly in continental Europe, where fixed income has traditionally played a more important role in investors’ portfolios than in the Anglo-Saxon countries.
The challenge is for ETF issuers and bond index compilers to work on creating better benchmarks, which reflect more closely what investors really need. That’s not to say that existing ETFs are badly-designed – there’s already substantial investor participation in a number of benchmark passive European bond funds, whether inflation-linked, long-duration, or corporate. The inflows into corporate bond ETFs in the last nine months have been phenomenal. Even with all the problems I’ve listed above, these funds have given investors an easy and cheap way to access corporate debt, an area which was historically the preserve of large institutions. But that’s not to say that the design of bond ETFs couldn’t be improved.
I noticed a year ago that the US ETF provider Ameristock Ryan was forced to close its constant duration US Treasury bond ETFs due to lack of investor interest. Perhaps they didn’t devote enough marketing resources to their funds – I don’t know – but I always thought that these funds were well-designed, since they avoided the problem of maturity drift and offered useful building blocks for a portfolio based upon a measure that is probably the first port of call for a bond investor, duration (this measure weights the cash flows from a bond according to their present, discounted value).
Anyway, I’m confident that the area of bond index construction and the design of associated ETFs is a fertile area for people to work on, and we should expect plenty more innovation in this area. And I’d still use bond ETFs (though, as a UK-based investor, only in a non-taxable savings account like a pension plan).