When Rob Arnott, the creator of fundamental indexing, spoke in Toronto last week, his presentation focused exclusively on equities. When he took questions at the end of his presentation, I asked him a question about the lesser-known RAFI fundamental bond indexes.
Recall that the goal of fundamental indexing is to address the flaws in capitalization-weighted indexes, which give the most influence to stocks that may be overpriced. Traditional bond indexes are cap-weighted, too: the more bonds a country or corporation issues, the greater their weight in the index. Arnott and others have criticized this methodology, so I asked him to comment. Here’s what he said:
“If you’re bond investor, you’re a lender. If you’re cap-weighted, then you’re lending the most to whomever has the most debt. What is the rationale behind that? If you’re investing in a global sovereign bond fund and Greece decides it wants to double its debt, as an index investor you would have to own twice as much. What’s the sense of that?
“Cap-weighting in bonds is patently ridiculous. Cap-weighting in stocks has a lot of theoretical justification, but when you use the same theories and apply them to bonds, it just becomes more obvious that the rationale is stupid.”
Ridiculous, stupid — these are strong words. But before you rush to dump your bond index fund, let’s take a closer look at Arnott’s argument.
A red herring
On the surface, the criticism of cap-weighted bond indexes seems to make sense. If you’re a mortgage lender, do you want you lend more money to people who already have the largest home loans? If you’re credit card company, do you look for customers who owe the most on their existing cards? Of course not. You look for borrowers with good “fundamentals”: a reliable job, a low debt-to-income ratio, a good credit history and the like.
But this has no relevance to the vast majority of index funds and ETFs that hold government bonds.
For starters, most investors hold all of their bonds in their own country—as they should. Unlike with equities, there are sound reasons for keeping your fixed income in your home country and currency. Some investment gurus, such as David Swensen and Larry Swedroe, even argue that you should avoid corporate bonds and stick to bonds issued by your own government. In other words, global sovereign bond funds are not a major part of most index investors’ fixed income holdings.
Even if you did decide to add foreign bonds in your index portfolio, the reference to Greece (or any other country at high risk of default) is a red herring. Cap-weighted bond index funds screen for credit risk, so no investor is in danger of being unwittingly saddled with a bunch of fiscal basket cases.
For example, the CIBC Global Bond Index Fund, the only one of its kind in Canada, tracks the cap-weighted J.P. Morgan Global Government Bond Index, which includes only investment-grade bonds. Most of the fund is in Japan, the United States, France, the UK, Germany, the Netherlands and Germany. (There are small allocations to Italy and Spain, which may concern you, but this is a criticism of bond-rating agencies, not cap-weighted indexes.)
The SPDR Barclays Capital International Treasury Bond ETF (BWX), listed in the US, is a perfect example of the methodology in action. This ETF used to have an allocation to Greece (according to this February 2010 article), but it doesn’t anymore, because the country’s credit rating has since been lowered and it’s no longer eligible for the index.
Other sovereign bond ETFs use different methods to make sure deeply indebted countries are not overrepresented. The US-listed iShares S&P/Citigroup International Treasury Bond Fund (IGOV) starts with a cap-weighted index but makes adjustments “designed to distribute the weights of each country within the index by limiting the weights of countries with higher debt outstanding and reallocating this excess to countries with lower debt outstanding.”
If you’re an investor who wants high-risk bonds issued by emerging countries, you’re still not likely to run into any problems with index ETFs. BMO’s Emerging Markets Bond ETF tracks an index “weighted by gross domestic product as a measure of economy size.” The recently launched iShares J.P. Morgan USD Emerging Markets Bond Index Fund (XEB) uses the same safeguard as IGOV.
Designing a bond index fund that systematically overweights countries in danger of defaulting would indeed be “ridiculous” and “stupid.” Which explains why no bond index fund seems to be doing it.
Clearly Mr Arnott is, among other things I’m sure, a salesman. Either that, or you are way smarter than he is! Presumably he knows this stuff as well as you, but can’t offer the balanced, thoughtful approach that you do, because he’s got something to sell. Thanks (again)
Thank you again for another good article.
I have a general question in my investing education. I have say 5,000$ to invest in an RRSP account, another 5k in TSFA account and 5K in completely unsheltered. I have a basic idea of what I’d put in which accounts, but you guys are leaps and bounds ahead of my understandings. Could anyone mention which types of etfs they’d put in each category or even better if there has been a past article or future article on this.
@Doug: Let’s make it clear that Arnott is infinitely smarter than I’ll ever be. :) But, as you say, I think that sometimes he pushes the marketing too far, and I like to encourage investors to have a healthy skepticism about financial innovation. Cap-weighted indexes are not ideal for bonds, but in practice, ETF providers seem to realize this and have taken steps to mitigate many of the problems.
And as always, the proof is in the track record. Well run bond index funds are almost always in the top quartile over long periods compared with actively manged bond funds. They are still the best we have.
@inglishmagor: I have done a previous article about investing small amounts:
You may also want to join the Canadian Money Forum and/or the Financial Webring to get more specific advice.
Thanks for another great analysis. I’m torn between your usual apt advice, and that of popular Rob Carrick’s take on the bond issue which he published just 2 days ago (http://www.theglobeandmail.com/globe-investor/investment-ideas/portfolio-strategy/canadas-bond-market-too-insular-for-its-own-good/article2030384/), pointing out that that the Canadian bond market is heavily skewed to government and financials – he actually promotes the global bond market for its diversity. I’ve actually never understood why Canadians seem to only like Canadian bonds – I had assumed that it had something to do with taxes or avoiding currency risk (which could be mitigated with the hedged funds in Rob’s article). Could you elaborate on your statement: “there are sound reasons for keeping your fixed income in your home country”?
@sleepydoc: Thanks for the link to Carrick’s article, which is really interesting. I think I will have to do a post on this in the future. But the short answer is that most of the reading I have done on asset allocation stresses the importance of keeping your bonds as safe as possible. That means no currency risk and no default risk (as you would have with emerging market bonds). Your diversification comes mainly from spreading out maturities.
And as I mention in the post, some even recommend avoiding corporate bonds and sticking to governments only, so when events like 2008 occur, your fixed-income saves you rather than plummeting along with your equities. Government bonds have the lowest correlation with equities, which is what you want in a portfolio.
That said, if you do invest in corporate bonds (either investment grade or high-yield), I see no problem with adding an allocation to an ETF of US bonds. As Carrick’s article points out, the corporate bond market is quite small. But I would use currency hedging — all of the US bond offerings from iShares, BMO and Claymore use hedging.
Well, whatever his words, the proof of Arnott’s hypothesis is always in the pudding. Will his index / an ETF that adheres to it, beat the close counterpart “index” based loosely on cap-weighting (whose looseness you have well pointed out)? It’s the active bond security selection process of RAFI vs that of the sort-of cap-weighters. Well done for a good article to stir up controversy on a topic that well deserves some strong discussion.