Michael Edesess has a new piece called “Is the Bond Index Broken?” His article is an answer to some criticisms of the Barclays Aggregate (link to Barclays info here). He seems to be mainly referring to its use as a benchmark, but since he references a discussion with John Bogle, it’s safe to assume that he also is including its use as an index for direct investment (such as through funds or ETFs, like AGG).
The three criticisms of the bond index
Inquiries made to people with long experience in the investment field identified three criticisms that have been leveled at the bond index:
I’m a huge Edesess fan, but that list does not include some of my biggest concerns with the index. Let’s start with his list, then I will add my concerns:
- Overweighting of heavily indebted issuers. Edessess believes this is not a problem. However, the problem he describes is one of mispricing. This is how cap weighted equity indeies become “overweighted.” That is, if an equity has too high a price, it will be over-represented in a cap weighted index. I don’t think this is the problem in fixed income. Say company A and company B are essentially identical. Both have identical cash flows, and identical amounts of cash on hand, but company A issues twice as much debt as company B. AGG will have twice as much debt from company A. The problem with this is that it would seem on its face that company A is the company you would want less exposure to, not more. Mispricing can add to this problem, but it’s not the problem.
- Double counting of some securities. I agree with him here. It may be a problem, but it shouldn’t be significant.
- Not representative of investor portfolios. Here’s what he has to say about that:
Here, the criticisms are those leveled by Bogle himself, and they do bite. About 70% of the U.S. Aggregate Index, according to Bogle, consists of U.S. federal government-issued securities. The total of those securities, according to Bogle, is about $16 trillion. Of that amount, however, at least half is owned by national governments – notably China, Japan, the UK, and, in fact, the United States itself. …
The demand for U.S. federal debt for such purposes has pushed up its price relative to less highly-rated corporate bonds. Hence, funds that replicate the Barclays U.S. Aggregate Index have substantially lower yields than U.S. corporate bond funds.
For example, Vanguard’s Total Bond Market Index mutual fund (VBTLX), which is invested, according to Vanguard’s web site, “about 30% in corporate bonds and 70% in U.S. government bonds of all maturities,” currently has a yield of 2.3% while its Intermediate-Term Corporate Bond Index fund (VICSX) has a yield of 3.5% – a difference of 1.2%.
Bogle argues that the small additional risk of default and slightly longer durations in the corporate bond fund (6.4 years for VICSX versus 5.8 years for VBTLX) does not justify forgoing 1.2% of yield, especially when – in the current interest rate environment – that 1.2% represents a reduction of yield of more than a third. To the extent that standard deviation measures risk, for example, the Total Bond Market Index fund VBTLX has a standard deviation of about 3.0% while the standard deviation of VICSX is 4% to 4.5%. Bogle questions whether this is enough difference in risk to justify a 1.2% premium for corporates.
I agree, in a sense. However, the AGG is intended to include all the investment grade debt that is out there. The lower grade the debt, the more it acts like equity from a variability and correlations standpoint. People own bonds, in general, for income and safety. Portfolio managers add bonds for diversification. Neither of these groups want the bonds to act more like equity, and as is pointed out above, funds with those kinds of bonds are available for those who do. The Barclays Aggregate Index is the default specifically for investment grade issues. The issue here, I guess, is that it’s used as the default index for “bonds”. If point #1 is not a problem, this shouldn’t be either. And if this is a problem, then the overweighting of more heavily leveraged entities’ debt should also be looked at. It depends on what you are using the index for, I guess.
So what are some of the other concerns?
4. Debt is removed from the index one year before it matures. Why? I suppose as debt matures, the bond just starts to act like cash, but that’s the reality of bond ownership. I would think that when all the index funds and ETFs sell off their one year from maturity debt, that those issues tend to sell slightly below value (supply and demand, right?).
5. The maturity and size constraints can cause huge turnover within these funds. If your goal is to own something that acts like an investment grade bond, but with the diversification of a fund, then you don’t want high turnover. Turnover for bond index funds range from 72% (BND) to 318% (!) (AGG). With an average duration around 5.4 and average maturity of 7.5, why so much turnover?
6. The details of the rebalancing rules make this index easy to front run. They publish daily both a backward looking index, consisting of current constituents, and a forward looking index. The index is rebalanced to match the forward looking index on the last day of the month. This is not to say that the fund managers can’t use their discretion to trade in and out of bonds as they see them going in and out of the index, but their mandate is to match the index performance.
I think it’s important to keep in mind that this index was invented in 1973. Prior to all of our computers and internet. Data was both more difficult to obtain and to analyze. If a person were to make up the ideal bond index today, the rules would probably not be these. In fact, people are trying to make something like “smart beta” style indices for bonds. It’s not simple. Bonds are not like stocks, and the people who invented this index back in the dark ages accomplished a rather remarkable feat. That doesn’t mean we have to continue to use it unchanged forever, either as a benchmark or as an investment guide.