(This is an updated version of a blog that appeared previously. The charts in the earlier version don't appear in this new version.)
Investors considering the Vanguard Total Bond Market ETF (NYSEArca:BND), the iShares Barclays Aggregate Bond Fund (NYSEArca:AGG), the Schwab U.S Aggregate Bond ETF (NYSEArca:SCHZ) and the SPDR Barclays Capital Aggregate Bond ETF (NYSEArca:LAG) may look at the index descriptions of the various funds and wonder:Why would I choose anything but the cheapest fund in the group?
After all, the funds track the same index. They should all provide the same performance, right?
Well, not necessarily. Bond funds are an entirely different animal than their equity brethren.
The indexes they track often have thousands of constituents, many of which are either highly illiquid or impossible to find altogether. Furthermore, some fund managers will own bonds that aren't included in the index but whose return and duration characteristics are deemed similar to those included in the index.
The benchmark for these four funds, the Barclays Capital U.S. Aggregate Index, is a great example of this. The index has 7,854 bonds, many of which are either hard to secure or highly illiquid. It's nearly impossible to fully replicate the index, and anyone attempting to do so would have such a high expense ratio that a few basis points of better tracking would be a drop in the bucket.
So all of the funds optimizeâholding some, but not all, of the securities in the index, in an attempt to match the index's overall return. And while optimization is a necessary evilâfrom a tracking perspectiveâit's employed to varying degrees. So, a quick analysis is in order.
Optimization By Degree:From SCHZ To BND
On one extreme of the optimization scale you have SCHZ, which holds only 414 bonds. That's just 5 percent of the securities in the Barclays Index. At the other extreme is BND, which holds 4,991 issues, or upward of two-thirds of the credits in the index. Then there are the middleweights, LAG and AGG, which hold 603 and 1,260 securities, respectively.
Ultimately it's up to individual investors to decide how much optimization they are willing to stomach, and at what tracking cost. In the case of SCHZ, you get a segment-best expense ratio at just 10 basis points, but you also get the most aggressive optimization strategy.
Is 1 basis pointâthe difference in price between BND and SCHZâenough to justify owning 58 percent less of the underlying index? If so, SCHZ is a fine choice for investors looking to shed costs.
That said, a closer look at the relative returns of each fund paints an entirely different picture.
The table below shows the number of holdings for each fund, the percentage of the benchmark they are optimizing, and each fund's expense ratio. On the right, it shows what the total return was for each fund, and how that compared to the total return of the index over the same period of time. Expected returns of each of the funds, all else equal, are the returns of the index minus each fund's expense ratio.
Index Return -ExpensesFor The Period
Does it matter?
One would expect each fund, in a perfect world, to provide the index's return minus its expense ratio.
As you can see, that hasn't been the case. In a five-month periodâthat's the longest stretch possible, as SCHZ is a new fundâthere's been an 18 basis point swing from the best to the worst performer. That may not sound like much, but it's larger than the 12 basis point difference between the lowest and highest expense ratio.
Five months isn't enough time to make definitive judgments, but even over the past year, the three funds that have been in existence in this period exhibited quite different patterns of returns. Specifically, in the past year, LAG has outperformed the benchmark slightly, while BND and AGG have lagged.
If your head is swimming at this point, it won't be helped by looking at the funds' yields.
These are, after all, fixed- income products, and many investors will start with the income yield when evaluating funds. According to Bloomberg, the 30-day SEC yield for the funds are 2.38 percent for BND, 2.10 percent for AGG, 2.38 percent for LAG and 1.55 for SCHZ. If that spread is too wide to believe, that's because it is:SEC yield is one of the industry-standard ways of evaluating yield, but like all yield measures, it is subject to statistical anomalies. That's likely what we're seeing here.
What all of this means is that these funds are extremely complex and have many moving parts. Throw in the fact that there is no consensus pricing on bonds, and it's clear the investment decision is multifaceted.
As my colleagues pointed out recently, investors seem to get this dynamic. BND is now the largest fund by assets in the segment, which suggests that full(er) replication is something that investors are looking for.
If investors were solely focused on expense ratio, SCHZ would be gaining assets against peers like AGG and BND. It seems something else is at play here.
It's important to note, of course, that tracking error can go both ways. SCHZ was the worst-performing fund in the group based on total return NAV over the past five months, and BND was the best.
That could just as easily reverse in the future. But if the level of optimization stays the same, one thing should happen consistently:SCHZ should show more variance against its index than BND.
This blog isn't intended to be a comprehensive discussion of bond fund analysis. A host of other factors influence the attractiveness of a bond fund, from the tradability and liquidity of the fund itself to premiums and discounts, and the timing and management of capital gains distributions and so on. Those are great topics for future blogs.
But the overarching theme is obvious:approach bond ETFs with caution.
Expense ratio alone isn't enough to evaluate these products. Carefully consider the index tracked, the optimization strategy, portfolio management style, expenses and potential capital gain before you buy.
Copyright ® 2011 IndexUniverse LLC . All Rights Reserved.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.