Those underperforming bond funds

By Felix Salmon
August 22, 2009
SPIVA report -- by far the best comparison of fund performance to underlying indices -- nearly all of those bond funds have underperformed their indices:

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Most investors have a significant exposure to bond funds. But according to S&P’s latest SPIVA report — by far the best comparison of fund performance to underlying indices — nearly all of those bond funds have underperformed their indices:

bondfunds.tiff

In case you can’t see this clearly, it says, among other things, that over the five-year period ending June 2009, 92% of investment-grade long funds have underperformed the index, 98% of mortgage-bond funds have underperformed the index, and 94% of general muni funds have underperformed the index. If you bought a California or New York muni fund, you had a 100% chance of underperforming the index.

Sam Mamudi, reporting on these results for the WSJ, puts them in the context of “the debate over passive versus active mutual-fund investing” — but he leaves out a crucial piece of information: passive investing in bonds is non-trivial. If you look at the performance of equity funds against the S&P 500, say, you can do so in the knowledge that it’s pretty trivial to buy an index fund or an ETF which will give you something very, very close to the performance of the index. With these bond indices, however, that’s not the case.

There are some bond ETFs: iShares, for instance, has a pretty broad suite of them. Most of them are pretty young, having been launched in 2007 or later, but there are a couple of older ones, such as the 20+ year Treasury bond fund (TLT). If you look at the information iShares provides, however, it’s really hard to tell how good the fund is at replicating the return of the index. We know that on June 30 its net asset value stood at 94.62, up from 83.41 five years previously; we can also find out that over that time the fund paid out 20.435 in dividends. You don’t want dividends, or you find them expensive to reinvest? Tough luck. We also know that the benchmark index increased from 237.77 to 388.29 over that time.

The index, then, increased over the course of those five years by an annualized 10.3%. If you just add the dividends onto the net asset value for the fund, you get an increase from 83.41 to 115, or 6.6% annualized. That’s pretty much in line with the 6.94% weighted average annualized return for long government bond funds generally.

What happens if you run the same exercise for IVV, iShares’ S&P 500 index fund? The index went from 1661.53 to 1498.94. The fund, meanwhile, went from 113.26 to 92.24 with 12.3185 in dividends. The index’s annualized return was -2%, while adding the dividends to the NAV of the fund gives an annualized return of -1.6%. The index fund outperformed the index, on this methodology, partly because you weren’t reinvesting dividends in a falling asset, but also because dividends are lower on stocks than on bonds, even government bonds.

But all of that is just doodling, really. The important thing to remember when it comes to bond funds is that realistically you should only be comparing managed bond funds to index bond funds, rather than to indices directly. After all, your choice isn’t between investing in a managed bond fund and investing in an index, it’s between investing in a managed bond fund and investing in an index bond fund. So while it’s true that a startlingly high percentage of managed bond funds underperform their index, it’s not necessarily true that the same percentage of those funds underperforms an index fund linked to the same benchmark.

Update: Wcw, in the comments, points me to this document, in which iShares says that the five-year return to end-June 2009 of the TLT fund was 7.25%, while the index return was 7.32%. The return on IVV was -2.28%, while the return on the S&P 500 was -2.24%. Even the TIPS bond fund, which Pimco disparages so, returned 4.80% to the index’s 4.94%. I would assume, of course, that these returns assume no commissions or bid-offer spreads when buying or selling the funds, or reinvesting dividends.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
5 comments so far

Hmmm. I’m sure I’m missing something, because I read your blog all the time and you know a ton more about bonds than I do. But…

According to Vanguard’s website, their Intermediate Term Treasury Index fund has returned within a few bps of their Barclay’s benchmark over the life of the fund. They show at most ten or twenty bps of shortfall.

Obviously for foreign/emerging market bonds, passive investing is tough to impossible, but for US bonds, I thought passive options were numerous and straightforward and could be counted on to track their benchmark +- a few bps for expenses.

ETFs are 40-Act funds in disguise, which means they have to present you standardized performance — and they do. See http://us.ishares.com/content/stream.jsp  ?url=/content/repository/material/perfo rmance_report/monthly_performance_report .pdf&mimeType=application/pdf The ishares AGG over 5 years is 17 bps behind its index over 5 years, and TLT is 7 bps behind (Treasuries are easier to trade than corporates). Yes, bonds are harder to index. But it can be done, and BGI is doing it.

Full disclosure: I know a number of people who work for BGI, and I like them.

Right, Felix. There aren’t many, if any, true index funds for bonds — managers try to replicate the index, because one can’t always source the exact bonds in the index, so a manager buys brothers and cousins, and attempts to match the major risk factors of the index, taking small tactical mismatches where he dares.

Expenses are the main difference in long term for bond funds. Security selection is secondary.

One other note: during bull phases, active managers and enhanced indexers tend to beat the indexes — research that neglects big bear phases such as we have just had gets trotted out to justify fees.

Unless you are really clever, and don’t have a constrained mandate, it is very hard to beat the indexes. Fuss and Gross do it, but few others.

I would think it is nearly impossible to beat an index that doesn’t have a realistic expense structure built into it.

Any bond fund that beats its index is likely to engage in:
a.) Madoff’ing the books; or
b.) Taking on additional risk.

Posted by Brad Ford | Report as abusive

If high frequency trading programs are front-running big bond funds, and they doubtless are since this is a massive market to exploit, then this would help explain their across-the-board below-index performance.

Posted by Dan | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/