The Decoupling

Of all the mutual fund categories that have undergone changes since the 2008 financial crisis, few if any can lay claim to the combination endured by the intermediate-term bond category, home to most core offerings in the bond-fund universe. Back then a notable group of funds claiming just under $500 billion in assets, the category has mushroomed as investors have fled equity market volatility, and its size now registers at nearly $930 billion.

The group has done much more than simply grow, however. As we discussed in a piece this past July, fund managers have also meaningfully changed the way that they've been managing core portfolios. Most intermediate-term bond funds use the Barclays Capital U.S. Aggregate Bond Index as their benchmark, and one trend that was especially visible in 2011 was a willingness on the part of managers to run their portfolios with less interest-rate sensitivity than that bogy. Many were worried that, even after a late 2010 spike, Treasury yields were still too low and that a nascent economic recovery would result in a yield spike that would push down the prices of longer-duration portfolios. The economy proved less robust than expected, however. At the same time, fears of contagion from Europe's problems drove down the prices of riskier assets in the late summer, while triggering a massive Treasury bond rally that saw the 10-year note go from 3.65% in February to levels around 2% in August (and ever since). Fewer than 12% of funds in the category managed to match or beat the return of the Barclays Aggregate Index for the year.

Even more has changed under the hood of the average core bond fund, though. On the heels of the financial crisis, many managers went bargain hunting for beaten-down bonds of all kinds that were too tempting to pass up. In many cases, that meant bulking up in areas of the market that are either lightly represented in the Barclays Aggregate benchmark--such as commercial mortgages--or those that don't make it into the index at all, such as nonagency mortgages and high-yield corporate bonds.

What's particularly notable, however, is that the shift away from the benchmark has been sustained, even after many of these sectors have returned to more normal levels of trading and pricing. One way we've been able to observe the shift is by looking at historical correlations between funds and the Barclays Aggregate benchmark. In particular, the R-squared statistic is a useful way of examining how much of a fund's returns over a given period can be explained by the returns of the index. The scale for R-squared statistics runs from zero to 100, with the lower bound suggesting no correlation to the benchmark while a value of 100 implies that 100% of a fund's returns can be explained by returns of the index.

In order to get a feel for how much things have changed, we went back and calculated R-squared data for the average fund in the intermediate-term bond category on a rolling 36-month basis, and began the exercise with the period ended December 2006. What it shows is fairly dramatic: R-squared figures for the average fund in the group began dropping in 2008, much as one would have expected given the opportunistic buying that occurred as the market swooned. But while the figure leveled out in the low 70s starting in early 2009, it never returned to the formerly high--that is, very close to 100--levels that it occupied in years prior to the crisis. Moreover, the statistic began to fall even further beginning last summer. On a trailing three-year basis ended February 2012, the average fund in the intermediate-term bond category now carries an R-squared of roughly 62. To be sure, one can come up with different numbers depending on how time periods are sliced and diced; the period beginning 2009 was particularly unusual because of the aforementioned opportunistic buying after the crisis, and that time period still factors in to the category's three-year history. But while shorter, more-recent periods may not produce as dramatic a picture, they still reflect a divergence from the index.

The phenomenon is not limited to a group of small outlier funds, either. A survey of the largest funds in the category shows that several actively managed funds that previously bore relatively high R-squared figures before the crisis have since dropped markedly.

A perusal of their portfolios, meanwhile, helps illustrate why so many of these funds are acting differently from the index. The most well-known example is  PIMCO Total Return (PTTRX), which, while famously short of its benchmark and bereft of Treasury bonds at some points during 2011, has also distinguished itself by branching out more aggressively to sectors such as high-yield, non-U.S. developed- and emerging-markets bonds, and even some municipal debt. Other high-profile funds, such as  Metropolitan West Total Return Bond (MWTRX), have always been a bit more willing to deviate from the benchmark but have been even more intrepid since the crisis. That portfolio, for example, has a large exposure to nonagency mortgages and also holds an overweighting in commercial mortgage-backed securities.

It's not just the traditionally adventurous who have taken formerly less-traveled pathways. Vanguard's intermediate-term investment-grade fund has historically fashioned itself a mostly corporate-bond portfolio but still had a high correlation to the benchmark prior to the financial crisis. It, too, holds 7% in asset-backed securities and 4.4% in commercial mortgages, for example, and its R-squared in relation to the Barclays Aggregate has dropped to 54 from 98 prior to the financial crisis.

Managers have described a variety of reasons for deviating from the benchmark, but a central theme has to do with the rock-bottom yields currently being paid by Treasury bonds. There is clearly some concern among managers about an eventual spike in rates, but the general consensus seems to revolve as much around the fact that Treasuries don't appear to offer much return potential at this stage, either from capital appreciation--which would require yields to fall even further--or from income. The latter is a crucial point in the context of afterinflation returns, in particular. At current levels, many Treasury bond yields won't be sufficient to overcome, much less keep up with inflation at its current run rate.

One can reasonably argue that it isn't only fund managers straying from the index that's been causing this phenomenon, but that the index, too, is straying from them to some degree. That owes much to the fact that increased Treasury bond issuance has consumed a larger percentage of the benchmark. That allocation numbered just below 25% at the end of 2006, for example, but clocked in at 35% by the end of February 2012. Conversely, while index rules have long kept exposures to commercial mortgages and asset-backed securities relatively modest, they've dropped to a combined 2.25% of the index as of February, down from more than 6% at the end of 2006.

Whatever the root cause of the shift, however, the differences have real implications for how investors evaluate their funds. On a relatively basic level it means that other so-called Modern Portfolio Theory statistics such as beta and alpha--which are often used as proxies for a fund's volatility and outperformance, respectively--are much less useful. That's because they only tend to be as statistically meaningful as the benchmark is correlated to the investment in question. The lower a fund's R-squared, therefore, the less useful are the rest of its stats.

That raises the broader issue of benchmarking in general. Generally speaking, a benchmark index ought to contain all of the securities in a manager's investable universe. That the Barclays Aggregate hasn't really lived up to that need for some time isn't really news, of course. What's clearly different today, though, is that we appear to be transitioning into a new phase in which even broad outlines of the index's performance behavior are proving less useful as signposts than they once were. There's no easy solution to this problem. There's little indication that Wall Street is anxious to begin work on building broader, more well-represented indexes that better match what fund managers are doing. For all of their importance, meanwhile, benchmarks are not necessarily designed and operated for the purposes or benefit of the end investor. Rather, they tend to reflect what investment banks find are the demands of their direct customers--many of whom are fund managers--and what it is that they're trying to sell to them.

Ultimately, though, the most pressing concern may be one of understanding portfolio risk. The good news is that the handful of catastrophes and close calls that occurred during the financial crisis appears to have encouraged many firms to take a second and third look at the risks they take in portfolios and how to try to make sure they don't spin out of control should we again experience a market disruption on the scale of 2008. Yet with so many funds branching out of their former comfort zones and investing in areas that once occupied little or no space in their core portfolios, the risk that investors will get something other than they bargained for has inevitably gone up.