Swedroe: Active Fails In Fixed Income

March 02, 2018

There is a myth that active bond fund managers want and need you to believe. It goes something like this: “Sure, active stock picking isn’t likely to work, but in fixed income, active management really shines.” This is a strange argument to make, because:

  1. Active management is a zero-sum game before expenses (and a negative-sum game after expenses). Thus, you have to be able to exploit the mistakes of others to generate alpha.
  2. Even more so than in equity markets, trading activity in bond markets is dominated by large, institutional investors. Thus, it’s hard to identify a likely supply of the necessary victims to exploit.
  3. The academic research has found that the vast majority of returns to fixed-income portfolios are well-explained by two common factors: term (duration) risk, and credit (default). In other words, any outperformance by active managers against benchmark indexes is likely to come from simply having greater exposure to these two common factors—not from security selection or market timing (two sources of true alpha).

With the sharp cyclical decline we have experienced in interest rates since the global financial crisis, active managers could claim to outperform a benchmark by constructing a portfolio with more duration than the declared benchmark.

Of course, you don’t need active management to have longer duration. Just increase your allocation to longer-duration indexes and invest in lower-cost, passively managed mutual funds and ETFs. The term premium (the difference in returns between long-term U.S. government bonds and one-month Treasury bills) from January 2009 through December 2017 was 4.1% per year. In contrast, over the prior period 1926 through 2008, the term premium was 2%.

The same is true for credit risk, which also has been well rewarded since 2009. The credit premium (the difference in returns between long-term corporate bonds and long-term U.S. government bonds) from January 2009 through December 2017 was 3.1% per year. In contrast, over the prior period from 1926 through 2008, the credit premium was 0%. The default premium over that same period was about 11.6% per year (based on the Bloomberg Barclays US Corporate High-Yield Index).

Again, if you desire exposure to credit risk, you can obtain it in lower-cost, passively managed mutual funds and ETFs. In other words, you don’t have to pay the typically high fees of active management to gain exposure to these factors.

Making matters worse is that the primary reason for investing in fixed income should be to dampen the overall risk of the portfolio to an acceptable level, and while safe bonds offer effective diversification from the equity risk that dominates most investors’ portfolios, credit risk is more highly correlated with equities. The lower the credit rating, the higher the correlation becomes.

The Evidence

Fortunately, we have plenty of evidence to help dispel the myth that active management is likely to generate alpha in fixed-income markets. The most recent comes from S&P Dow Jones Indices, which has published its S&P Indices Versus Active (SPIVA) scorecards, comparing the performance of actively managed mutual funds to their appropriate index benchmarks, since 2002. The 2017 midyear report, the latest available, includes 15 years of data on the performance of active bond funds.

Following is a summary of the results:

  • The worst performance was in long-term government bond funds, long-term investment-grade bond funds and high-yield bond funds. In each case, just 2% of active funds beat their respective benchmarks. On an equal-weighted (asset-weighted) basis, long-term government bond funds underperformed by a shocking 3.5%age points (3.0%age points), long-term investment-grade bond funds underperformed by 2.6%age points (2.2%age points) and high-yield bond funds underperformed by 2.3%age points (1.7%age points).
  • For domestic bond funds, the least poor performance was in intermediate- and short-term investment-grade funds, where 76% and 71% of actively managed funds underperformed, respectively. On an equal-weighted basis, their underperformance was 0.5 percentage points and 0.7 percentage points, respectively. However, in both cases, on an asset-weighted basis, they outperformed by 0.3 percentage points. This result possibly could be explained by the funds having held longer maturities (that is, taking more risk) than their benchmarks.
  • Active municipal bond funds also fared poorly, with between 84% and 92% of them underperforming. On an equal-weighted (asset-weighted) basis, the underperformance was between 0.6 percentage points and 0.7 percentage points (0.2 percentage points and 0.4 percentage points).
  • Actively managed emerging market bond funds fared poorly as well, as 67% of them underperformed. On an equal-weighted basis, their underperformance was 1.4 percentage points. On an asset-weighted basis, their underperformance was 0.4 percentage points.

We even have evidence from a 1993 study, which covered a period when, arguably, the markets were less efficient than they are today. Christopher Blake, Edwin Elton and Martin Gruber, authors of the study “The Performance of Bond Mutual Funds,” which appeared in the July 1993 issue of the Journal of Business, examined the performance of 361 bond funds and found that the average actively managed bond fund underperforms its benchmark index by 0.85% per year.

In addition, Kevin Stephenson, author of the study “Just How Bad Are Economists at Predicting Interest Rates?”, which appeared in the Summer 1997 issue of the Journal of Investing, found that only 128 out of 800 fixed-income funds (or 16%) beat their relevant benchmark over the 10-year period he covered.

Is Past Performance Predictive?

Being a loser’s game does not mean there aren’t some winners, offering the hope that somehow we can identify the few winners ahead of time. Unfortunately, the evidence suggests that believing this is possible in a reliable way would be the triumph of hope over experience.

For example, in his 1994 book “Bogle on Mutual Funds,” John Bogle studied the performance of bond funds and concluded that “although past absolute returns of bond funds are a flawed predictor of future returns, there is a fairly easy way to predict future relative returns.”

He continues: “The superior funds could have been systemically identified based solely on their lower expense ratios.” Other studies on the subject, including those on municipal bond funds, all reach the same conclusions:

  • Past performance cannot be used to predict future performance.
  • Actively managed funds do not, on average, provide added value in terms of returns.
  • The major cause of underperformance is expenses; there is a consistent one-for-one negative relationship between expense ratios and net returns.

The results of a December 2004 study by Morningstar’s Russel Kinnel, “Time to Clear Out the Dead Wood,” further demonstrate the importance of costs and that the past performance of actively managed funds is a poor predictor of future performance. Kinnel tested funds with strong past performance and high costs against those with poor past performance and low costs. He writes: “Sure enough, those with low costs outperformed in the following period.”


Latest Research

AQR Capital Management contributes to the literature with their December 2017 study, “The Illusion of Active Fixed Income Diversification.” The firm’s researchers studied the performance of institutional fixed-income managers from 1997 through 2017.

AQR extracted monthly manager and benchmark returns that belonged to three categories: global aggregate, core-plus (U.S. aggregate credit index-benchmarked portfolios with allowance for out-of-benchmark exposures) and unconstrained bond (go-anywhere) funds. They limited themselves only to funds within a category that had a benchmark clearly mirroring the category. In addition, they required the base currency to be U.S. dollars.

For global aggregate managers, they were able to find 89 bond funds in their database, and ended up with 53 funds that had the global aggregate index benchmark as well as sufficient returns data for the analysis. For core-plus managers, the same filtering criteria yielded 115 funds. For unconstrained bond managers, it yielded 27 funds.

The funds they eventually used represented 70% of the number of eligible funds available on eVestment in these three categories (and 69% of the assets under management).

Following is a summary of their findings:

  • On a returns basis, active fixed-income managers have outperformed their benchmarks. However, the majority of active returns for fixed-income managers can be explained by exposure to credit markets, not security selection or market timing. For example, active returns for all three categories have a very strong correlation with high-yield excess returns: 0.76 for global aggregate, 0.95 for core-plus and 0.82 for unconstrained bond.
  • Credit tilts are consistently positive and do not vary significantly over time. Inasmuch as managers adjust their exposure to credit, they tend to vary between long and very long (they tend not to take short positions). To the extent that there are deviations in credit tilts, they only weakly predict future credit returns. In other words, there is scant evidence of timing ability.
  • Most active fixed-income returns are a result of the credit risk premium, which is related to the equity risk premium. The resulting diversification loss can dampen the risk-adjusted performance of an investor’s overall portfolio. And this has been the case.

These findings led AQR to conclude that “a significant portion of [fixed income] manager active returns comes from being overweight, structurally and permanently, sources of return that are highly correlated with [high-yield] credit.”

They also warned that “there is a downside of this effective credit overweight to overall strategic allocations, namely a reduction in overall portfolio diversification.”

AQR’s researchers showed that, while over their full sample the U.S. aggregate credit index has tended to provide excellent diversification to equities, realizing a -0.33 correlation with the S&P 500 Index, an equal-weighted portfolio of core-plus managers has actually realized a correlation to equities of 0.05 over the full sample, changing the sign of the correlation of fixed-income returns to equity returns from negative to slightly positive.

This effect has been even stronger since 2008, when active fixed-income managers across categories have tended to have an even higher effective credit exposure. During this time period, the correlation of U.S. aggregate credit index returns to the S&P 500 Index was -0.22, while an equal-weighted portfolio of core-plus managers had realized a correlation of 0.33.

Making matters worse is that the correlation tends to turn strongly positive at exactly the worst time, when equities are crashing, increasing portfolio drawdowns.

Bottom Line

The bottom line is pretty simple: There is an overwhelming body of evidence that active management in fixed-income markets is even more of a loser’s game than it is in equity markets.

The winning strategy for investors is the same for both stocks and bonds: Decide how much exposure you desire to common factors, and then implement the appropriate allocations using the lower-cost, passively managed mutual funds and ETFs that give you the most effective exposure (considering not only expense ratios but also fund construction rules and trading strategies) to those factors.

Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.

Find your next ETF