The illusion of alpha in active bond management

By LARRY SWEDROE

Over the past 20 years, active fixed income (FI) managers have tended to deliver returns in excess of their benchmarks. This has generated a popular notion that while active investing in equities is a loser’s game, active bond management is a winning strategy. Before you jump to that conclusion, it’s important to understand that active fixed income managers may be outperforming their chosen benchmarks simply because they have higher exposures to well-documented factors. Such factors include duration and credit. That’s not alpha, just another form of beta (exposure to a common factor).

Jordan Brooks, Tony Gould and Scott Richardson contribute to the literature on active management in fixed income markets with their study Active Fixed Income Illusions, which was published in the Spring 2020 issue of The Journal of Fixed Income.

Their data sample included 454 U.S. Aggregate benchmarked managers, 53 Global Aggregate managers and 158 unconstrained bond managers. It covered 82%, 35% and 90% of the assets under management across the three categories, respectively, as captured in the eVestment database. The study covers the period March 1998 through September 2019.

The researchers required at least five years of returns data and retained only USD denominated funds funds that have a benchmark that reflects that category. For example they removed Global Aggregate and U.S. Aggregate funds where the tracking error exceeds the volatility of the benchmark itself.

How managers can take active risk

They began by noting that there are several dimensions across which FI managers can take active risk:

— Security selection within government bonds, securitised assets, and corporate bonds included in the benchmark.

— Asset allocation across FI sectors (e.g., duration timing or sector rotation toward or away from the credit/ spread risk embedded in corporate bonds or the prepayment risk embedded in asset-backed securities).

— Out-of-benchmark tilts, including riskier high-yield corporate bonds, corporate loans, CLOs, inflation-linked securities, emerging market debt, and/ or non-agency mortgages.

Little evidence of manager skill

Brooks, Gould and Richardson adjusted the performance of the active FI managers for exposure to an economically intuitive set of traditional bond risk premia — including duration, corporate credit, emerging markets and volatility risks. Following is a summary of their conclusions:

— Across multiple categories of active FI managers, there is little evidence of manager skill — either in aggregate or individually.

— While active FI funds outperform their benchmarks, traditional bond risk premia explain most of the active returns across all three FI manager categories — the outperformance isn’t alpha. Instead its source is a different beta (exposure to different risks than the self-declared benchmark).

— Across all FI categories, a large portion of active FI manager returns can be explained by exposure to credit markets. Exposure to emerging market currency risk also explains returns of active FI managers.

— Despite the common perception that adding excess returns in FI markets is relatively easy, active returns of FI managers largely represent a repackaging of traditional risk premia — FI alpha may be merely an illusion.

— The increased exposure to traditional bond risk premia reduced diversification benefits. For example, over the sample period, the correlation of U.S. Aggregate returns to the S&P was -0.19, while an equal-weighted portfolio of U.S. Aggregate active managers had a positive correlation of +0.15.

Don’t pay active fees for passive exposures

The authors stated: “The repackaging of traditional risk premia appears to be a pervasive phenomenon in active management, and investors should be wary of paying active management fees for passive exposures.”

In other words, just as an equity manager can outperform a market benchmark that has a beta of 1 by building a portfolio with a beta of 1.5 (taking 50 percent more risk), FI managers may tilt their portfolios to higher risk and higher yielding segments of the FI markets.

In either case, active returns are simply a manifestation of traditional risk premia. It’s not alpha, just beta, which can generally be obtained at a lower cost through passive, transparent, systematic strategies.

Increased correlation with equities

Brooks, Gould and Richardson also noted that the fact that a large portion of returns to active managers was explained by exposure to credit risk “is hardly comforting for an investment into an asset class that is meant to provide diversification from equity markets”. The exposures to credit and currency risk increased the correlation with equity returns, a risk that raises its ugly head at the wrong times.

For example, during the fourth quarter of 2008 in which U.S. equities returned -23%, U.S. Aggregate returns were 4.4%, somewhat mitigating the equity drawdown experienced during the crisis. U.S. Aggregate FI manager returns, on the other hand, were roughly half that of the U.S. Aggregate Index (2.1%) due to the underperformance of credit. The authors noted: “During the quarter in which it was needed most, active FI diversification was elusive.”

“The true average alpha is zero”

These findings led them to conclude: “Stated simply, the distribution of alphas we observe within each category is broadly consistent with a world in which the true average alpha across managers is zero, and likely negative net of fees. Again, the inference is that the historically positive active FI returns are only an illusion of alpha — active FI investors were paying high fees for alpha but getting beta instead.”

These findings are consistent with those of Stephen Laipply, Ananth Madhavan, Aleksander Sobczyk and Matthew Tucker, authors of the study Sources of Excess Return and Implications for Active Fixed-Income Portfolio Construction, published in the 2020 Special Issue of The Journal of Portfolio Management. They analysed quarterly holdings data for a broad sample of fixed income mutual funds — a comprehensive dataset of 221 U.S. active investment-grade and high-yield mutual funds with more than $525 billion in assets in the period December 31, 2015, to June 30, 2018. They attributed excess returns to (1) static factor exposures, such as a structural tilt to duration, mortgage spreads and credit spreads; (2) time-varying factor exposures, such as varying duration over the cycle; and (3) individual bond security selection.

More efficient to employ index products

They found that, although bond funds in aggregate demonstrated positive alpha, a significant amount of their performance was driven by exposure to static factors as opposed to dynamic timing or security selection. The return contribution of security selection alpha is smaller than that of factor tilts across managers of all skill levels. The authors demonstrated how active portfolio managers can employ index products to more efficiently express factor views and help capture more excess return through reduced costs and frictions.

In summary, while active managers have outperformed their benchmark indexes, the evidence demonstrates it has not been due to either security selection or market timing. The alpha was an illusion, as it was explained by exposure to common factors that could be accessed at lower costs through structured portfolios that gain systematic exposure to those factors.