According to the capital asset pricing model (CAPM), riskier securities should yield higher returns. The authors assess distressed corporate bond portfolios and conclude that the CAPM holds under a theoretical time-independent framework in that market when overall volatility is not controlled for and in a more practical time-dependent framework where investors are assumed to possess market-timing skills. Investors are better off, however, using a buy-and-hold strategy and investing in low-volatility distressed securities.
The US high-yield bond market has increased significantly over the past two decades (from $214 billion in 1990 to $1.21 trillion in 2012). Part of this market includes distressed securities, which are securities that have not defaulted but whose current yield exceeds the equivalent US Treasury bond yield by at least 10%. The size of the distressed debt market changes with economic circumstances, increasing during times of stress.
The authors use the option-adjusted spread (OAS) to categorize distressed corporate bonds as low or high volatility and then apply several types of analyses to assess which bond group outperforms. They conclude that, in practice, unless market-timing skills can be applied correctly, investing in lower-volatility securities is better for buy-and-hold investors over the long run.
How Is This Research Useful to Practitioners?
Although the capital asset pricing model (CAPM) is widely known in the investment world, its strengths and limitations are not always well-defined and understood. The authors conclude that, contrary to what the CAPM suggests, lower-volatility distressed debt portfolios provide higher returns. They also show how to use OAS volatility to construct better-performing buy-and-hold portfolios.
Using a time-independent analysis, they create portfolios from different market circumstances by equally weighting securities that pass the criterion for distressed debt. The timing of distress is adjusted so that all securities in the portfolio become distressed at the same time. When this time-independent case is not adjusted for overall market volatility, the highest-volatility portfolio outperforms the lowest-volatility portfolio, fulfilling CAPM expectations. When a time-dependent model is used, in which the portfolios are updated continuously as securities become distressed, the lowest-volatility portfolio outperforms because of lower default rates and higher terminal values. The authors normalize their dataset to ensure that the overall market volatility does not hide the volatility characteristics of individual bonds. They conclude that the lowest-volatility bonds perform better in both time-independent and time-dependent scenarios.
The rare investors who possess market-timing skills can increase their returns by investing in high-volatility securities after market dislocations (e.g., the dislocations in the US market in 1998, 2002, or 2008) and then rebalancing to low-volatility securities before economic conditions start to deteriorate.
How Did the Authors Conduct This Research?
Because of the increased size of the high-yield market, research into this topic has increased over the last three decades. Vasicek (KMV 1984) measured probabilities of default (PDs) and viewed firm equity value as a call option. Altman (Institute of Chartered Financial Analysts 1989) used a capital-based model for deriving PDs based on a bond’s credit rating and duration; his method is still used by credit rating agencies. Iben and Litterman (Journal of Portfolio Management 1991) looked at term structures of credit spreads, and Duffie and Singleton (Review of Financial Studies 1999) explored credit derivatives.
The authors do not focus on PD itself but instead use volatility—in particular, the option-adjusted spread—as their criterion for identifying different risk–return portfolios. They use data from 1997 to 2012 from the Bank of America/Merrill Lynch US High-Yield Master II Index to extract securities that exceed the 10% spread threshold for at least two months. Bonds mature, default, or become non-distressed, and bonds can reenter the dataset. The authors use realized OAS volatility prior to the debt becoming distressed to create four distinct portfolios and examine the portfolios with the highest and lowest volatilities.
They define terminal value as a means of dealing with the eventual recovery rate of bonds that default and introduce the concept of annualized return distribution analysis to remove any market-timing effects for the time-dependent analysis.
Finally, they review regime-shifting techniques by changing the portfolios as economic conditions improve, deteriorate, or remain static. The authors show that there is little performance difference between high-volatility and low-volatility bonds in static markets, that low-volatility securities outperform high-volatility ones during deteriorating conditions, and that high-volatility portfolios outperform during improving conditions.
The authors provide an interesting view regarding the ideas behind and application of the CAPM—a model most investment professionals are familiar with. They clearly show the significance of the high-yield debt market and give a concise overview of the literature before outlining their process and the several methods they apply. Because their dataset includes some tumultuous periods in the market, it is not too surprising that the lower-volatility securities outperform the higher-volatility ones. Nevertheless, their conclusions about when the CAPM holds and does not hold apply to any dataset and are thus valuable.