Seniority Differentials in High Yield Bonds: Evolution, Valuation, and Ratings

Almost 20 years ago, one of the coauthors of this article published a study that reported finding systematically wider yield spreads on senior corporate bonds than on subordinated bonds with the same credit rating, but issued by different companies. The study also showed that this difference in spreads did not represent a market “anomaly” or failure to price risk correctly, but instead reflected differences in the actual, and hence the expected, loss rates of the securities. And such differences were in turn shown to stem from the practice of the rating agencies—which was abandoned about ten years ago—of rating a given issuer's subordinated debt two “notches” below that of its senior debt. Partly in response to this finding, all of the major agencies modified their use of this “two‐notch” convention by initiating in‐depth fundamental analysis of subordinated issuers on a case‐by‐case basis. In the meantime, the near disappearance of subordinated debt in the high yield market since the global financial crisis and its partial replacement by secured debt has furnished the authors of this article with a seemingly related “anomaly” to explore—namely, the tendency of secured bonds to have higher yields than samerated senior unsecured bonds. As in the earlier study of the senior‐subordinated puzzle, the authors' analysis confirms that the market has been properly pricing the relative risks of the different securities by showing that the actual loss rates of the secured issues have been systematically higher than those of like‐rated senior unsecured issues. The clear suggestion of these findings, as in the case of the earlier study, is that those investors who have chosen to incur the costs of analyzing expected loss rates instead of relying solely on the ratings have been rewarded for their efforts. And if the past is a guide to the future, this article may also succeed in spurring the rating agencies to make further refinements to their methods.