The Time-Varying Leading Properties of the High Yield Spread in the United States

We propose a comprehensive empirical examination of the time-varying leading properties of two high yield spreads in the United States and compare them with the leading properties of the term spread between the mid-1980s and the end of 2011. We show that high yield spreads are not reliable predictors of future economic activity, as measured by real gross domestic product and industrial production. Their predictive content for economic growth, statistically and economically signi ficant between the end of the 1980s and the beginning of the new century, vanishes in the second half of the 2000s. This disappearance is coincident (i) with structural breaks in the relationship that largely occurred in the early years of the past decade and during the 2007-2009 fi nancial crisis and (ii) with the reappearance of the leading properties of the term spread in recent years. In general, high yield spreads outperform the term spread at predicting economic growth at horizons up to one year