Credit Spreads as Predictors of Real-Time Economic Activity: A Bayesian Model-Averaging Approach

We use Bayesian Model Averaging (BMA) to forecast real-time measures of economic activity using a large set of possible predictors. The set of potential predictors includes option-adjusted credit spreads—in addition to a large number of other asset market indicators—based on bond portfolios sorted by maturity and credit risk as measured by the issuer’s distance-to-default. The portfolios are constructed directly from the secondary market prices of outstanding senior unsecured bonds issued by a large number of U.S. nonfinancial corporations. Our results indicate that relative to a direct autoregression, BMA yields consistent improvements in the prediction of the growth rates of real GDP, industrial production, employment, and business fixed investment, as well as of the changes in the unemployment rate, at horizons from the current quarter (i.e., “nowcasting”) out to four quarters hence. The gains in forecast accuracy are statistically significant and economically important and owe almost exclusively to the inclusion of our portfolio credit spreads in the set of predictors—BMA consistently assigns a high posterior weight to models that include these financial indicators. JEL Classification: C11, C53