Credit Spreads and Real Activity

Abstract

This paper explores the transmission of credit conditions into the real economy. Specifically, I examine the forecasting power of the term structure of credit spreads for future GDP growth. I find that the whole term structure of credit spreads has predictive power, while the term structure of Treasury yields has none. Using a parsimonious macro-finance term structure model that captures the joint dynamics of GDP, inflation, Treasury yields and credit spreads, I decompose the spreads and identify the drivers of this transmission effect. I show that there is a pure credit component orthogonal to macroeconomic information that accounts for a large part of the forecasting power of credit spreads. The macro factors themselves also contribute to the predictive power, especially for long maturity spreads. Additional factors affecting Treasury yields and credit spreads are irrelevant for predicting future economic activity. The credit factor is highly correlated with the index of tighter loan standards, thus lending support to the existence of a transmission channel from borrowing conditions to the economy. Using data from 2006-2008, I capture the ongoing crisis, during which credit conditions have heavily tightened and I show that the model provides reasonably accurate out-of-sample predictions for this period. As of year-end 2008, the model predicts a contraction of -2% in real GDP growth for 2009, which is lower than comparable survey forecasts.

Keywords: term structure, credit spreads, forecasting, macro-finance, no-arbitrage model, credit conditions

JEL Classification: E43, E44, E47, G12