Asymmetry in the bivariate relationship between output and interest rates

This paper investigates whether an asymmetry is present in the Granger-causal relationship between output and a set of interest rates and their spreads, across expansionary and contractionary business cycle phases in post 1950 U.S. Non-structural VAR models of monthly industrial production and three interest rates and four spreads are estimated for expansion and contraction samples. This study finds asymmetry in the bivariate relationship between the output and the financial variables across the two samples. Most of the interest rates and the spreads that were observed to Granger-cause output in the full sample continues to do so over the expansion sample but lose their predictive power in the contraction samples. ; Earlier version: Asymmetry in the relationship between output and interest rates (Paper no. 93-13)