The yield curve and the macro-economy across time and frequencies

Abstract

This paper assesses the relation between the yield curve and the main macroeconomic variables in the U.S. between early 1960s and 2009 across time and frequencies, using wavelet analyses. The shape of the yield curve is modelled by latent factors corresponding to its level, slope and curvature, estimated by maximum likelihood with the Kalman filter. The macroeconomic variables measure econmic activity, unemployment, inflation and the fed funds rate. The cross wavelet tools employed - coherency and phase difference - , the set of variables and the length of the sample, allow for a thorough appraisal of the time- variation and structural breaks in the direction,intensity,synchronization and periodicity of the relation between the yield curve and the macro-economy. Our evidence establishes a number of new stylized facts on the yield curve-macro relation; and sheds light on several results found in the literature, which could not have been achieved with analyses conducted strictly in the time-domain(as most of the literature)or purely in the frequency-domain.