摘要:This paper deals with the analysis of US unemployment by means of I(d) techniques and using two different measures, the unemployment rate and the initial claims at different data frequencies. The results indicate that the unemployment rate series are I(d) with d constrained between 0.5 and 1, while the initial claim series display orders of integration close to 1. Thus, mean reversion is only obtained in case of the unemployment rate while the hysteresis hypothesis seems to be supported by the claimant counts.