Has long become longer or short become shorter? Evidence from a censored quantile regression analysis of the changes in the distribution of U.S. unemployment duration

There is conflicting evidence regarding the recent evolution of unemployment duration in the U.S. In this study we rely on censored quantile regression methods to analyze the changes in the US unemployment duration distribution. We employed the decomposition method proposed by Machado and Mata (2003)to disentangle the contribution of the changes generated by the covariate distribution and by the conditional distribution and adapted it to a duration analysis framework.The data used in this inquiry are taken from the nationally representative Displaced Worker Survey of 1988 and 1998. We provide evidence that the unemployment duration distribution shifted leftward. The main driving force behind that shift was the sharp leftward move in the unemployment rate distribution. This force was partially counteracted by the ageing of the displaced population, the striking absence of impact from being displaced via a plant shutdown, and the higher sensitivity of unemployment duration to unemployment rates.