We apply censored quantile regressions to representative German register data with more than 91,000 observations in order to determine crucial factors that influence the distribution of unemployment duration in West Germany during the 1980s and 1990s. We find that the effect of some regressors varies and has different sign depending on the quantiles of the unemployment duration distribution – a violation of the classical proportional hazard assumption which is very popular in unemployment duration analysis. We also find that variables reflecting the (un-)employment history of an individual such as the length of tenure, recall to the same employer in the past, recent unemployment and the position in the population income distribution before unemployment have the strongest effects on unemployment duration. We conclude that work history variables are most suitable in characterizing the job search behavior of an individual. The macroeconomic environment and the educational degree seem to have a limited effect only.
censored quantile regression unemployment duration register data