Counterfactual Analysis Using Censored Duration Data

We propose standardization techniques for the duration distribution in a population with respect to another taken as standard using right censored data, which forms a basis for counterfactual comparisons between distributional features of interest. Alternative standardizations are based on either a proportional hazard semiparametric specification or a nonparametric specification of the underlying conditional distribution. Applications to the restricted mean survival time and the hazard rate are discussed in detail. The proposal is applied to the counterfactual analysis of spells of unemployment duration gender gaps in Spain between 2004-2007. The behavior in small samples is investigated using Monte Carlo experiments.

Blogs Publicados