Application of a Simple Nonparametric Conditional Quantile Function Estimator in Unemployment Duration Analysis

In many econometric applications it is unclear from the very beginning whether a parametric functional of a continuous regressor should be specified as a linear, as a higher order polynomial or as a piecewise linear. Nonparametric estimators can provide relevant information as they are a convenient tool for data exploration. For this purpose we consider an extension of the conventional univariate Kaplan-Meier estimator for the hazard rate to multivariate right censored duration data and truncation of the marginal distributions of random regressors. It is a combination of nearest neighbor estimator and the Nelson-Aalen type estimator. It is a Akritas (1994) type estimator which adapts the nonparametric conditional hazard rate estimator of Beran (1981) to more typical data situations in applied analysis. We show with simulations that the estimator has nice finite sample properties and our implementation appears to be very fast. A small application to German unemployment duration data demonstrates the need for flexible specifications of conditional quantile functions. The results indicate that the level of social benefits has a strong impact on the length of long term unemployment for low earners and fertility is related with longer unemployment periods of females.

Wichert, Laura and (2005), Application of a Simple Nonparametric Conditional Quantile Function Estimator in Unemployment Duration Analysis, ZEW Discussion Paper No. 05-67, Mannheim.