Use this URL to cite or link to this record in EThOS:
Title: Alternative econometric methods for the analysis of unemployment duration, with applications to the UK job seeker's allowance reform
Author: Bernardinelli, Daniele
Awarding Body: University of Essex
Current Institution: University of Essex
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
In this work, I present different methods for analyzing the effect of unemployment compensation on the duration of unemployment and subsequent employment stability, showing how going beyond standard models can provide useful insights to better understand the effects of policy interventions. Throughout the work, 1 use data from the British Household Panel Survey (BHPS), covering the period from September 1993 to August 1999, a period that includes t he Job Seekers' Allowance (JSA) reform in 1996. In Chapter 1, I investigate the effect of unemployment compensation on the duration of unemployment and subsequent employment duration using standard parametric binary choice models applied to discrete transition data, and I find that UC is associated with longer unemployment spells and longer subsequent employment spells. However, results show no significant effect of the reform either on unemployment duration or on post-unemployment job tenure. In Chapter 2, I present methods for analyzing duration data using censored quantile regression. The analysis confirms the basic findings from Chapter 1. However, what the quantile regression analysis allows to capture is the substantial heterogeneity across quantiles, with shorter durations being more sensitive to the policy variables considered. Finally, in Chapter 3 I present a method for applying quantile regression to censored duration data that are discrete in nature. The implementation of the model is illustrated using data from Chapter 2, where the durations are expressed in months. I find that the model produces estimates comparable to those of standard OR models, with generally smaller standard errors. My work shows that using quantile regression models for duration analysis can provide a richer characterization of the effect of policy intervention across different quantiles of the distribution of the outcome of interest.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available