Estimating the latent effect of unemployment benefits on unemployment duration

Simon M. S. LO, Gesine STEPHAN, Wilke RALF

Research output: Working paperDiscussion paper seriesResearch

Abstract

We estimate the effect of a shortening of unemployment benefit entitlements on unemployment duration. Previous studies on the same or related problems have not taken into account that the competing risks duration model is not identified and we shed first light on the question whether the non identification problem may preclude informative results. It turns out that the identification bounds for the parameters of interest are very wide in the absence of strong assumptions. We suggest an assumption on the dependence structure between risks which is milder than what conventional duration models assume. Under this assumption, the identification bounds are tighter and become informative for the direction of the treatment effect. We find evidence that the unemployed with higher pre-unemployment earnings are more likely to enter full-time employment and, in particular, subsidized self-employment.
Original languageEnglish
Place of PublicationGermany
PublisherThe Institute of the Study of Labor (IZA)
Number of pages32
Publication statusPublished - Jun 2012

Publication series

NameIZA Discussion Paper Series
No.6650

Fingerprint

Unemployment duration
Duration models
Unemployment benefits
Dependence structure
Treatment effects
Identification problem
Self-employment
Unemployment
Competing risks
Entitlement

Bibliographical note

Revised version published as 'Competing Risks Copula Models for Unemployment Duration: An Application to a German Hartz Reform' in Journal of Econometric Methods, 2017, 6, 1-20.

Keywords

  • unemployment benefits
  • unemployment duration
  • difference-in-differences
  • partial identification
  • dependent censoring

Cite this

LO, S. M. S., STEPHAN, G., & RALF, W. (2012). Estimating the latent effect of unemployment benefits on unemployment duration. (IZA Discussion Paper Series; No. 6650). Germany: The Institute of the Study of Labor (IZA).
LO, Simon M. S. ; STEPHAN, Gesine ; RALF, Wilke. / Estimating the latent effect of unemployment benefits on unemployment duration. Germany : The Institute of the Study of Labor (IZA), 2012. (IZA Discussion Paper Series; 6650).
@techreport{7ce56520c23b40e2945b2f73e5e158d1,
title = "Estimating the latent effect of unemployment benefits on unemployment duration",
abstract = "We estimate the effect of a shortening of unemployment benefit entitlements on unemployment duration. Previous studies on the same or related problems have not taken into account that the competing risks duration model is not identified and we shed first light on the question whether the non identification problem may preclude informative results. It turns out that the identification bounds for the parameters of interest are very wide in the absence of strong assumptions. We suggest an assumption on the dependence structure between risks which is milder than what conventional duration models assume. Under this assumption, the identification bounds are tighter and become informative for the direction of the treatment effect. We find evidence that the unemployed with higher pre-unemployment earnings are more likely to enter full-time employment and, in particular, subsidized self-employment.",
keywords = "unemployment benefits, unemployment duration, difference-in-differences, partial identification, dependent censoring",
author = "LO, {Simon M. S.} and Gesine STEPHAN and Wilke RALF",
note = "Revised version published as 'Competing Risks Copula Models for Unemployment Duration: An Application to a German Hartz Reform' in Journal of Econometric Methods, 2017, 6, 1-20.",
year = "2012",
month = "6",
language = "English",
series = "IZA Discussion Paper Series",
publisher = "The Institute of the Study of Labor (IZA)",
number = "6650",
type = "WorkingPaper",
institution = "The Institute of the Study of Labor (IZA)",

}

LO, SMS, STEPHAN, G & RALF, W 2012 'Estimating the latent effect of unemployment benefits on unemployment duration' IZA Discussion Paper Series, no. 6650, The Institute of the Study of Labor (IZA), Germany.

Estimating the latent effect of unemployment benefits on unemployment duration. / LO, Simon M. S.; STEPHAN, Gesine; RALF, Wilke.

Germany : The Institute of the Study of Labor (IZA), 2012. (IZA Discussion Paper Series; No. 6650).

Research output: Working paperDiscussion paper seriesResearch

TY - UNPB

T1 - Estimating the latent effect of unemployment benefits on unemployment duration

AU - LO, Simon M. S.

AU - STEPHAN, Gesine

AU - RALF, Wilke

N1 - Revised version published as 'Competing Risks Copula Models for Unemployment Duration: An Application to a German Hartz Reform' in Journal of Econometric Methods, 2017, 6, 1-20.

PY - 2012/6

Y1 - 2012/6

N2 - We estimate the effect of a shortening of unemployment benefit entitlements on unemployment duration. Previous studies on the same or related problems have not taken into account that the competing risks duration model is not identified and we shed first light on the question whether the non identification problem may preclude informative results. It turns out that the identification bounds for the parameters of interest are very wide in the absence of strong assumptions. We suggest an assumption on the dependence structure between risks which is milder than what conventional duration models assume. Under this assumption, the identification bounds are tighter and become informative for the direction of the treatment effect. We find evidence that the unemployed with higher pre-unemployment earnings are more likely to enter full-time employment and, in particular, subsidized self-employment.

AB - We estimate the effect of a shortening of unemployment benefit entitlements on unemployment duration. Previous studies on the same or related problems have not taken into account that the competing risks duration model is not identified and we shed first light on the question whether the non identification problem may preclude informative results. It turns out that the identification bounds for the parameters of interest are very wide in the absence of strong assumptions. We suggest an assumption on the dependence structure between risks which is milder than what conventional duration models assume. Under this assumption, the identification bounds are tighter and become informative for the direction of the treatment effect. We find evidence that the unemployed with higher pre-unemployment earnings are more likely to enter full-time employment and, in particular, subsidized self-employment.

KW - unemployment benefits

KW - unemployment duration

KW - difference-in-differences

KW - partial identification

KW - dependent censoring

M3 - Discussion paper series

T3 - IZA Discussion Paper Series

BT - Estimating the latent effect of unemployment benefits on unemployment duration

PB - The Institute of the Study of Labor (IZA)

CY - Germany

ER -

LO SMS, STEPHAN G, RALF W. Estimating the latent effect of unemployment benefits on unemployment duration. Germany: The Institute of the Study of Labor (IZA). 2012 Jun. (IZA Discussion Paper Series; 6650).