Subsidized temporary jobs: lock-in and stepping stone effects

Abstract

This article evaluates the effectiveness of subsidized temporary jobs as stepping stones to regular employment. We study a French program that allows job seekers to work part-time while remaining registered with the unemployment agency. In this program, insured individuals concurrently receive part of their unemployment benefits and wage income. Using administrative data, we find that subsidized temporary jobs have both a significant lock-in effect and a significant positive post-treatment impact on the hazard rate to employment. Since individuals facing a high implicit tax rate have incentives to self-select into better part-time jobs, we also find that a higher tax rate leads to a weaker lock-in effect and a stronger post-treatment effect. Simulations suggest that the lock-in effect first dominates, but that the overall effect eventually becomes positive. They also point to ways of improving the effectiveness of the policy.

Keywords:

• unemployment duration
• duration models
• policy evaluation
• treatment effects
• part-time work
• lock-in effects

JEL Classification:

• J64
• J68
• C41

Acknowledgements

We are grateful to Bruno Crépon, Christian Göbel, Francis Kramarz, Rafael Lalive, David Margolis, Jean-Marc Robin, and Josef Zweimüller for helpful suggestions. We also would like to thank participants at the ESEM 2008, EALE 2008, and JMA 2008 conferences, the Third Workshop on Evaluation of Passive and Active Labour Market Policies, the CREST, CEE, University of Évry, University of Orleans and University Paris 1 – Panthéon-Sorbonne seminars, and the Dutch Ministry of Social Affairs for their comments. All remaining errors are ours.

Notes

1 See Munts (1970) and Holen and Horowitz (1974) for early studies of the American partial benefit program.

2 Hereafter we alternate the expressions ‘reduced activities’ (RAs), ‘temporary jobs’ and ‘part-time jobs’ when referring to occasional employment under the French Activité Réduite program.

3 Here, the term ‘labour market spell’ refers to one unemployment spell (including time to treatment, in treatment and to nonrandom censoring), and possible subsequent employment spells.

4 The coefficient attached to the replacement ratio in the equation for the duration until treatment is −1.148 and is significant at the 1% level.