What can we learn from labour force surveys about the impact of temporary employment?

Over the last ten years the Italian labour market experienced several changes in the regulation of temporary contracts. Using short panels from the Italian Labour Force Survey (Lfs) we identify the causal effect of experiencing a spell of temporary employment vs. a spell of unemployment on short-term labour market outcomes. The parameter of interest is recovered by imposing that, conditional on a suitable set of observable characteristics of the workers, of their households as well as of the local labour market, the treatment status is ignorable for the outcome. We carry out the analyses over three two-year periods characterized by increasingly open arrangements about the use of temporary contracts.

We exploit the features of the Lfs rotating sampling scheme to build a test for the ignorability assumption. It turns out that for the 1995/96 and 2000/01 samples, based on the old quarterly Lfs, ignorability is rejected, while for the 2005/06 sample, based on the new continuous Lfs, the matched treated and control groups turn out statistically balanced.

As for the estimate of the impact, experiencing a spell of temporary work takes to a 30% higher employment rate one year later for men, 35% for women. Most of this impact is due to temporary and unsatisfactory jobs, though. When we look at the impact on the probability of experiencing a transition to a permanent job, the effect is not significant for men, and just marginally significant for women. As for the impact on the probability to get a satisfactory job, it is significant and as large as 9.5% both for men and women.

Finally, there is a sizeable heterogeneity of the effects across areas for men: for permanent employment there is a positive effect in the North, while it is negative in the South; for satisfactory employment the effect is about 15% in the North, while it is nil in the South.