Discussion Paper

No. 2018-13 | February 02, 2018
Labour contracts and stepping-stone effect in Italy: a multinomial analysis


Do short-term contracts facilitate the transition to permanent contracts? The authors use a rich administrative database for Italy to run a stepping stone analysis and evaluate which contractual agreements have more chances to lead to a permanent working position. They find that individual specific characteristics make it more likely for a worker to be employed with a specific contractual agreement and that the contribution toward more working stability varies with the previous contract. The authors conclude that fixed term positions act more as stumbling blocks than building blocks for open-ended contracts.

JEL Classification:

J20, J21, J41


  • Downloads: 838


Cite As

Maria Giovanna Bosco and Elisa Valeriani (2018). Labour contracts and stepping-stone effect in Italy: a multinomial analysis. Economics Discussion Papers, No 2018-13, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2018-13

Comments and Questions

Anonymous - Evaluation
February 03, 2018 - 12:40

I have two major concerns.

1/ This analysis is not state-of-the-art in a sense that it does not correct for unobserved heterogeneity, which is a huge issue in this context -- those in short-term contracts may differ from the overall population in factors such as ability and motivation. Why ...[more]

... don't the author follow the literature and run models that do correct for this issue? See the following studies. At least the authors should acknowledge this problem.

Baert, S., Cockx, B., Verhaest, D. (2013): Overeducation at the Start of the Career: Stepping Stone or Trap? Labour Economics, 25, 123-140.

van den Berg, G. J., Holm, A., van Ours, J. C. (2002): Do stepping-stone jobs exist? Early career paths in the medical profession, Journal of Population Economics, 15, 647-665.

2/ The study's literature review is very ad hoc and incomplete. It is very strange that the aforementioned study were not cite. I have the impression that the authors didn't put much effort in reading the literature.

Maria Giovanna Bosco - alternative methodologies
April 09, 2018 - 10:00

The literature we refer to is wide and generally suffers from the bane of identification of the starting population. We are perfectly aware of that, and also that finding the ultimate methodology to overcome the issue is almost impossible. We performed a test about the potential bias that would arise ...[more]

... in case of endogeneity in our published article here: http://rajournals.in/images/ijararticle/v3-i10/2rajar.pdf - to find that thanks to the huge size of starting population, the issue is likely to be minimized. Therefore, we are likely on the correct way even in a multinomial set-up, as these results are consistent with our previous findings.

I am grateful for the literature suggestion.

Anonymous - Not really a stepping stone analysis
February 06, 2018 - 09:22

The topic of this study sounds interesting, but the analysis is not convincing. The choice of the control group is strange and probably not appropriate for analyzing stepping stone effects. It is also unclear whether the outcome variable is suitable for the analysis.

Maria Giovanna Bosco - Incomplete comment
April 09, 2018 - 10:07

The outcome variable is a status in terms of probability with respect to other statuses, as widely known in the literature, for instance see Cai, L., Law, V., and Bathgate, M., (2014 ). As for the "strange" selection of population, we took all the dependent workers in a given Italian ...[more]

... region. That is - the entire universe - in a given area. If this is strange, well, perhaps any other choice would be a narrower one. I agree with the fact that most of the times stepping stone analyses consider flowing from unemployment to employment, but we have observational data and not a smaller sample of workers, as most other studies do. Since we are not able to see what happens when a person is not employed, we would introduce a serious bias by supposing unemployment where independent work or informal work could actually be the the real statuses.

Anonymous - Referee Report 1
April 09, 2018 - 16:11
See attached file

Maria Giovanna Bosco - Reply to Referee Report 1
November 12, 2018 - 09:58

We might change the paper title and abstract as to meet the regional dimension of the article, although the methodology applies to large sample of workers that could fairly represent a sample of the Italian dependent workers. We will re-write the Introduction, Conclusion and Implications as to the reflect also ...[more]

... the status of the art of the current debate over the transitions to permanent jobs.

Anonymous - Referee Report 2
April 27, 2018 - 21:46

The paper studies whether different forms of short-term contracts increase the probability that workers transit to a permanent contract.

My main comments regarding the paper are the following:
1) The article is not much reader friendly, I think the exposition of the material could be improved. At some points, ...[more]

... connections are made to irrelevant points - some described below in the minor comments.

2) It is not clear the contribution to the literature. The authors correctly report that the literature have reached some mixed results. However, it is not clear how the article is advancing the understanding on how short-term contracts help or damage workers. Is the authors' finding based on better empirical method? Is the data used superior? or else? Why? I believe the authors should make the paper's contribution clear from the beginning.

3) The empirical implementation is problematic. Restricting the analysis to workers who have had two transitions may create a serious problem for the analysis because the worker's number of transitions is a consequence of this contractual type. The authors should discuss this aspect.

4) Another key issue is that workers self-select into different contracts, creating a clear issue with endogeneity. The authors should discuss such issue at length and explain why controlling for the covariates they observe should be enough to deal with that.

Other points:
pg2. The paper initially makes a connection with the Jobs Act, that mostly concern firing regulations. It is not clear the connection with temporary contracts.

pg5. Why not working on a sample instead of dropping out workers with man spells? This might be problematic because the number of spells is a quite endogenous variables, as it depends on the type of job contract (according to hypothesis raised by the authors and in the literature.

typo pg2 "in the 2008-2012 rime span to"
typo pg3 "On the other, the “trap” or “deadlock”"
typo/informal pg5 "We don’t consider here"

confusing passage "We don’t consider here those workers who have been working since 2007 or earlier and are still on the job in the five years 2008- 2012; and those who start a job in 2009-2012, but not in 2008, (so perhaps they start in 2009 and keep working - they do not appear in the present study)."

Maria Giovanna Bosco - Reply to Referee Report 2
November 12, 2018 - 10:01

1) + 2) A rewriting of the article may be helpful in making the text more readable. As for point 2) specifically, we rely on definetely superior and rich individual-level observational data. We will stress this point further in case it was not so clear from the beginning.

3) ...[more]

... "Restricting the analysis to workers who have had two transitions may create a serious problem ". We performed the analysis also for workers with more than two transitions, even if we did nor report the results, as the they do not provide further insights but confirm what we found using only two transitions.

4) Self-selection is a problem we are aware of. In a different article we proved that even controling for endogeneity with a different methodology, result are robust, as open-ended jobs are the best predictors for subsequence.