Page 1 of 1

Discussant comment

Posted: Mon Jun 11, 2018 1:14 pm
by jdavalos
Comments on : Youth unemployment and transition from school to work in Benin

The paper explores the relationship between work experience and the probability of being employed. More specifically, it studies the effect of working while studying on the transition into the labor market.

Bjorn (2015) found that those who worked while studying had a lower probability to transition. This is not an academic document but an ILO report.

Regarding the introduction. It is simpler to convey the message in one, eventually two pages. You go beyond that and provide too many details. More experienced researchers may choose to merge the essentials of the literature background into the introduction. This is not what you do.

Regarding the conceptual framework and literature review.

• Your paper makes clear that there is a lack of evidence for developing countries while mentioning that there is already some evidence for Benin’s case (Björn, 2015).

• My recommendation is to summarize the length of these sections and to provide some details on Björn methodological approach and conceptual framework.

o We have no explanation about the transmission channel that is supposed to explain Björn findings for Benin.

Models and methods

Helpful to place the problem under the IE terminology
• Untreated: Those who finished their education without having worked while studying & found a job + those who quitted in order to seek for a job & found a job
• Treated : Those who where working while studying and found a job.
• Outcomes: Duration of the transition

Your system of equations deliver a different message.
• You should clarify since the beginning that the duration is only observed for those who have left-school and transitioned. This is only clarified after several paragraphs.

Instruments
• Unfortunately parents education is also a proxy of unobserved abilities caused by the household environment and also a proxy of parental networks that may ease the transition.
o Also, students facing economic vulnerability may be pushed to transition faster because they cannot afford the cost of a long job-search

• You write “WE believe that”… in order to justify your arguments. You must argue. The purpose of this section is avoid any “beliefs”
• You mention the local unemployment rate at the time of studying as an instrument. I do not “believe” it. A high unemployment rate may have pushed to WS (or the opposite), but at the same time this could have affected the unemployment spell ambiguosly. From the supply side, high unemployment rates could have increased the wiilingnes to work at lower wages in order to lower the unemployment spell. From the demand side, the unemployment spell should increase. The effect is then ambiguous. This might be explaining your falsification test result.
• Actually you say that you control for both instruments into the main equation. By definition, they need to be excluded (they are included only for the purpose of the falsification test, not for the final model).


Data

Many doubts were clarified in this section
• What about alternative definitions of treatment… What is the definition used by Björn (2015) ? Might this explain the discrepancy in results?

Results
• Now you explain the instruments were included to perform a falsification test, which means that they should not have been included in the final model as you stated in the methodological section.

• The falsification test only makes sense if you convince the reader about its null effect on the unemployment spell.

Others

There are missing references (Di Falco, etc.)