some comments and suggestions

Luca Tiberti
Posts: 33
Joined: Thu Jun 07, 2018 6:31 am

some comments and suggestions

Post by Luca Tiberti » Thu May 30, 2019 1:31 pm

Thank you, Jennifer and team, for your really interesting and strong paper. Let me share some comments:

- Introduction: I would add:
o the possible endogeneity biases and how you addressed them
o possible pathways you tested
o main results (punctual estimations)
Also, I would probably move the literature review therein and make your contribution even clearer (e.g., I am not sure whether the novelty in your paper – in addition to the index – is the country or also that you considered “different stages in life”).
- Section 2:
o I would title “Context and data”
o 2.1 “The 1991-2002 Civil war in Sierra Leone” (Figure 1 should be shown in 2.3, as we still do not know about the definition of the conflict variable)
o 2.2 “Describing the Data”
o 2.3 “Defining conflict” (where you also show some desc stats on conflict data)
o 2.4 “Defining Labour Market Outcomes” (where you also show some desc stats on LMO). BTW, are the employment and LMP for any economic sectors? I would probably distinguish between sectors (wage, self-employment in agricultural and non-agricultural sectors) as the effect (and the pathways you identified) may differ.
o 2.5 “Defining the timing of key variables” (this is something is sparse in the paper, but would be better to have a clear description of the timing of your key variables (outcomes and conflict) to understand well how you set the causal relationship you estimate. Here, I would add I would add some non-parametric investigation of the relationship between the intensity of violence and your continuous outcomes
- Section 3:
o “Identifying the effect of conflict on LMO: empirical strategy”
o 3.1 “Econometric specification”
o 3.2 “Challenges to Identification”
- Sections 4 and 5 (current 6 and 7 are fine to me)
- Section 6: “Conclusions” (rather than 7c)
In the discussion of potential pathways (currect section 3), you identified two possible conflicting mechanisms (“discouragement” and “added worker” effects). You should try to link this discussion to your results more explicitly.
Econometric specification:
- Explain µ in the text
- This specification seems valid only for those who did not migrate (as “j” just indicates the place where he/she was born). Do you know exactly (year) when the person eventually migrated and where?
- I am not sure to understand why the reference cohort should be between 9 and 34
I generally liked the identification strategy. However:
- You state that the DDR helped some former soldiers to join the labour market. Might it have direct implications on the individuals you observed (if young/child soldiers)? Is the programme homogenous across the country?
- it seems excluding any possible “spill-over” effects. Is that a realistic assumption? (a village in conflict does not affect other villages in the country and, especially, its neighbouring villages).
- Related to this, I would control for a few more pre-conflict socio-economic (e.g., LMP at j level?) village variables (in its current specification, the model does not include any socio-economic variables which might affect more directly labour market outcomes).
- Also, removing migrants and compare the results (with and w/out migrants) is partially satisfactory. You probably do not have the exact time of migration and the new destination, so you cannot do better. But it should be said, as well as how many people migrated.
Table 2: col 5 and 6, control variables should be “yes”?
Table 6: why number of chiefdom just 39?
Results on wages: -2/0 and 0-3 seem having the largest impact here – contrarily from the employment/LMP estimates. Can this be explained that here, I guess, you just look at wage workers and not to all the labour force/employees? If so, I would reiterate my suggestion to also run sectorial estimations for the employment/LMP
- IV: please say more about the “measure of quality of education”. Also, a few lines on the Wooldridge (2002) would be probably necessary.
- You state that education is the main impact pathway. However, it seems that this is the real impact mechanism you tested.
Table 8: please add a note for each specification.
Impact pathways. E.g., I would also look at, if available:
o Psycological-stress related variables
o Family assets
o Supply/demand side variables affecting employment and wages?


Return to “PMMA-20257 - Civil conflict and labor market outcomes in Sierra Leone”

Who is online

Users browsing this forum: No registered users and 1 guest