comments from discussant

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

comments from discussant

Post by Luca Tiberti » Tue Jun 12, 2018 4:51 am

Thanks for your presentation! As said, this proposal has two main merits: the topic (urbanization, old topic but not fully explored) and the use of the satellite data. My detailed comments below:

- I would suggest to look at the very recent literature looking at the role of secondary towns (and migration to secondary towns) on poverty (Christiansen and Kanbur’s works, and special issue on World Development). Your paper may probably inscribe in this literature. It is important to look at this literature as it tells us that the welfare effects can differ by type of urban site.
- The definition of “urbanization” is still unclear to me from your proposal.
- My major concern is about the variability of your urbanization variable. Given your time span (relatively short - 2011/2016), do you think that the luminosity data vary enough to capture urbanization dynamics effects? Have you evidence of significant urbanization trend between 2011-2016? otherwise, you may think to use earlier dataset and have pooled datasets. More technically, which is the variability of your "treatment" variable? this is a key condition for identification.
- Luminosity as proxy of urbanization and development. Good idea. However, I suggest you look at the paper just published on Annual Review of Economics (attached), where the authors provide an interesting discussion on the use of this data. In particular, they discuss some possible caveats with this data, such as: comparability of these data across time (as satellites sensors have recently changed, data subject to saturation, the data are highly skewed (but nonlinear techniques like MLE or log transformations can be done), they are subject to blooming.
- It would be useful to see the correlation between the change over time between population density and luminosity. Population density - another possible proxy for urbanization - after a given threshold might negatively affect welfare and labour market conditions.
- The proposal identifies a series of positive impact channels. What about the negative channels? Think to leakages in terms of sanitation, electricity, and so on due to urbanization. Also, it is stated “it is worth noting that urbanization involves some other structural and nutritional transitions …”: please be more explicit here
- related to the previous point, you should develop a strong conceptual framework to derive your testable hypotheses. I see competing effects - theoretically, the overall effect is undetermined. But the conceptual/theoretical framework should help to understand such mechanisms.
- At some point, you talk about “urbanization programs”: what do you refer exactly?
- How do you relate rural-to-urban migration with urbanization? It seems to me that such a link is only explored in a secondary analysis. In such a case, you propose to identify internal movements through an exogenous shocks such as positive and negative rainfall deviations. It captures only a part of migration, and specifically male migration (as women are generally much less mobile). Also, to identify the movements, you need lagged variables, which reduces your sample. Hence, check whether it is a convenient approach.
(4.6 MiB) Downloaded 69 times


Return to “Can urbanization improve household welfare and provide inclusive opportunities? The case of urban expansion in Ethiopia – PMMA 20210”

Who is online

Users browsing this forum: No registered users and 1 guest