Page 1 of 1

some comments

Posted: Thu May 30, 2019 12:38 pm
by Luca Tiberti
Thanks Kamal and team for your interesting and challenging paper!

Just a few comments:
- an attrition rate of 18.4% is not low. Please discuss better why we should be confident that the final sample selection is not biased. Or, try to address it through standard ways
- Selection into labour market participation: how is this addressed in the paper (the Lewbel approach is not enough, but it just deals with the endogeneity of the treatment)
- please discuss the direction of the bias
- "household shock" does not seem a convincing IV external variable (reverse causality, directly related to your outcomes, etc).

I am also sharing a few papers you may find interesting:
- Fitzgerald et al. (to address attrition)
- Semykina (paper + codes here
- my paper on Brazil (to see how I interpreted the Lewbel IV)