Dear all,
I am asking a novice question regarding the appropriate level of IV when you have individual level data. Specifically, I am thinking of using a group-level instrument (i.e. trade volume) to instrument for endogenous group-level treatment (degree of recession) on outcomes (i.e. public spending). Does this make sense at all? I am asking this because my first stage looks ridiculously large, probably due to the clustering. Could anyone suggest any readings? Thanks much in advance!
I am asking a novice question regarding the appropriate level of IV when you have individual level data. Specifically, I am thinking of using a group-level instrument (i.e. trade volume) to instrument for endogenous group-level treatment (degree of recession) on outcomes (i.e. public spending). Does this make sense at all? I am asking this because my first stage looks ridiculously large, probably due to the clustering. Could anyone suggest any readings? Thanks much in advance!