Dear All,
I am attempting to estimate the relationship between the change in poverty and economic growth. In an attempt to control endogeniety, I am using a two-step GMM procedure through xtabond2.
However, I am having issues with how to correctly specify the code. I am trying to replicate a paper in the literature - the paper states that "models are estimated using two-step GMM with instruments of lagged log income and difference of logged population. Hence, I ran the following specification:
grlhcountglob - logged difference in poverty headcount
grlminc - logged difference in mean income
lpop - logged population
I am unsure if this is correctly instrumenting for the variables as stated in the literature. Additionally, I am not sure with we have to include the lagged dependent variable in the instrument set/or as a covariate as well to correctly extract an estimate.
I appreciate any help with this matter,
Thank you very much.
I am attempting to estimate the relationship between the change in poverty and economic growth. In an attempt to control endogeniety, I am using a two-step GMM procedure through xtabond2.
However, I am having issues with how to correctly specify the code. I am trying to replicate a paper in the literature - the paper states that "models are estimated using two-step GMM with instruments of lagged log income and difference of logged population. Hence, I ran the following specification:
Code:
xtabond2 grlhcountglob grlminc, twostep gmm(L.lminc D.lpop, collapse) small robust
grlminc - logged difference in mean income
lpop - logged population
I am unsure if this is correctly instrumenting for the variables as stated in the literature. Additionally, I am not sure with we have to include the lagged dependent variable in the instrument set/or as a covariate as well to correctly extract an estimate.
I appreciate any help with this matter,
Thank you very much.
