Dear all,

I am doing my undergraduate dissertation on the relationship between natural resource rents as % of GDP (Tnrgdp) and corruption. Since I investigate 160 countries over 21 year period it looks appropriate to use xtabond2 (command large N small T). I will include control variables: ln(gdp per capita), ln(population), government effectiveness (geffect), rule of law (rol), civil liberties (civlib). The code I intend to use is presented below. Do you think I will explain away my effect with by including such related controls? Also, how can I deal with the clustered nature of time-series cross-sectional data?

My code looks like this:
Code:
xtabond2 corruption l.corruption Tnrgdp lngdp lnpopulation civlib geffect rol y*, gmm(l.corruption, collapse) iv(Tnrgdp lngdp lnpopulation geffect rol civlib) iv(y*, eq(level)) twostep robust orthogonal small
I would be happy if I get any advice or suggestions from more experienced people on my code or model, chosen controls etc.

Sincerely,

Mateo Radman