BDM(2003) operation problems

 BDM operation problems

 

Q: Why BDM does not use log(GDP) directly, instead, they use the predicted value of GDP?

IncWB, this is “the residual components of log(real per capita income) that is independent of the selection institutions W and S.” They said this procedure allows them to “separate the effects of political institutions from wealth effects that cannot be attributed to those institutions” (BDM 2003:220)

 

Q: Why use region-year fixed effect?

BDM includes 6 regions, Europe, South, and Central America, North America and the Caribbean, Asia, the Middle East, and Africa. The main reason to use the region-year fixed effect is to remove “any temporal and spatial factors that might be the actual explanation for shifts in the values of our dependent variables. (BDM 2003: 217-218)

However, they do not explain why they use region-year instead of country-year. My guess is their independent variable (W, S, W/S) rarely changes within a country, so using country fixed effects are not ideal. 

BDM 2003 stata code

reg lgdpWB W S

predict incWB,residual

reg lgdpWB demaut

predict demincWB, residual

regr demaut W S

predict demres, residual

label var demres "Residual democracy not accounted for by W"