[Asad Rind]

Hello everyone,

I have one question similar to this. How can we compare whether the coefficients of two different groups are statistically different or not when using "IV regression". I have searched alot but unfortunately can not find the results for such a test. I am using an instrumental variable so its not possible for me to test through OLS.

i also used this code

ivregress y (a = x y z) b c
est store one
ivregress y (a = x y z) b c d
est store two
suest one two

but got this error

"unable to generate scores for model one suest requires that predict allow the score option"

[Bimal Adhikari]

I have a similar question regarding split-sample/interaction model analysis. I am trying to understand whether major power countries allocate foreign aid differently than minor power donors based on recipient's UN voting. I run the following code which gives me the following graph. The question (if I have been clear enough) is that, does this mean that minor powers do not base their aid-giving on recipient's voting behavior, whereas major powers reduce government-to-government aid when recipients vote differently. Also, does the overlap in confidence interval between major donors and minor powers affect the results? PS. The graph is based on the interaction term between "major power donors and UN voting."

eststo: xtreg F.lnpublic absidealdiff major am i.major#c.(loggdppc  growth govtecon physint  governance_all_mean  polity2 prio2 lndistance colony  logtrade  sc2  lnpopulationtotal2 defense lnall ), re robust

         
     
     
cap postclose aei_s

postfile aei_s major me_a se_a lo_a hi_a using "aei_s.dta", replace
foreach n of numlist 0(1)1 {

lincom absidealdiff + am*`n'
local me_a = r(estimate)
local se_a = r(se)
local lo_a = `me_a' - 1.96*`se_a'
local hi_a = `me_a' + 1.96*`se_a'

post aei_s (`n') (`me_a') (`se_a') (`lo_a') (`hi_a') 

}
postclose aei_s

    preserve
    use aei_s,clear
    twoway (rcap hi_a lo_a major)
    restore