Good morning to everyone. I am working with IPUMS data to study some aspects about immigration. And now, I am facing an issue about ancestries.
To simplify, imagine that the US has 400 counties and 4 regions.
I basically count the number of ancestries in counties 1 -100 located in region 1 and then regress this dependent variable on the total number of immigrants located in regions 2,3,4 to overcome an endogeneity problem. Again, I do the same for ancestries located in counties 101…200 in region 2 and they regress it on the total number of regions 1,3,4 and so on for the other counties/regions.
Of course each regression has its own coefficients.
Now my question is whether is possible to aggregate these regressions into one regression and get one coefficient.

This is the code to give an example:

Code:
*Ancestry construction in counties of region 1*
egen ancestry1_1980 = count(id) if year==1980 & region==1, by(birthplace county)
*Immigrants of other regions*
egen outsiderimmigrants234_1980=count(id) if  region !=1 & year==1980, by(birthplace)
*regression
reg ancestry1_1980 outsiderimmigrants234_1980

*Ancestry construction in counties of region 2*
egen ancestry2_1980 = count(id) if year==1980  & region==2, by(birthplace county)
*Immigrants of other regions*
egen outsiderimmigrants134_1980=count(id) if  region !=2 & year==1980, by(birthplace)
*regression
reg ancestry2_1980 outsiderimmigrants134_1980
Thank you.