Dear Andrea,
To generate a variable with one of the e(b) values:
gen beta_var1 = _b[var1]
This will generate a constant.
However I think it is not necessary for you to convert the vector e(b) into
variables.
After estimation of the first equation you can save e(b)
-matrix mymatrix = e(b)-
Then you can use all or SOME of the values of -mymatrix- to estimate
predicted values using a second set of Xs. Imagine that in the second step
you only want to use the first two estimated betas from the first equation.
