Artur T. schrieb: > Hello, > > I am trying to replicate Verbeek's schooling example (Ch. 5.3). He does > an IV estimation for a wage equation (p. 150, table 5.3) with 7 > regressors (incl. the constant) but instruments only 3 variables and > hence only uses three instruments. > In stata one can associate instruments to specific variables like: > > ivreg lwage76 (ed76 exp76 exp762 = age76 age762 nearc4) black smsa76 south76 > > where only "ed76 exp76 exp762" are "instrumented". > Is this possible in gretl? I did not find any hints... > > IIRC, you just have to include the *non*-instrumented variables in the instrument list. That way they "trivially" serve as their own instruments, and the "true" instruments then can only refer to the variables treated as endogenous. HTH, sven _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users