why the results from the seperated 2-stage regressions are different from IV-regression?

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

why the results from the seperated 2-stage regressions are different from IV-regression?

Qian Yu
This post has NOT been accepted by the mailing list yet.
 Dear all
       my equation of regression is:
         value_real=a0*china_exp_real+a1*gdp_imp+a2*gdp_capita_imp+a3*gdp_exp+a4*gdp_capita_exp+a5*dist+a6*product_area +a7*land+a8*border+a9*comlang_off+a10*col45+a11*colony
since a0*china_exp_real is an endogenous variable, then it's instrumented by two IV variables:gdp_china and varydistance_china
       I obtained different results from the separated 2-stage regression and IV-2SLS, ie:
       First stage:

xtreg china_exp_real gdp_china varydistance_china gdp_imp gdp_capita_imp gdp_exp gdp_capita_exp dist product_area land border comlang_off col45 colony

predict china_exp_realhat

       Second stage:

xtreg value_real china_exp_realhat gdp_imp gdp_capita_imp gdp_exp gdp_capita_exp dist product_area land border comlang_off col45 colony


       IV-2SLS
xtivreg value_real gdp_imp gdp_capita_imp gdp_exp gdp_capita_exp dist product_area land border comlang_off col45 colony ( china_exp_real= gdp_china varydistance_china)


Why the final results of regression are different????


Thanks in advance,
Qian