I'm using instrumental variable to estimate a regression with IVLASSO. I have a problem with weak instrument. In some specifications the first stage F statistic is smaller than 10, I heard that in this case, I should adjust the statistics for weak instrument and look at a different table for standard error and F statistics, because for weak IV a different test should be used. Can anyone tell me how can I get the right statistics for weak IV? What is the command in stata?
No I'm using the following command:
Code:
ivlasso cm15 drt19-drt22 (ir_part_any = dr_1012 dr2_0810 dr_0810 dr2_0608 dr_0608 any_shock0810 all_shock0810 any_shock0608 all_shock0608 dr2_1012 dr5-dr12 drt5-drt12 dr_0609 dr_0610 dr2_0609 dr2_0610 dr_0612 dr2_0612 any_shock0612 all_shock0612 most_shock0612) if female==1 & cost>0 & cost!=. & current_age!=. & current_age>=15 & ma8_inf>12 & ma8_inf!=., cluster(id_geo_ex) first idstats
Regards,
Jiajing
0 Response to How to adjust right test statistics for weak instrument?
Post a Comment