-1

I am trying to use the twostepweakiv command in Stata to get weak-iv robust confidence intervals for an IV model with two endogenous regressors and two instrumental variables.

I am struggling to interpret the output of the twostepweakiv command.

For one endogenous regressor, the output does not report the upper and lower bound of the confidence interval. Only the lower bound is reported, while the upper bound is reported as "...".

   Test    Conf. level     Confidence Set        
-----------------------------------------------------------------------
LC_2sls        95%         [ .000322,   ...  ]

This sentence is also reported:

LC test gamma_min is 5%; distortion cutoff is 11% based on the given grid, obtained by 10^6 simulation draws.

For the second endogenous regressor, instead of the upper and lower bound of the confidence interval, the expression "entire grid" appears.

   Test    Conf. level     Confidence Set        
-----------------------------------------------------------------------
LC_2sls        95%         entire grid

Any help with the interpretation of this output and suggestion on how to get the command to produce a full confidence set would be extremely welcome!

Nick Cox
  • 35,529
  • 6
  • 31
  • 47
Silvia
  • 9
  • 1
  • This is more statistical than about programming, but with no specification of data or even your exact code, even the program author (not me) might struggle with this. Why not ask on CV or Statalist? – Nick Cox Aug 07 '23 at 21:10
  • thanks a lot Nick, I was just about to post on statalist indeed. – Silvia Aug 08 '23 at 09:36

0 Answers0