With thanks to Isaiah Andrews and Kit Baum, I am happy to announce that twostepweakiv is available on SSC. This package implements two-step weak-instrument-robust confidence sets based on Andrews (2018) and refined projection method for subvector inference based on Chaudhuri and Zivot (2011) for linear instrumental-variable (IV) models. twostepweakiv supports a range of variance-covariance estimators including heteroskedastic, autocorrelation, and one- and two-way cluster-robust VCEs. twostepweakiv can be installed via ssc install twostepweakiv. Examples and options can be found in its help file via help twostepweakiv. More supporting material, including a forthcoming Stata Journal article that describes two-step weak-instrument-robust confidence sets and refined projection method, can be found at https://github.com/lsun20/TwoStep.
I would also like to thank Keith Finlay, Leandro Magnusson and Mark Schaffer because twostepweakiv builds on and extends their weakiv. Unlike weakiv, twostepweakiv should be used as a standalone routine. twostepweakiv works by calling ivreg2 first to parse the specification provided by the user and then estimate a minimum-distance model depending on what the user has specified as estimator.
Requires: Stata version 11, avar, ivreg2, moremata and ranktest from SSC.
To check whether you have the required packages, you can use
If you encounter error messages such as
please install the necessary packages from SSC as twostepweakiv depends on them.
Further examples and options can be found in the help file. Error reports and suggestions are welcome as twostepweakiv is my very first SSC package. I hope to improve twostepweakiv with your input.
I would also like to thank Keith Finlay, Leandro Magnusson and Mark Schaffer because twostepweakiv builds on and extends their weakiv. Unlike weakiv, twostepweakiv should be used as a standalone routine. twostepweakiv works by calling ivreg2 first to parse the specification provided by the user and then estimate a minimum-distance model depending on what the user has specified as estimator.
Requires: Stata version 11, avar, ivreg2, moremata and ranktest from SSC.
To check whether you have the required packages, you can use
Code:
which avar which ivreg2 which ranktest mata: mata which mm_quantile()
Code:
command {avar ivreg2 ranktest} not found as either built-in ado-file
function mmsdf() not found
Further examples and options can be found in the help file. Error reports and suggestions are welcome as twostepweakiv is my very first SSC package. I hope to improve twostepweakiv with your input.
