Easy way to do a Generated Regressor correction (standard errors)?

[WilliamVigil]

I am doing a Control Function approach to deal with a case of endogeneity. While doing this I noted that the OLS standard errors in the revised regression are likely wrong.

I was wondering if there's a straightforward way to do a correction for the sampling error in this case, basically one that gives the correct standard errors of generated regressors?

What I'm looking for is something like the correction that Stata does automatically in the context of 2SLS (specifically in the 2nd stage of the 2SLS procedure).

Thanks

[Jeff Wooldridge]

It's easiest to use a bootstrapping method, where you include both parts of the estimation for every bootstrap sample. This is fairly standard for CF applications.

[Stephen Jenkins]

In addition to approaches based on bootstrapping, see also:
(1) "The robust variance estimator for two–stage models", by James W. Hardin, The Stata Journal, 2002, Volume 2 Number 3: pp. 253-266 http://www.stata-journal.com/article...article=st0018. (Free download): Abstract. This article discusses estimates of variance for two-stage models. We present the sandwich estimate of variance as an alternative to the Murphy–Topel estimate. The sandwich estimator has a simple formula that is similar to the formula for the Murphy–Topel estimator, and the two estimators are asymptotically equal when the assumed model distributions are true. The advantages of the sandwich estimate of variance are that it may be calculated for the complete parameter vector, and that it requires estimating equations instead of fully specified log likelihoods.
(2) "Calculating Murphy–Topel variance estimates in Stata: A simplified procedure" by Arne Rise Hole, The Stata Journal, 2002, Volume 6 Number 4: pp. 521-529, http://www.stata-journal.com/article...article=st0114 (free download) Abstract. Building on the work by Hardin (Stata Journal 2: 253–266), this note shows how the calculation of the Murphy–Topel variance estimator for two-step models can be simplified in Stata by using the scores option of predict.
(3) The Stata Journal Volume 10 Number 2: pp. 252-258, http://www.stata-journal.com/article...article=st0191
And so on.
Lesson: Stata's search or findit commands (and Google) are your friends. Make contact with them ...

[Sayoree Gooptu]

The Murphy topel correction of variances commands are given when the endogenous variable is discrete as it involves score in the predict command. Is it possible to do the same for a continuous variable? I am unable to find the commands anywhere.