2
$\begingroup$

I want to estimate tobit marginal effects using multiply imputed data, however I see that tobit is not among the estimation commands supported by Stata's MI prefix - I understand that the validity of multiple imputation estimates hinges on the asymptotic normality of the parameter being estimated, but I haven't been able to find an answer (that I understand) which specifies whether this is the case here. Does anybody know what the asymptotic distribution of the tobit marginal effects is, or under what conditions it would/would not be normal?

This discussion on Statalist deals with the same question but doesn't resolve the issue. I have also replied there regarding the question.

https://www.statalist.org/forums/forum/general-stata-discussion/general/1291111-tobit-model-and-multiple-imputation?_=1527118657364

$\endgroup$
0
$\begingroup$

On page 674 of Wooldridge's Econometric Analysis of Cross Section and Panel Data, he states that the delta method can be used to find the standard error, and the delta method uses a Wald statistic which assumes normality.

Wooldridge p674

However, there must be a good reason why Stata doesn't include Tobit in the default list of estimators.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.