Long time reader, first-time poster!

I am analysing a data set where my dependent variable (expenses claimed) is censored at 0 and thus wish to use a Tobit model to analyse my results. I am interested in testing both a univariate version (i.e. Tobit type-1 or one-step Tobit) and bivariate version (i.e. Tobit type-2 or double hurdle model). These are the current commands I am using:

Code:
* Tobit Type-1
tobit depvar varlist1 varlist2, ll(0)
* Tobit Type-2
probit depvar varlist1 varlist2
truncreg depvar varlist1 varlist2, ll(0)
However, this becomes complicated when I try to include an instrumented variable for varlist2, which I believe is endogenous. Whilst STATA appears to have a command for Tobity Type-1, I cannot find the equivalent for Tobit Type-2 in STATA or extensions (e.g. craggit)

Code:
* IV Tobit Type-1
ivtobit depvar varlist1 (varlist2 = varlistiv), ll(0)
Would it be valid to estimate my two-steps separately using the same IV?

Code:
* IV Tobit Type-2
ivprobit depvar varlist1 (varlist2 = varlistiv)
ivregress depvar varlist1 (varlist2 = varlistiv) if depvar>0
Many thanks!

Luca