ATE, ATT and checking for collinearity when using Endogenous Switching Regression

[Johan van Zyl]

Hello
I am using an Endogenous Switching Regression model (using the movestay command - Lokshin and Sajaia (2004)) to analyse the impact of credit on smallholder farmers' productivity. I please need assistance with the following:

1. After using the movestay command, the msat command must used to calculate the Average Treatment Effect and the Average Treatment Effect on the Treated (Lokshin and Sajaia). However, I am struggling to install the msat command. It is not available in the ssc directory of STATA. My understanding is that one needs mspredict_ar.ado and msat.ado files that needs to be saved in a certain directory, but I have no idea where to obtain these files from. Does anyone perhaps know where I can get these files or advise on some other way to calculate ATE and ATT?
2. I need to check for collinearity. Below is the code that I have so far. Please advise on this. Should I consider the insturmental variables at all in the collinearity test, and if so how do I do it?

************************************************
***** Define variables to be used in models *****
*************************************************

* Define dependent variable, y1list - Crop yield (mzyld, soyld, mlyld or rcyld)
* Define endogenous variable - y2list - credit
* Define exogenous variables - x1list - region ageh female union family wkage hplotsiz noplots hcultsiz hfertd irrig icrop (mzfms sofms mlfms rcfms). Change this last variable as needed, depending on which crop is considered
* Define instrumental variables - x2list - dep1 dep2
* Define alternative set of instruments x2alt for overidentified case
* Define exogenous variables x12 for eq2, instrumental variable x22 for eq2

global y1list lnmzyld
global y2list credit
global x1list ageh female union family wkage hplotsiz noplots hcultsiz hfertd irrig icrop mzfms
global x2list dep1 dep2
global x2listalt dep1 dep2 hcultsiz
global x1list2 ageh female union family wkage hplotsiz noplots hfertd irrig icrop mzfms
global x2list2 hcultsiz

*************************
***** Tests on data *****
*************************
* Test for collinearity

* Multicollinearity: parents' average education is collinear with
* whether they completed grad school or college

* Correlation table
correlate $y2list $x1list

* Run regression, find VIF. If VIF>10 then drop the variable
reg $y1list $y2list $x1list
vif

* Run regression without variable that has high VIF
reg $y1list $y2list $x1list2
vif

* Run regression without the other variables
reg $y1list $x2list2
vif
*VIF=1 for simple regression


Thanks in advance for assistance!
Regards
Johan