I am estimating a model for credit access as a function of some exogenous and other interest variables using a household survey data. Based on survey questionnaire, I identified the following recursive binary dependent variables.
Demand for credit (D) = 1 or 0
Access to Credit (A) =1 or 0, if D=1
Credit sources (S) =1 or 0, if A=1
I estimated a three-stage probit regression as follows (suppose vector of variables X1, X2 and X3 for each stage)
Stage 1. probit D X1, vce(.)
Stage 2. Probit A X2 L if D==1, vce(.)
Stage 3. Peobit S X3 M if A==1, vce(.)
Note: L and M are estimated inverse Mill’s ratio from stage 1 and stage 2, respectively to correct some selectivity biases. However,
I am struggling to use David Roodman’s cmp stata command. Is there any one to help me out with the appropriate cmp syntax please?
Best Regards,
seid
Demand for credit (D) = 1 or 0
Access to Credit (A) =1 or 0, if D=1
Credit sources (S) =1 or 0, if A=1
I estimated a three-stage probit regression as follows (suppose vector of variables X1, X2 and X3 for each stage)
Stage 1. probit D X1, vce(.)
Stage 2. Probit A X2 L if D==1, vce(.)
Stage 3. Peobit S X3 M if A==1, vce(.)
Note: L and M are estimated inverse Mill’s ratio from stage 1 and stage 2, respectively to correct some selectivity biases. However,
I am struggling to use David Roodman’s cmp stata command. Is there any one to help me out with the appropriate cmp syntax please?
Best Regards,
seid
Comment