Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Conditional Mixed Process Code

    Dear Statalist,
    I have two equations
    Equation (1): y=a1+B1X1+B2X2 +B3X3+u1 (Y=continuous dependent variable)
    Equation (2): X1 =a2 1y+B2X2 +B3X3+u2 (X1= Binary variable).
    I have tried to implement Seemingly unrelated regression, Ivregress and Simultaneous Equation Estimation using CMP command in Stata . Following are my codes
    SUREG: cmp (y=X1 X2 X3) (X1 =y X2 X3), ind ($cmp_cont $cmp_probit) qui tech(dfp)
    Problem (1): while the results panel for the joint estimation look fine for the regression with continuous dependent variable, stata is not reporting standard errors for the probit regression with binary dependent variable. Moreover, if I remove the dfp options I only get coefficients for both the regressions but no standard errors.
    Ivregress using CMP
    Cmp (X1= z1 z2 X2 X3) ( y=X1 X2 X3 ), ind ($cmp_probit $cmp_cont) qui tech(dfp)
    Problem: I am not sure why I am not getting standard errors for the constant terms. Also I would be grateful if you could kindly explain ` Fitting constant-only model for LR test of overall model fit`.

    Simultaneous Equation Framework
    cmp (y=X1# X2 X3) (X1 1y# B2X2 B3X3) , ind ($cmp_cont $cmp_probit) qui tech(dfp)
    Problem: Not getting standard errors for constants and the same message as IV ` Fitting constant-only model for LR test of overall model fit`.
Working...
X