The RATS Software Forum

Hello,
I am trying to estimate a Simultaneous Equation Model using the SUR function with robusterrors.
I first identified the instruments, I then built the model using the equation function and the group function. I then estimated the model using the SUR function. I noticed that whatever the first equation in the SEM is, its coefficients are always insignificant. Meaning that the order of the equations in the VAR is affecting the var-cov matrix of the coefficients. The coefficients are not affected and I always get the same values.
Any idea why this is occuring?
Thank you for your help.
Charbel

We can't tell for sure unless you post the whole program and data. However, the key matrix in the robust standard errors will typically not be full rank if the number of instruments x the number of equations is larger than the number of data points.

Thank you Tom,
Attached the code and the data.
I want to estimate a simultaneous equation model using a 3SLS estimator. As explained in my first post, the standard deviations of the coefficients are sensitive to the order of the equations in the SEM (Group function). The number of instruments x the number of equations is less than the number of observations. Your help is appreciated.

Your model isn't identified---all your equations are identical (other than normalization).