Estimating a Non-linear model by FIML
Posted: Fri Jan 25, 2013 3:51 am
Hi,
I consider a simultaneous equation system of the following form:
y = a_0 + a_1*x + a_2*z + e_1
x = b_0 + b_1*y + b_2*y*w + b_3*q + e_2
where z, w and q are all uncorrelated with the two disturbances. Otherwise, the disturbances are assumed to be drawn from a joint normal. If b_2 = 0, this is just an ordinary simultaneous equation system that can be estimated quite ok by FIML. My question is:
Does EViews also handle the case where b_2 is not equal to 0, i.e. when there is a non-linearity in one of the endogenous variables in the model?
-A-
I consider a simultaneous equation system of the following form:
y = a_0 + a_1*x + a_2*z + e_1
x = b_0 + b_1*y + b_2*y*w + b_3*q + e_2
where z, w and q are all uncorrelated with the two disturbances. Otherwise, the disturbances are assumed to be drawn from a joint normal. If b_2 = 0, this is just an ordinary simultaneous equation system that can be estimated quite ok by FIML. My question is:
Does EViews also handle the case where b_2 is not equal to 0, i.e. when there is a non-linearity in one of the endogenous variables in the model?
-A-