Programming GMM

[Dennis]

We have a relatively complex GMM system. We would like to define variables in the GMM system that are then subsequently used in the GMM orthogonality conditions. For example, we would like to define a variable that is a function of parameters and data series. Then we would like to use that variable later in the orthogonality conditions. We know this can be done with likelihood functions object. Can it be done with GMM objects.

For example, suppose we define variabls x, x1, and x2 as a function of parameters (beta, rho, tau) and series (q1, q2, K1, and K2) as follows. Can we do this in the GMM object and then use x, x1 and x2 later in the GMM object to form our orthogonality conditions?
X=(beta(1)*(q1*K1(-1)^tau(1))^rho(1) + beta(2)*(q2*K2(-1)^tau(1))^rho(1))
X1=beta(1)*(K1(-1)^tau(1)*q1)^rho(1)
X2=beta(2)*(K2(-1)^tau(1)*q2)^rho(1)

[EViews Gareth]

No.