Error using CMP for estimation of SURE model with endogenous fixed effects

Dear Stata-listers,

I have the following problem in using the CMP command (CMP most updated version, Stata 13 MP, system = Windows).


I need to estimate the following system of equations

y1 = alpha1_J + beta1* X + error1 , J = 1...300 [main outcome equation]

alpha1_J = gamma1 * Z + gamma2 * Z * X [multinomial (probit) first stage equations]


In short, I have a problem when the set of dummies alpha1_J in my outcome model are potentially endogenous.
In order to control for this potential endogeneity, I want to estimate a multinomial probit first stage, with instruments Z.

Using CMP, I assume that this problem should be feasible, if I try to estimate:

y1 = alpha1_J_hat + beta1* X + error1 , J = 1...300 [main outcome equation]
alpha1_J = gamma1 * Z + gamma2 * Z * X [multinomial (probit) first stage equations]

where now the alpha1_J_hat are the references latent variables coming from the first stage multinomial probit.

The problem is that, when I try to do so, by writing

cmp ( eq1: y2 = ( X )##( Z ) ) ///
( eq2: y1 = X eq1# ) , ind( $cmp_mprobit $cmp_cont )

I get the following error:

Equation eq1 not found.
invalid syntax
r(111);

Can you please help me to understand if want I am trying to do is feasible by using CMP, and whether I am making any mistake in the syntax?

Thanks in advance,
Giuseppe