nonlinear igmm estimation with interaction of coefficients

[b_smith_aub]

I have an equation I want to estimate using the nonlinear iterated gmm:

yt = a + (1 − p)bX1 + (1 − p)cX2 + pX3 + et.

The code I have written and run in STATA is:

gmm (Y-({a}+{b}*X1+{c}*X2+{p}*X3)),igmm instruments (Some matrix of instruments).

The returned results are b_hat=(1-p)b and similar for c. Is there a way to dis-aggregate these coefficients? Also, my professor thinks that I am not using a non-linear gmm estimator based on my returned coefficients. Am I using the wrong syntax/code for a nonlinear model?

Thanks in advance for your help.

[Tom Jefferson]

Since gmm allows you to specify nonlinear equations, there's no need to estimate a linear equation and then recover p, b, and c after the fact. Instead, try this:

gmm (Y - ({a} + (1-{rho})*{b}*X1 + (1-{rho})*{c}*X2 + {rho}*X3)), igmm instruments(...)

Hope this helps.

Tom