Using mata functions to define GMM equations

[Geoff Zheng]

I'd like the use the cumulative distribution function of the multivariate normal as a moment condition in a GMM estimation. It seems like there is a mata command for this:

mvnormal(L, U, M, V)

, but I don't know how to use this function in vanilla stata. For many reasons, I'd like to use the GMM estimation command rather than write it myself in mata. Is there a way to do this?

If more detail is necessary, the parameters I'd like to estimate are parameters of the multivariate normal, M and V

[Geoff Zheng]

For posterity: the solution I figured out was, rather than to specify equations using substitutable expressions, specify moments using a custom program (it's all documented in the gmm helpfile). Inside of the custom program, I was free to call mata functions to my heart's content.