GMM says "near singular matrix"

[logsu]

Hi experts,

I want to estimate a complicated non-linear equation with GMM, but it replies "near singular matrix", could you tell me why this happens?

the problem is about intratemporal and intertemporal substituion of consumption.

(I put the original equation, a picture cut from the paper in the attachment)

the equation I put into eviews is:

(p/(r*p(+1)))*((1.5*m(+1)^(1-1/c(1))+n(+1)^(1-1/c(1)) )/(1.5*m^(1-1/c(1))+n^(1-1/c(1))))^((c(2)-c(1))/(c(2)*(c(1)-1)))*(m(+1)/m)^(-1/c(1))-1/c(3)
where c(1) --c(3) are parameters,

instrument varables: C , m(-1)/m(-2), n(-1)/n(-2), p(-2)/(r(-2)*p(-1)) , m(-1) , n(-1). as used in the paper.

The paper is also attached below.

Thank you very much!!