Hello,

I have a system of equations of the following form:

p = beta * X + error_1
q = alpha * beta + error_2

I observe X, p, and q. I do not observe beta, and estimate it in a first regression. I want to use the second regression to estimate alpha.

the two datasets corresponding to the two regressions are distinct, and the error terms error_1 and error_2 are uncorrelated.

any help on how to conduct this with SAS would be highly appreciated, thanks!

PS: In the original form, the first regression is a logit. I would be satisfied to run a procedure where both regressions are linear, but if I could keep the logit specification in the first regression this could be even better.