[Limdep Nlogit List] Grouped data models

[jkirkley at vims.edu]

I am not even sure I am asking this question correctly (and no offense to authors of econometric software and texts), but does LIMDEP (9.0) have an option for calculating the variance involving generated regressors (another term is apparently constructed regressors).  To me, this seems like a proxy variable situation, but others have informed me it is different.  I have an equation (atually a production model) in which I do not have observations of one variable on the right hand side (fishing effort).  This variable is being predicted from another equation.  One econometric text indicates that the conventional variances for the coefficients are incorrect and then offers a lengthy and to me complicated procedure for obtaining the correct variances.

Does anyone know whether or not LIMDEP 9.0 has options to calculate this variance or has anyone prepared code to calculate the variance.

Jim Kirkley