Only when these
parameters could be identified separately given the knowledge of the joint probability
distribution of the variables (both policy and non-policy variables), would it be possible
to carry out an econometric analysis of alternative policy options.