ESTIMATION AND TESTING WHEN EXPLANATORY VARIABLES ARE ENDOGENOUS - AN APPLICATION TO A DEMAND SYSTEM

In a system of linear equations where some of the right-hand-side explanatory variables are correlated with the equation errors a priori information in the form of linear restrictions is used to transform the model into a system where all right-hand variables are exogenous. The resulting equations can be estimated using standard multivariate methods. It is shown how the simple errors-in-variables model places testable restrictions on some of the parameters of the model. The methods are applied to a system of demand equations using quarterly data for the U.K.