Estimation of censored demand equations from large cross-section data

The article develops an alternative econometric methodology to estimate a s ystem of censored demand equations using a large cross-section data from Co lombian urban households. The approach preserves the behavioral information expressed by zero expenditures and conforms with the requirements imposed by consumer theory in a way consistent with the random utility hypothesis. We motivate the choice of the Tobit model as a statistical representation o f consumer behavior and introduce the methodology by specifying the AIDS mo del modified according to both a translating and scaling demographic transf ormation. We propose to estimate each demand equation in unrestricted form using the jackknife technique. We then recover the demand parameters imposi ng the cross-equations restrictions by using minimum distance estimation. T he empirical results of the censored demand system for specific households of policy relevance are reported.