Box-Cox Type Models: A Proposal for an Estimation Strategy

Abstract

This paper proposes a strategy for evaluating and estimating Box-Cox models which are widely used for example in estimation of hedonic price equations in urban economics. The proposition has in mind some difficulties arising in practical implementation of Spitzer's [1982] conclusion about the equivalence between results obtained with the maximun likelihood technique, by one hand, and the ones obtained with the iterative and non linear least squared technique, by the other. The strategy is shown estimating a sites and services's hedonic price equation.