Consumer behaviour analysis for luxury goods: a technical note for empirical studies

Today, many econometric tools are provided for studying consumer expenditures. Less attention has been paid to the fact that, when analysing expenditures for superior goods, these tools have to be adapted to some particularities that have a serious impact on the estimation outcome. First, the considered data, in our case household expenditures for jewellery in Spain, exhibit a strong censoring that must not be ignored. Second, our confidence bands show serious nonlinearities for basically all continuous covariates. Finally, it turns out that especially in the case of luxury goods, one has to control for endogeneity of the covariate ‘total expenditure’.