Inflation and inequality bias in the presence of bulk purchases for food and drinks

Abstract

The aim of this article is the estimation of annual food expenditures with limited information about bulk purchases with data from a Spanish household-budgest survey for 1990-1991. Three alternatives are compared. The first, currently used for official purposes, does not use all the information. The second uses all the available information in a rough way. The third assumes a formal model for the unknown frequency of purchases. The three alternatives are compared by a regression model that should be homogeneous with respect to the dummy variables that represent the partial information of the groups and should show a distinct pattern of outliers under each alternative. Finally, we study the effect of the official and the best alternative on food inflation and inequality measures. We find that they lead to similar inflation rates but to different inequality estimates.