Analysis of Price Level Heterogeneity across Households based on the Geary-Khamis Price Index

We construct a household price index based on household-level scanner data (homescan data). Instead of the conventional Paasche type price index, in this paper, we apply the Geary-Khamis (GK) method that is commonly used to measure the values of different currencies in international trade. The mathematical properties of the GK price index help identify the complex factors behind household-level price differentials. We ?d that households that buy cereals, manufactured foods, and daily products at weekend bargain rates, reduce their price levels signi?antly. Further, households with children make purchases at lower prices than households without children through frequent bulk purchases at bargain sales. Certain types of shopping behaviors of the elderly lead them to make purchases at higher prices than the young. Speci?ally, shopping at specialized stores, not shopping at home improvement stores, spending less on Sundays, and purchasing the same goods repeatedly increase the price levels for the elderly.