Combining micro and macro data in hedonic price indexes

Abstract

This paper proposes arithmetic and geometric Paasche quality-adjusted price indexes that combine micro data from the base period with macro data on the averages of asset prices and characteristics at the index period. The suggested indexes have two types of advantages relative to traditional Paasche indexes: (i) simplification and cost reduction of data acquisition and manipulation; and (ii) potentially greater efficiency and robustness to sampling problems. A Monte Carlo simulation study and an empirical application concerning the housing market illustrate some of those advantages.