Panel data models of differentiated product markets
Panel data models of differentiated product markets
How do product demand and market shares react if product prices are changed? This traditional question of Industrial Organization has experienced a rebirth in recent years. Starting from the hedonic price measurement, economic models have developed that allow the structural analysis of market structures. Even though the literature on the modelling of the demand for differentiated products made considerable progress with respect to the theoretical foundations and the empirical methods, two main deficits still exist: first, panel data econometrics is rarely used and second, there is a general lack of empirical studies.This research project aimed at removing both deficits. It was conducted in close cooperation with Ariel Pakes (Harvard University und NBER) and has lead to many interesting results. From a methodological perspective, it has been shown that the identification of certain parameters can be achieved by using the panel structure of data explicitly so that the use of instrumental-variables methods becomes obsolete. From an economic perspective, this project has produced new knowledge concerning the relatinship between virtual and real markets. Details on that count can be looked up in the research report on the Consequences of digitalisation on classical markets.