A Stochastic Brand Choice Framework for Econometric Modeling of Time Series Market Share Behavior

Abstract
The two main purposes of this article are (1) to develop a model of individual behavior in the choice of brands which can be aggregated over consumers to produce a model of market share behavior of brands and (2) to show how modern multiple time series analysis can be applied within the general structure we develop to choose among specific alternative models. Starting at the individual level, we devise a stochastic brand choice model. The character of consumer heterogeneity is specified, permitting aggregation. The aggregation allows a model of aggregate brand switching to be derived as well as a model of market share behavior through time. Panel data are analyzed in terms of brand switching and market share behavior.