TIME SERIES ANALYSIS OF COMPETITION BETWEEN RELATED CATEGORIES AND BRANDS
Abstract
The purpose of this research is to build a model that diagnoses and explains the degree of competition between related categories and brands and that uses as inputs only aggregate time series data. The specification of the nature of the input data is important. From a practical point of view satisfactory methods have been developed using brand switching and/or perceptual data. Approaches and models that use aggregate time series data also have been developed but their practical application presents difficulties, especially when the number of categories and brands involved is large. The model we develop consists of two parts. The first part is a system of market share equations in which each category or brand is modelled by an equation. Unlike the traditional approaches, each equation contains only variables related to the product category or brand of that equation. The second part develops a measure of degree of competition and uses the estimated parameters of the first part to compute a set of correlation coefficients that are likely to include in their range the true value of degree of competition between a pair of categories or brands. This set of correlation coefficients also allows us to look for brand and/or form effects in the competitive structure, and therefore allows us to explain why two brands compete more with each other than with others. Finally we show how this set of correlation coefficients can be used as inputs for perceptual mapping routines and produce a market map. Since we have prior expectations with respect to the signs of each estimated parameter and with respect to the relative values of the correlation coefficients the model provides ample opportunities for securing the quality of the results. In two empirical applications we tested the theoretical implications of the model and cross-validated its results against a traditional econometric model. In both applications the fitted market shares also satisfy the requirements for logical consistency.
Degree
Ph.D.
Subject Area
Marketing
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.