MULTIPLE TIME SERIES ANALYSIS OF COMPETITIVE MARKETING BEHAVIOR (ECONOMETRICS, MODELS)

HIROKAZU TAKADA, Purdue University

Abstract

The primary objective of this dissertation is to investigate the competitive behavior of firms in one industry. The focus is on identifying the direction of causality between variables. The functional forms of models are specified without a priori assumptions by the framework of multiple time series analysis. Data are available on a large number of variables for the industry. The product category belongs to a general class of products sold predominantly in retail food outlets. A group of the bivariate market share models are developed for each firm, followed by the construction of the models for the pairs between the significant decision making variables. These models provide the information to detect the direction of causality between the variables and to describe the competitive structure of the market. Analysis of the models indicates mixed results. The market share of firm 1, for example, has an effect on the relative price and the advertising expenditure of the firm. Local advertising activity, on the other hand, has an impact on the market share. The results are found to be consistent to the decision making processes of the firms. Interfirm effects and intrafirm effects are also identified to depict the competitive structure of this industry. The feedback effect among the variables is not found in this study suggesting that the bimonthly data interval is short enough to detect the unidirectional causality. Comparisons of the results with the econometric study indicate the differences in the direction of causality between the variables and in the functional forms of the models. The thesis concludes with the discussions of limitations of the study and suggestions for further research.

Degree

Ph.D.

Subject Area

Marketing

COinS