Diagnostics, Expectations, and Endogeneity
Published 2/1/2005
Author: Philip Hans Franses
View this content
Executive Summary
The author argues that for econometric models to be really useful to answer managerially relevant “what if” questions, three important issues must be addressed. The first concerns the statistical properties of the econometric model. There are several such properties, and sometimes it seems that some of these are overlooked in practical work. The second issue recognizes that consumers and competitors might make predictions about the consequences of marketing mix actions, and in fact they may change their behaviour due to these expectations. Neglecting this issue leads to biased estimates of the effects of a marketing strategy. Finally, the third issue concerns endogeneity, which, loosely speaking, is the possibility that marketing mix decisions depend on managers’ views of consumers’ and competititors’ expectations. This issue makes it more difficult to separate the net effect of the marketing mix from the observed effect. Fortunately, as the author argues, the quality of data in present-day marketing research should enable modelers to put forward models that address all three issues in a satisfactory way.
Biography
Philip Hans Franses is Professor of Applied Econometrics and Professor of Marketing Research, Erasmus University Rotterdam. Currently, he is Director of the Econometric Institute. His research interests include marketing research models and techniques.
J Marketing Research, Volume 42, Number 1, February 2005
View Table of Contents.