A Reconsideration of the Advertising/Consumption Relationships-A Cointegration Analysis Using Taiwan Aggregate Data

Abstract

This paper reinvestigates the relationships between advertising and consumption by employing multivariate cointegration and error-correction modeling. A usual practice of bi-variate model is also modified using VAR (vector autoregression). In this model, wealth and income effects are adopted. The empirical results are consistent with previous research that there is a two-way causality between advertising and consumption but also highlight the income and wealth effects. Cointegration reveals that there exists a long-run relationship among the variables.