Estimating the long-term effects of advertising on sales: A co-integration perspective

Abstract

Co-integration econometrics have important theoretical advantages over more traditional approaches to estimating the long-term effects of one variable on another. When advertising and sales data are co-integrated, adaptations of common econometric procedures may be used to estimate the long-term effects directly rather than inferring them from short-term relationships. This paper presents a method for detecting and estimating co-integrating relationships using ordinary least-squares regression procedures. The method is illustrated with the well-known Lydia Pinkham data from 1907 through to 1961. The results show that there was a strong positive relationship between Lydia Pinkham advertising and sales in the long-term. Implications of the method and findings are discussed.