Review of Industrial Organization 18: 417–426, 2001.
© 2001 Kluwer Academic Publishers. Printed in the Netherlands.
A Cointegration Analysis of Advertising and Sales
Department of Economics, The Management School, Lancaster University, Lancaster LA1 4YX,
Abstract. It is argued that the nature of the industry level relationship between advertising and
sales can give some indication of the form of competition in an industry. Hence, this paper examines
whether there is a long-run, stable, equilibrium relationship between advertising and sales for food
and soft drinks industries. Results suggest that the variables are non-stationary, but do not contain
seasonal unit roots. Cointegration is not identiﬁed between soft drinks industry advertising and sales,
which, together with the results of differenced variable regressions, suggests that rivalry between
ﬁrms in this industry may be intense.
Key words: Advertising, cointegration, sales, unit roots.
JEL Classiﬁcations: C2, C5, L1.
The nature of the industry level relationship between advertising and sales can
give some indication of the form of competition in an industry. If industry de-
mand has not yet reached saturation point, a positive, stable, long-run relationship
between advertising and sales may reﬂect ﬁrms’ use of advertising primarily to
attract new consumers and to increase purchases by existing consumers. Alternat-
ively, in mature, saturated industries, advertising may be used to maintain customer
loyalty and/or redistribute market shares. Rivalry between ﬁrms may be intense and
cointegration between advertising and sales will be less likely. Hence, this paper
examines whether advertising and sales in the U.K. food industry and the more
closely deﬁned soft drinks industry are cointegrated, in order to help determine the
nature of competition in these industries.
In existing research, Baghestani (1991) and Zanias (1994) used the U.S. Lydia
Pinkham Company data to conﬁrm that a long-run equilibrium, cointegrating, re-
lationship existed between advertising expenditure and sales. The results of error
correction mechanism (ECM) modelling indicated that movements in advertising
I would like to thank Lisa Hart, Geraint Johnes, Denise Osborn, David Sapsford, Mohammed
Salisu, David Young, the editor and two referees for helpful comments at various stages of the writing
of this paper. The usual disclaimer, of course, applies.