THE effect of concentration on advertising has long been the subject of debate. On the theoretical side two approaches may be distinguished. Starting with Kaldor [I 949-50], an analysis in which the assumed purpose of advertising is to make profits, but not necessarily maximum profits, has been developed by Sutton [I974]. Additionally, building on the original work of Dorfman and Steiner [I954] and others, Cable [I972] has provided a profit maximising model of the way in which concentration affects the optimal amount of advertising. Both approaches predict a systematic, non-linear, relationship between concentration and advertising. In Sutton's terms this is rationalised by pointing to the different opportunities and incentives for advertising under different market structures; Cable's analysis is based on various hypotheses concerning the magnitudes of the price and advertising demand elasticities and reactions under different market structures. Basically, they both argue that it is difficult to internalise the benefits of advertising in low concentration industries, and monopoly gives no incentive for competitive advertising. In between, oligopoly provides the structure most conducive to advertising. As reviewed in a later section, however, the empirical evidence is somewhat conflicting. Some reject the hypothesis of a positive relationship, but other evidence has confirmed the existence of a linear or inverted U-shaped relation. Three main issues have been examined in these studies; the existence and shape of the underlying relationship between concentration and advertising; the influence of product and market characteristics; and more recently the direction of causality. The second though has probably been the source of most debate. The issue has been that differences in product and market characteristics will, ceteris paribus, affect the amount of advertising carried out. Samples are then chosen (and/or variables included) which attempt to account for these differences. In particular, researchers generally attempt to distinguish between consumer and non-consumer goods samples. It is argued in this paper that such apparently judicious selection of samples may be misguided and seriously affect the results. The main reason is that, even if it were possible to define a consumer good in principle, in general, firms as well
Read full abstract