Advertising and Profitability: Further Implications of the Null Hypothesis

Abstract

THIS essay is concerned with studies of the impact of advertising outlays on profitability via their effects on entry barriers or ease of collusion. The null hypothesis is that no such effects exist. A number of authors have argued the plausibility of the alternative hypothesis that advertising does affect profitability in this fashion.1 Following the pioneering work of Comanor and Wilson [2], a number of attempts have been made to compare these hypotheses by means of cross-section regressions in which profitability, as measured by the ratio of profit to assets or equity, is the dependent variable, and advertising intensity, as measured by the advertising/sales ratio, is one of the independent variables.2 If received economic theory is to be taken seriously, I would hold that profit maximization must be part of the maintained hypothesis in any study of this sort. On this assumption, I have argued elsewhere [6, pp. 223-6] that the existence of a correlation between advertising intensity and profitability cannot serve to cast doubt on the null hypothesis. I indicated that such a correlation is predicted by a model in which price and production cost are determined exogenously, while advertising levels are chosen to maximize profit. The greater the price-cost difference, the higher are both profits and advertising in that simple model. Causation, however, runs from profitability to advertising, not the other way around. This argument is considerably amplified in what follows. In an interesting recent paper in this journal, Vernon and Nourse [I I] attempt to distinguish between these competing hypotheses by considering both firm and industry advertising intensities. They argue that if advertising can erect barriers to entry, the magnitude of this effect must logically depend on industry advertising intensity. Thus, their alternative hypothesis implies positive correlation between industry advertising intensity and firm profitability. On the other hand, they contend that the null hypothesis (with the assumption of profit-maximizing behavior) implies a correlation between firm profitability and firm advertising intensity. For a sample of 57 large U.S. firms primarily engaged in the manufacture of consumer non-durable goods,