We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Advertising and Market Concentration: A Re-examination of Ornstein's Spurious Correlation Hypothesis.
- Authors
Bradburd, Ralph M.
- Abstract
In a recent article in this Journal [1] and in a more recent book [2], Stanley Ornstein examined the relationship between advertising and market concentration--with advertising the dependent variable--and concluded that the observed significant positive relation between the two is probably the result of spurious correlation due to "large firm effects" [1, 900-901]. His conclusion rests in large part on his finding a positive and significant relation between advertising and market concentration in samples of producer goods industries, a category where buyers should be too well informed for advertising to be significantly related to market power. <BR> In this essay, we will show that Orntein's results can be explained by the substantial portion of some "producer goods" industries' output actually going to consumer demand. We will also show that a log linear specification provides a better fit than the linear in estimating the relation between advertising and market concentration. <BR> We have shown in this essay that Ornstein's finding of a positive and significant relation between advertising and concentration in producer goods subsamples can be explained by the substantial portion of some "producer goods" industries' output actually going to consumer demand. When we eliminate those industries from the producer goods subsamples, the positive relation is no longer significant, and in some cases is insignificantly negative. Since Ornstein's hypothesis of spurious correlation was based primarily on the empirical foundation of a positive and significant relation between advertising and concentration in producer goods subsamples, we must conclude that the Ornstein hypothesis is unsupported. <BR> We have also shown that a log linear specification provides a better fit than the linear in estimating the relation between advertising and concentration, and finally, that this relation, in samples of consumer goods industries, is not insubstantial.
- Subjects
ADVERTISING; INDUSTRIAL concentration; MANUFACTURING industries
- Publication
Southern Economic Journal, 1980, Vol 47, Issue 2, p531
- ISSN
0038-4038
- Publication type
Article
- DOI
10.2307/1057544