Discussion

Ward A. Hanson, University of Chicago
[ to cite ]:
Ward A. Hanson (1987) ,"Discussion", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 308.

Advances in Consumer Research Volume 14, 1987      Page 308

DISCUSSION

Ward A. Hanson, University of Chicago

The context of the Gaski paper is the old suspicion that much of marketing can be considered waste and inefficiency, at least from the point of view of the consumer. This view arose most strongly in the populist reactions to the growth of mass marketing. It has never totally been displaced in the general public's mind, but probably has diminished in importance relative to concerns about the impact of advertising.

The Gaski paper argues that a measure of relative market "efficiency" is important to understand consumer satisfactions with the bundle of products they consume. There are several possible uses of such a measure. The first would be to detect "problem areas" where a particular brand provides low satisfaction. Another would be to identify disparate consumer segments. A final use would be to provide policy oriented "benchmarks" for such issues as taxation.

There are several problems with the method suggested by Gaski. The first is that "efficiency" is very difficult to assess using only total consumer valuation and expenditure, especially between very different supply conditions. It is not particularly fruitful to compare the expenditures on consumer goods between an urban area such as Toronto with the outback of the Northwest Territories. Even the most efficient marketing system must reflect the higher costs of doing business.

A second problem area is that the measure provided uses average satisfaction, while the most important decision criteria is marginal satisfaction. For consumer choice between products it is the marginal utility curve (weighted by price) which provides the means to allocate funds. This can also be seen by comparing a very steep MU curve with a flat MU curve; both could have the same average satisfaction while having quite different implications.

A third problem is that the measure provided can be seriously skewed by differences in the economic positions of the respondents. Consider an extremely wealthy person versus someone poor. In this case we would have VA($) - 0 and VB($) = k, where the index B refers to the poor consumer. But then

(VA(P) - VA($)) / VA($) - *.      (1)

These are some of the main reasons why researchers utilize the concept of consumer surplus to handle these efficiency issues.

Given the definitional problems, the main use of the measure proposed by Gaski appears to be to identify different consumer segments and to serve as an adjunct to such approaches as conjoint analysis or "Defender" type models.

The Ratchford and Gupta paper also considers "efficiency", but in a quite different sense. Here the issue is whether choice theory can be used to estimate how costly it is to consumers when they have incomplete information about the purchase options available to them. Combined with the Srinivasan paper they represent a healthy step in the direction of merging competition and behavioral models. There is much fruitful work to be done in this area, as a consideration of the choices and actions by firms in response to the decision rules of consumers (whatever they are) has the possibility of combining much of the work in Marketing, which now is too often restricted to discussions within the different camps.

The goal of these papers is to estimate the losses to consumers which stem from two sources; the opportunity loss from buying the wrong product and the search costs incurred during shopping. As in the earlier paper, the most serious technical questions arise in trying to measure consumer losses as opposed to measuring the underlying demand. The demand determination follows some straightforward logic. The utility of the product minus its price gives a value distribution across products. Differences in individual position and search costs determine how much search each consumers will engage in. These together determine expected demands for the products and exacted prices paid by consumers.

The most striking of the empirical findings in the Ratchford and Gupta paper is the magnitude of the search costs. They seen very high. There are two probable reasons for this. There is a implicit assumption in the authors' calculations that the price elasticities of consumers are symmetric around the average elasticity. It is this assumption, rather than their stated reliance on large N, which gives the particularly simple expression for costs of search. This assumption will tend to understate the true "average" elasticity, which consequently causes the search estimates to be too hi p. Second, the elasticities calculated using the Consumer Reports data may be biased.

A more restrictive problem with the approach is the specification of maximum opportunity loss Vmax Unless the product category is hi ply segmented and all consumers consider all of the products (despite historical consumption patterns) the definition of the appropriate consumption benchmark will be difficult. If the product category isn't cleanly split then it will be very difficult to get empirical estimates of the welfare losses.

The Srinivasan paper contains a number of useful behavioral qualifications to the approach used by Ratchford and Gupta. First, the gain calculation should be based on perceived benefits rather than on ex post actual benefits. Second, the gain calculation should be a search over the consumer's relevant set, not necessarily the total set. Third, risk neutrality is a strong assumption. And finally, internal search parameters will influence loss calculations. Unfortunately, the pilot nature of the surveys presented in Srinivasan paper are not yet sufficient to adequately test the importance of these qualifications.

----------------------------------------