Printer Friendly

More on the search for Giffen goods.

In a recent paper in this journal (Gilley and Karels |1991~), we showed that observed behavior consistent with the Giffen paradox can occur largely independent of the form of the utility function. Our model took the potatoes and meat fable of the Irish potato famine literally. We added a nutrition constraint to the consumer's utility maximization problem, and found that with two binding constraints the consumer's feasible set contains a corner that leads to demand curves with upward sloping segments.

Papers by Johnston and Larson |1994~ and Wichers |1994~ in this issue extend and sharpen our knowledge of Giffen goods. The first paper builds on our model by extending the potential "arenas" where economists should look for evidence of Giffen behavior. The second paper raises some philosophical issues concerning the definition of Giffen goods and the role of a second constraint in the utility-maximizing problem. These issues apply to both our paper and Johnston and Larson's.

Johnston and Larson employ a model that is essentially the same as ours, except that the second binding constraint places an upper limit (local satiation) instead of a lower limit (a minimum nutritional requirement) on the optimization problem. In the Gilley-Karels model, a Giffen-like good is obtained because the consumer's income is so limited as to preclude survival on the consumption of meat alone. Potatoes exhibit Giffen characteristics because they are more efficient than meat at providing nutrition. When the price of potatoes increases, this nutritional efficiency causes consumers to purchase more potatoes to avoid violating the nutritional constraint. Johnston and Larson obtain Giffen-like behavior in their model for very similar reasons. As local satiation is obtained, the least-cost source of satiation drives the consumer's choice. When the price of this good falls, the consumer continues to be satiated and purchases less of the good that is more efficient in providing satiation. Johnston and Larson interpret this local satiation as a type of time rationing problem, arguing that there is an upper limit to the amount of food that a consumer can ingest in a short period. This was essentially the mathematical problem addressed in earlier papers by Goldfarb |1977~ and Wichers |1979~. We were unaware of these papers and failed to cite them in our original paper. What is interesting and important about Johnston and Larson's paper is that it provides another very plausible set of circumstances where Giffen-like behavior could be found.

One of the more interesting aspects of the Giffen problem is the dearth of published empirical evidence supporting the existence of such goods. Dougan |1982~ quite convincingly points out that little evidence would come from an analysis of market demand curves, Johnston and Larson suggest, however, that a substantial body of econometric evidence on Giffen goods exists in the agricultural economics literature. Several studies find evidence of a positive own-price elasticity of demand. More evidence is likely to exist, but is not often published since the result is usually dismissed as an identification problem in the econometric study. Johnston and Larson's interpretation of the consumer's maximization problem suggests we should not dismiss those studies so readily. In fact, their study provides experimental economists with a new set of conditions to test for evidence of Giffen behavior.

The Wichers comment has two important messages. First, the models Gilley and Karels |1991~ and Johnston and Larson |1994~ use to obtain Giffen-like behavior do not follow the standard Hicksian single constraint optimization problem. The Giffen result obtained in traditional utility theory comes from a negative income effect that is sufficiently strong to override the substitution effect. In the Gilley-Karel and Johnston-Larson models there is no substitution effect because of the binding second constraint. The Giffen-like behavior is due exclusively to a negative income effect that results from the relative efficiency of a critical characteristic of the goods. Hence, the upward sloping demand curve is obtained as in traditional utility analysis, but by eliminating the substitution potential in the problem. This does not invalidate the conclusions generated by the two-constraint models. The question of interest is how to define a Giffen good. In the original sense of Hicksian utility analysis, potential Giffen goods are generated by very unusual preferences. In the Gilley-Karel and Johnston-Larson models, all goods can potentially be Giffen goods. A distinction between these two types of Giffen goods is that the Hicksian Giffen good is a global concept, whereas the Giffen-like behavior we have described should disappear when the second constraint is no longer binding.

An equally interesting point in both the Gilley-Karels and the Johnston-Larson models is the intuition behind a strongly inferior good. Following classical utility theory, the quantity purchased of an inferior good is inversely related to the real income of the agent. According to the standard explanation in many textbooks, consumption of inferior goods rises as a consumer's income goes down because the consumer can no longer afford the preferred alternative, forcing a substitution toward cheaper goods. That explanation is precisely the cause of inferior goods in our model, but it cannot be an explanation in Hicksian utility analysis because there is only a budget constraining individual behavior. The agent can still afford both commodities, but chooses to buy less of a commodity because preferences dictate such a result to be consistent with utility-maximizing behavior. There is little intuition one can obtain about inferior goods in a single-constraint model. The second point in Wichers's paper concerns the shape of the indifference curves around the area of the binding nutritional constraint. Wichers suggests that because the second constraint is imposed by nature instead of purely economic forces, it is invalid. He argues instead that the indifference curves will already reflect the nutrition constraint by flattening out as the minimum nutritional level is approached. In other words, he argues that the consumer's indifference curve will coincide with the nutritional constraint we imposed.

Two comments seem reasonable here. First, we traditionally think of a utility function as a mapping of the consumer's preferences. The underlying preference ordering serves as a ranking instrument by satisfying a group of mathematical and behavioral conditions (completeness, transitivity, etc.). Viewed purely in this manner, the mental experiment used to achieve this ordering does not consider any other influences. The nutritional constraint is necessary to specify where the ranking process is applicable.

If the indifference curves flatten out as suggested by Wichers, the goods exhibit greater and greater substitutability. As the indifference curves flatten out, the likelihood of a corner solution in the traditional Hicksian model increases. Ultimately, the consumer would only consume one of the goods--the good that provided subsistence at the cheapest cost. Thus, we should observe consumers subsisting on only one of the goods and no Giffen-like price response would be obtained. Whether we observe such behavior is then an empirical question. The best place to observe such behavior would be in a laboratory setting. As cited in our paper and the Johnston and Larson's paper, there is a substantial body of evidence that Giffen-like behavior is observed among animals in laboratory situations. A recent paper by DeGrandpre, Bickel, Rizvi, and Hughes |1993~ confirms similar behavior among human subjects in a cigarette smoking study.

The original purpose of our paper was to attempt to answer Boland's criticism of utility theory--namely that if we cannot dismiss the existence of Giffen goods we should at least be able to predict the conditions that generate them. Johnston and Larson's paper adds another place to look for Giffen goods. In both the Gilley-Karels and Johnston-Larson models, Giffen goods arise not from traditional Hicksian analysis, but from the imposition of a second binding constraint. This results in goods having efficiency characteristics in the production of nutrition or satiation. In this sense, the models are more closely aligned with the attribute analysis of Lipsey and Rosenbluth |1971~ who suggested that Giffen goods can arise with greater likelihood than traditional utility analysis predicts. These papers, along with the growing experimental work, confirm their suggestion and provide consumer theorists with a likely set of places to search.


Boland, L. A. "Giffen Goods, Market Prices and Testability." Australian Economic Papers, June 1977, 72-85.

DeGrandpre, R. J., Warren K. Bickel, S. Adu Turab Rizvi, and John R. Hughes. "Effects of Income on Drug Choice in Humans." Journal of the Experimental Analysis of Behavior, May 1993, 483-500.

Dougan, William R. "Giffen Goods and the Law of Demand." Journal of Political Economy, August 1982, 809-15.

Gilley, Otis W., and Gordon V. Karels. "In Search of Giffen Behavior." Economic Inquiry, January 1991, 182-89.

Goldfarb, Robert S. "The Demand Curve for an 'Otherwise-Normal' Good Can Have an Upward Sloping Segment Under Coupon or Time Rationing." Atlantic Economic Journal, December 1977, 68-70.

Johnston, Richard S., and Douglas M. Larson. "Focusing the Search for Giffen Behavior." Economic Inquiry, January 1994, 168-74.

Lipsey, Richard G., and Gideon Rosenbluth. "A Contribution to the New Theory of Demand: A Rehabilitation of the Giffen Good." Canadian Journal of Economics, May 1971, 131-63.

Wichers, C. Robert. "Upward Sloping Demand Curves for Normal Goods." Atlantic Economic Journal, July 1979, 85-87.

-----. "In Search of Giffen Behavior: Comment." Economic Inquiry, January 1994, 166-67.
COPYRIGHT 1994 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1994 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Gilley, Otis W.; Karels, Gordon V.
Publication:Economic Inquiry
Date:Jan 1, 1994
Previous Article:Focusing the search for Giffen behavior.
Next Article:Dealing with arguably legitimate exam gripes: a possible Pareto improvement.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters