Search

This paper is an empirical investigation into the validity of some theses posed in [5]. There it is stated (1) that the individual is able to evaluate his income level on a zero-one scale in a cardinal way, and (2) that the resulting evaluation U(y) of a steady income system at a level y under specific assumptions is approximately equal to A (y: uo2) where A is a lognormal distribution function with parameters u, o2. In this paper these statements are empirically investigated on the basis of a consumer survey conducted by the Belgian Consumer Union. Moreover, approximate relationships between the values of u, o2 and several variables, such as income, family size, age, job, stability of income, etc., have been derived. It turns out that o2 is an individual constant, but that u can be explained for a considerable part by the variables family size and income level. Most notable is the dependence of the individual's welfare evaluation or income, a phenomenon that is called "preference drift." It might lead to major changes in parts of the existing economic theory on consumer behavior and social welfare.

In this paper the consumer problem is considered as a decision problem. At first consumer behaviour is described by the usual standard model, well-known in economic theory. The main criticism of this model is that it presumes a consumer acting rationally in the sense that he can choose simultaneously from a multitude of commodities, whereas ordinary intuition tells us that even the choice from a small set of alternatives is very difficult. Things are becoming even more difficult if goods available at different times are considered as different options, and when randomness is introduced in terms of 'contingent claims'. We call this the assumption of 'irrational rationality'. Things become more natural if the ordinal utility concept is replaced by a cardinal concept. This is introduced as a normed measure W on the service space of commodities Ω. Then it becomes possible to fragment consumer behaviour into several consecutive stages. At the first stage the consumer decides on how much is consumed and how much is saved, at the second stage a broad distribution is made over a few subgroups of expenditures, while in later stages money is gradually 'differentiated' up to the point where it is actually spend and transformed into purchases. If the consumer acts in this way, he most probably will reach a suboptimal solution when compared to the 'master plan' where everything is planned in one stroke. However, it is the more realistic description, since it is rational for the consumer to consider one decision at a time. We call this feature 'rational irrationality'. Probability and time preference are introduced by introducing new coordinate spaces. The Cartesian product of the state space S, the commodity space Ω and the time space T, with their corresponding marginal measures P, W, T, becomes the natural space to describe consumer behaviour.