980916

_{[Note for bibliographic reference: Melberg, Hans O. (1999): Consumer theory - As
evaluated by three economists: Blaug, Sugden and Hausman, www.oocities.org/hmelberg/papers/990611.htm]

Consumer theory - As evaluated by three economists

Blaug, Sugden and Hausman

by Hans O. Melberg

Blaug: Does the theory yield testable predictions?

The main purpose of the theory of consumer behaviour is, according to M. Blaug, to
"justify the notion of a negatively inclined demand curve from fundamental and
compelling axioms of individual behavior" (Blaug 1992, p. 140). Unfortunately it
turns out that the axioms (see appendix below) are not enough to exclude the theoretical
possibility of so called Giffen goods, that is a good for which demand increases when the
price increases. The problem is that in theory there is nothing within the axioms that
prevent the income effect from a price change to outweigh the substitution effect. One
could always argue that at least the substitution effect is always negative after a price
change, but this is to loose focus of the real question: how can we justify downward
sloping (uncompensated, Marshallian) demand curves theoretically (not just empirically).

Given the fact that consumer theory cannot derive the downward sloping demand curve from
individually rational behaviour, one is entitled to ask why we should spend so much time
studying the theory. Why can't we just start from the empirical generalization that
usually demand falls when price increases without spending a lot of effort learning about
the unsuccessful effort to justify the downward sloping curve? This was the position of
the economist Gustav Cassell long time ago, and it seems to be close to Blaug's position
today. As he writes:

"The Slutsky-Allen-Hicks decomposition of price responses into income and
substitution effects, and the invariably negative sign of the substitution effect, are the
only substantive achievements of the immense intellectual efforts of literally hundreds of
economists applied for a century or more to the pure theory of consumer behavior"
(Blaug 1992, p. 141)

Sugden: Three (testable?) implications and was it worth the effort?

Robert Sugden (1992) believes that consumer theory has a wider scope than Blaug. His
definition is that "consumer theory is the branch of economics that explains the
demand for goods, and in particular, how demands are influenced by changes in prices and
income" (Sugden 1992, p. 26). In addition to the usual assumptions, Sugden notes that
those trying to establish downward sloping demand curves from assumptions about individual
demand, need to use some kind of aggregating assumption. That is, there is no
"law" that makes sure that the market demand functions have the same properties
as the individual demand functions. A simple example should be enough to convince the
reader. A person standing up in a room full of people sitting down gets a better view of
whatever is happening on the stage. The same is not true in the aggregate: if everybody
stands up they do not all improve their ability to view the events on the stage.

Do the usual assumptions about consumer behaviour lead to any surprising and testable
implications? Sugden lists three implications of the axioms. First, if price and income
increase by the same percentage there should be no change in demand. Technically this is
known as the property of homogeneity. Second, there is the law of demand which states that
unless we have an inferior good we can be certain then demand will increase if price
decreases. (An inferior good is a good which we want less of as our income grows). Third,
there is the property of symmetry. Once again we have to assume away the income effect of
a price change. As long as we do this, the assumptions of consumer theory says that the
cross-substitution effect between good x and y must be the same as the cross-substitution
effect between y and x. Here is an example. Assume the price of cheese increases by $1 pr.
kilo. The increased price of cheese may affect the demand of wine, say the demand of wine
pr. week falls by 2 liters. This implies that the cross-substitution effect between a
change in price of cheese and a change in the demand for wine is: 2 kg. / 1 liter = 2
kg/liter. The property of symmetry then implies that we can use this to predict the change
in the demand for cheese if the price of wine increases (as opposed to the demand for wine
when the price of cheese increased). In short, if the price of wine increases by $1 pr.
liter, then the demand of cheese will fall by 2 kg. Thus, the axioms of consumer theory
have at least three implications: homothetic demand functions, the law of demand and the
property of symmetry.

The three implications raise two issues. First, Sugden echoes Blaug in "wondering
whether all the initial assumptions about the rational consumer were really
necessary" since the theory yields so disappointingly few results. In fact, Samuelson
has showed that it is possible to justify downward sloping demand curves (and the first
two properties listed above) using only the weak axiom of revealed preference (without
using the concepts of indifference and metaphysical preferences). Blaug also notes this,
but he is very critical of Samuelson's claim that revealed preference theory represents a
new and better approach to consumer theory. The theory is no better than the old theory in
establishing the downward sloping demand curve.

Second, we could ask whether the implications really are testable. Blaug notes that we
never observe only the substitution effect. A change in price always has an income effect
and as such the prediction that isolated the substitution effect is always negative is not
testable (as implied by what Sugden the law of demand). Rødseth (1992) argues that we
also need to assume that preferences are stable in order to test the theory since choices
are made at different times. it is not irrational to change your opinions, so preferring x
to y at one point in time and y to x at another point in time is only irrational if we
also assume that preferences are stable. In short, consumer theory is associated with much
work, few implications, even fewer testable implications (unless further assumptions are
added), and the same implications are also derivable from alternative theories (even
random behaviour produces downward sloping demand curves as Becker has showed).

Hausman: A partially successful effort to provide microfoundations

D. Hausman is less critical than Blaug of the force of the conclusions above. "It
would be nice" he admits "to have a quantitative account of market demand, and
it would be nice to make use of a less idealized model than that of a simple consumption
system …" Still, we have enough, he believes, to argue that "the descent
from the level of market generalization to supposed theoretical underpinnings appears to
be a success" (Hausman 1992, p. 39). Although it does not produce testable
predictions of novel facts (as the Popperian Blaug wants), it does provide a plausible but
incomplete microfoundations for aggregate behaviour (which Husman wants).

Conclusion

What does the contrast between these three economists tell us about consumer theory? In
three previous essays I have tried to discuss the value of formal modeling in economics. I
then claimed that formal modeling was justified if it gave us "surprising"
results that were difficult to find using verbal reasoning. Although this criterion is
very much in agreement with Blaug, I am more reluctant to accept his conclusions. For
instance, I think the discovery of the possibility of a Giffen good is a valuable
achievement that was made easier by formalization. One might argue that the theoretical
discovery of the possibility of a Giffen good may not be that important if it does not
correspond to a real phenomena (and economists have to my knowledge so far not been able
to prove the empirical existence of a single Giffen good). This is true. However, one
might try to save the theory, as hausman does, by arguing that the point of a theory is
not only to produce surprising, true and important predictions, but also to improve the
reliability of an explanation by providing the microfoundations on which the macro-theory
is build. That is, if an emphasis on individual rational behaviour makes us more aware of
the causal machinery that creates downward sloping demand curves. This is valuable in
itself. It may not be surprising, but it provides a better justification for the already
existing belief that when price increases demand will usually go down.

Appendix

* There are two sets of standards assumptions. First, about the space of feasible price
and good combinations - no negative prices or quantities are allowed, perfect divisibility
(continuity). Second, we have a set of assumptions about behaviour and preferences - that
people act as if they are rational maximizers of utility and that their utility functions
have certain properties - always increasing with more goods, diminishing marginal rate of
substitutions, independence and so on. If we introduce uncertainty we need to include
additional assumptions, such as the rational formation of beliefs.

References

Blaug, Mark (1992): The methodology of economics. Cambridge: Cambridge University
Press. (esp. ch. 6, The theory of consumer demand, pp. 137-149)

Hausman, Daniel M. (1992): The inexact and separate science of economics.
Cambridge: Cambridge University Press. (Ch. 1 and 2 on rationality and consumer choice)

Rødseth, Asbjørn (1992): Konsumentteori. Oslo: Universitetsforlaget.

Sugden, Roger (1992): "Consumer Theory" in Shaun Hargreaves Heap et al. : The
theory of choice: A critical guide. Oxford: Basil Blackwell, pp. 26-35.

For a more empirically and macro oriented study, see: Angus Deaton (1992): Understanding
Consumption. Oxford: Oxford University Press.

_{[Note for bibliographic reference: Melberg, Hans O. (1999): Consumer theory - As
evaluated by three economists: Blaug, Sugden and Hausman, www.oocities.org/hmelberg/papers/990611.htm]}}