[PEN-L:2357] Re: Giffen goods

[Terrence Mc Donough <TERRENCE.MCDONOUGH@xxxxxx>]

Thanks to Eugene C. for the citation.  Other people have observed the
apochryphal character of the famine Giffen goods story, so perhaps I
am not to blame for Barro.  :]  Along the lines of our JEI article,
Barro's assertion of the historical invariance of the law of demand
is philosophically problematic for neocl. economics because it raises
the question of the non-falsifiability of utility theory.  The "law
of demand" is one of the few empirical predictions of utility theory.
The Irish Giffen example is generally raised as an example of the
possible falsifiability of the law of demand which is then
immediately explained through the operation of the negative income
effect.  The fact that it is patently ridiculous that the Irish
increased their consumption of potatoes due to an increase in price
in the midst of a potato famine! is apparently beginning to sink in.

But all is not lost.  Experimental economics to the rescue.  One
study (Silverberg et al. 1987) induced hungry animals to consume more
of an unpleasant food alternative when its "price" rose because their
"income" was too low to provide sustenance otherwise.  To quote a
couple of paragraphs from our JEI article:

Battalio et al (1991) set out to confirm Silverberg et al's study by
reproducing a similar economic environment in their laboratory.  They
motivate their study by observing that "in constructing such an
environment within the bounds set by legal and ethical considerations
and our limited research budget (!), we have used rats as
individuals." They obtain similar results by depriving the rats of
liquid sustenance.  The conclude in part, "we believe that the
results of this experiment are important, not because they suggest
that Giffen goods are likely to be observed in aggregate market
demand, but because of their ability to test between competing
representations of consumer behavior and to validate the
counterintuitive prediction of the Slutsky-Hicks theory regarding the
existence of Giffen phenomena."
    Such attempts are not only a testimony to the enduring legacy of
the Giffen legend.  More generally they reflect a greater preference
for finding or contriving  data than for rejecting theories for which
there exists no supporting evidence.

Terry McDonough