individual brands or firms. Since the shares of total beer
consumption attributable to these three categories varied over time
and among areas, could one speak at all of a geographic market area
for beer? Consider, for example, the state of Wisconsin. The vast
bulk of beer consumed in Wisconsin was produced there; hence,
Wisconsin might be termed a distinct market. This reasoning,
usefully termed "Little In From Outside" (LIFO), fails to recognize
the importance of exports. In the case of Wisconsin, some 70
percent of beer produced there has traditionally been exported to
other states. An equally significant, but contrary, type of error
is the frequently encountered "Little Out From Inside" (LOFI)
argument. In this instance the error consists in an unduly literal
application of the firm market area analysis to entire markets and
industries. Thus, in a prominent antitrust case involving bank
mergers, the argument for a distinct geographic area was that
little business done by the merging banks was nonlocal. The
argument was erroneous because much of the bank business conducted
within the hypothetical market area involved banks from other
areas. In sum, the LIFO argument overlooks exports; the LOFI
argument overlooks imports. Only by simultaneously satisfying both
the LIFO and LOFI criteria do we ascertain with a reasonable degree
of certainty that a given hypothetical market area is, in fact,
geographically distinct.
METHODS OF ESTIMATION OF GEOGRAPHIC
MARKET AREAS
While
theoretical work on geographic market areas is at least 50 years
old, empirical work on this topic is relatively recent.
Paradoxically, almost none of the empirical work has been done by
scholars specializing in regional/urban economics. These
specialists have tended to concentrate efforts on phenomena such as
optimal plant location, while specialists in transportation
economics tend to abstract from decisions involving plant location.
There are at least two explanations for this state of affairs.
One
explanation is the inherent difficulty of estimating market areas.
While few experiments in economics are controlled, the role of
judgment (extent of uncertainty) in interpreting results in this
particular area is well above average. Put another way, estimating
geographic market areas is comparatively less scientific than most
empirical work in economics. Another explanation is that, since the
scope of geographic market areas is strongly affected by existing
plant locations and transportation networks, study of the latter
phenomena assumes greater importance. Changes in plant location and
transportation networks tend to be very long-run in nature,
while the study of geographic market areas is primarily concerned
with short and intermediate length periods.
Therefore,
while it is true that changes in plant location, sizes of plants,
transportation networks, etc. may change the geographic scope of
markets, abstracting from these factors in the present context is a
minor