Download Scissors or Horizon and more Lecture notes Law in PDF only on Docsity!
SouthernEconomic Journal 1997, 64(2), 517-
Scissors or Horizon: Neoclassical Debates
about Returns to
Scale, Costs,
and
Long-Run Supply,
Nahid
Aslanbeigui*
and Michele I.
Naplest
Modem treatmentof
long-run (U-shaped)
cost curves
developed
from reactions to Sraffa's
criticisms of Marshall.He
argued
that internal (dis)economies
were
incompatible
with
partial-
equilibrium analysis
under
perfect competition. Pigou
concurred and drew
L-shaped
cost
curves;
Viner realizedthat this made firm size indeterminateand
industryoutput
volatile.
Using
Austin and Joan Robinson's
analyses, Stigler justified rising costs/supply, determinacy,
and
stability by
irrational
entrepreneursenduring
coordinationfailure and
by
factor
price changes.
We conclude that
consistency requires
constantcosts but firm
employment,output,
and factor
incomes remain
theoretically
indeterminate.It becomes
likely
that
large
firms will undermine
perfect competition.
1. Introduction
The short-run
U-shaped
cost curves of economics textbooks have been the
subject
of a
growing
literature, suggesting
that firm costs are constant over a wide
range
before
they
rise
(Larson 1991;
Maxwell 1965;
Yordon 1992). Using Machlup's
distinction and
building
on Beattie's framework
(1988),
Larson (1991)
has shown that,
if factors are indivisible in
acquisition
but divisible in use,
firms will choose fixed
input proportions
at
output
levels below
full-capacity
utilization. This
implies
constant and
equal marginal
and
average
factor
produc-
tivity
as well as constant and
equal marginal
and
average
variable costs. With the
majority
of
the real-world cases
falling
in this
category,
Larson has
suggested
that short-run constant costs
require "reworking
of
many aspects
of traditional economic
theory" (p. 473).
In this
paper,
we focus on a related
aspect
of textbook
analysis,
the
prevailing
view that
long-run
cost curves are
U-shaped
as well. We document the
challenge posed by
Sraffa's
(1926)
seminal article,
which showed the
incompatibility
of downward- or
upward-sloping
cost curves
with
partial-equilibrium analysis
under
perfect competition.
We discuss and evaluate the
ensuing
debate,
which lasted for almost two decades and led to the
emergence
of
long-run U-shaped
costs/rising supply
curves as dominant. We show that the interwar debate was resolved at the
expense
of
consistency,
since
partial-equilibrium analysis
is
logically
coherent
only
for the
Department
of Economics
and Finance,
Monmouth
University,
West
Long
Branch,
NJ 07764, USA;
E-mail
[email protected].
t Economics,
The
College
of New
Jersey,
CN4700, Ewing,
NJ 08650-4700, USA;
E-mail
[email protected].
The authorswould like to thankJim Devine,
Frank
Thompson,
and the
anonymous
refereefor
insightful
comments
and Young
Back Choi and CristinaMarcuzzofor helpful
conversations.The usual disclaimer applies.
Received August
1996; acceptedApril
'On short-runcost curves,
see also Johnston (1958)
and Dean (1976).
2
The
empirical
evidence for
L-shaped/U-shaped
cost curves in the
long
run is
beyond
the
purview
of this
paper.
For
studies on this issue,
see Lee (1984),
Dean (1976),
and Johnston (1958, 1960).
518 Nahid Aslanbeigui
and Michele I.
Naples
constant-cost case. We suggest
that
consistency requires
a
reworking
of traditional
price theory
under constant costs,
as well as the theories of the functional distribution of income,
factor
employment,
and firm
output
in the
long
run.
The
paper begins
with
background
sections on Marshall's
(1920)
and
Pigou's (1912)
anal-
yses
of the laws of returns and Sraffa's
challenge
to
long-run
models of
perfect competition.
The next two sections show that Sraffa's essay triggered
two different reactions.3 Some
(Pigou
- argued
that
long-run
firm cost and
industry supply
curves were
L-shaped:
Costs
declined until they
reached a horizontal
range.
This
suggested
a return to the classical view that
price
was determined on the
supply
side of the
economy.
It also
implied
indeterminate firm
size. Others (Viner
E. A. G. Robinson 1931;
J. V. Robinson
- justified U-shaped
firm costs and/or upward-sloping industry supply
curves.
Stigler's (1942)
textbook
synthesis
secured the need for "both blades of the scissors,"
demand as well as
supply,
to determine
long-run price
and firm
size,
Sraffa's
arguments notwithstanding.
We then evaluate the
remaining
inconsistencies in Stigler's synthesis.
We conclude that
only
constant costs are
theoretically
consistent with perfect competition.
However,
constant costs make it
impossible
to determine
firm output, employment,
and factor incomes a priori
and make
perfect competition unlikely.
2. Costs and
Industry Supply
Before Sraffa
Marshall
According
to Marshall
(1920),
increased
production
in
agriculture yielded diminishing
returns since the supply
of land was limited;4 this led to an
upward-sloping supply
curve for
the industry.
In
manufacturing,
however,
increased scale of
production "improved organization,
which increase[d]
the
efficiency
of the work of labour and
capital" (p. 265), implying increasing
returns. Economies
of scale were internal
if
they depended
on "the resources of the individual
houses of business engaged
in it and the
efficiency
of their
management" (pp. 262-263).
These
were due to such factors as division of labor and improved
skill-task
fit, improved recognition
for the product
of one firm,
and
diminishing marketing
costs. Economies were external if
they
depended
on "the
general development
of the
industry" (p. 262). Examples
of such economies
included "the growth
of correlated branches of
industry
which
mutually
assist one another,
perhaps being
concentrated in the same localities,
but
anyhow availing
themselves of the modern
facilities for communication offered by
steam
transport, by
the
telegraph
and
by
the
printing-
press" (p. 264). Regardless
of their
origins,
economies of scale resulted in downward
sloping
supply
curves
(p.
Most of Marshall's
discussion was couched
in terms of cost and
supply
at the
industry
level. The firms within any industry
were
dynamic
in nature: Like the trees of a forest, they
were in different stages
of their
life-cycle. Any
discussion about individual
production
units
was therefore in terms of the
representative
firm,
which linked the static notion of
industry
to
the dynamic
and historical firm (Moss 1984, p.
The
representative
firm was a
production
3 In some quarters,
Sraffa's challenge
was met by
the development
of models of imperfect/monopolistic competition
(Chamberlin
1933; Robinson 1933).
This
has been
well documented elsewhere
and is
beyond
the
purview
of our
paper
(see Blaug
1978;
Marcuzzo 1994;
Sardoni 1994;
Shackle
1983).
4 Most
Cambridge
economists who were trained
by
Marshall or
Pigou
used the term
diminishing
returns to denote both
short-run and long-run phenomena.
520 Nahid Aslanbeigui
and Michele I. Naples
itself be at a point beyond
which it would
experience diminishing
returns (Pigou
p. 177).
This analysis
did not, however, specify
the
shape
of the
supply
curve for the individual
firm.
In the 1920s,
the Marshallian treatment of the laws of returns was criticized heavily
on
various grounds. Clapham (1922)
declared that the laws of returns were
empty
economic boxes,
representing
a
logical/theoretical apparatus
devoid of
empirical
content, precision,
and
clarity.
Others repeated Young's (1913)
concern with the
accuracy
of
Pigou's policy
conclusions
with
respect
to
decreasing
and
increasing
cost industries
(Knight
Robertson 1924).7 However,
everyone
seemed to
agree, explicitly
or
implicitly,
that cost curves were
U-shaped
and
supply
curves upward sloping.
3. Sraffa's Criticisms
Sraffa's (1926)
seminal article showed that
downward-sloping
or
upward-sloping industry
supply
curves violated the conditions
necessary
for a
competitive industry
studied
in a
partial-
equilibrium
context. Perfect
competition required
that the demand for and
the
supply
of a
commodity
be
independent
from each other and from the demand for and
supply
of
other
commodities. When conditions of
decreasing
or
increasing
costs were
present,
a
change
in the
output
of one
industry
introduced
precisely
such
interdependence.
First,
the Marshallian treatment assumed that
increasing
costs were due to the existence
of
a factor in limited
supply (e.g.,
land).
Increased
industry output implied
an increase in the
price
of this factor;
this then
implied
a rise in the cost of
production
in other industries that used
this
factor. If their
products
were
potential
substitutes,
the
change
in relative
prices
would affect
the
demand for the
product
of the
original industry (Sraffa
p.
Second, increasing
returns faced similar
problems.
Internal economies were
incompatible
with
perfect
competition
because
they
led to
monopoly.
A static framework
required assuming
away
economies
due to
general
economic
progress.
As a result,
one had to focus on economies
that were external
to the firm but internal to the
industry
(hereafter
referred to as external
economies).
Such
economies were rare in
reality, argued
Sraffa,
and thus so were
"supply
curves
showing decreasing
costs"
(Sraffa
p. 540).
The "cost of
production
of commodities
produced
competitively
... must be
regarded
as constant in
respect
of small variations in the
quantity produced"
(p. 541).
Sraffa's
article
sparked
a debate about the
shape
of cost and
supply
curves.
Although by
the
early
1930s, perfect competition
was a familiar term and its
prerequisites
were well
specified,
no such consensus
was
yet
reached over the
shape
of the cost and
supply
curves. The
following
two sections will discuss
these debates about
L-shaped
and
U-shaped
cost curves and their
implication
for the slope
of the
industry supply
curve.
Long-Run
Cost and
Supply
Curves
L-Shaped
Cost Curves
Pigou's
Reaction. Six months after the
publication
of Sraffa's article, Pigou accepted
his
argument
that
diminishing
returns
required general-equilibrium
analysis
and were
incompatible
7 For a comprehensivesurvey
of this controversy,
see Aslanbeigui(1996).
Scissors or Horizon 521
with perfect competition
under
partial equilibrium.
In two
articles, Pigou (1927, 1928) spelled
out the conditions for the existence of equilibrium
in that framework. Events
exogenous
to the
operation
of an
industry,
for
example, general technological changes,
inventions,
or
changes
in
tastes,
were assumed
away.
In addition,
the
scope
of the
analysis
was
necessarily
and
logically
confined to "commodities which individually employ
so small a
proportion
of each of the
several factors of production
that no
practicable changes
in the scale of their
output
could
sensibly
affect the relative values of these factors"
(Pigou
p.
With no limitation in
the supply
of factors of
production, including managerial ability, Pigou
concluded that
... it is impossible
for productionanywhere
to take place
underconditionsof increasing
costs. In this matter
my
conclusion agrees
with that reached by
Professor Sraffa in his recentarticle. (p. 193; emphasis
added)
However, Pigou
did not rule out the existence of external economies. A
change
in the
industry's
scale of
production
could lead to the use of more
productive techniques
and "in-
creased specialization among
the makers of the machines used in the
industry" (Pigou
p.
195).9 Therefore,
the
industry supply
curve could
slope
downward and/or be horizontal.
Where average
cost for the
industry
declined (i.e., marginal
cost was below the
average cost),
the supply
curve would be the same as the
average-cost
curve since no
industry
could survive
at
prices
that did not cover
average
costs.'
Where
marginal
costs were constant,
the
average-
and
marginal-cost
curves would be horizontal and coincide. In other words,
the
general
form
for the
industry supply
curve was a
rectangular hyperbola (p. 197).
In 1928, Pigou
turned his focus to the firm's cost curves. He concurred with Marshall that
the firms in an
industry
were at different
stages
of their
life-cycle.
He, too, accepted
the need
to
study
the cost and
supply
conditions for a firm
through
the medium of a
typical
unit,
which
he chose to call the
equilibrium
firm, implying
... that there can exist some one firm, which,
whenever the
industry
as a whole is in
equilibrium,
in the
sense that it is
producing
a
regularoutput...
in
response
to a normal
supply price
...
will itself also indi-
vidually
be in
equilibrium....
The conditions of the
industry
are
compatible
with the existence of such a
firm;
and the
implications
about these conditions,which,
whetherit in fact exists or not,
would hold
good
if it did exist,
must be valid.
(Pigou
1928, pp. 239-240)
Unlike Marshall's
representative
firm, however, Pigou's equilibrium
firm did not
enjoy
internal
economies;
it reached an
optimum
size, "trespass beyond
which
yields
no further internal econ-
omies"
(Pigou
p.
Its costs therefore declined for the initial
ranges
of
output
and
became constant later.
Pigou
was one of the first economists to
publish U-shaped
cost curves
for the firm; however,
it was
implicit
in his
diagrammatic analysis
and discussion that the firm
would not move
beyond
the lowest
point
of its
average
cost curve. Stable
equilibrium
for this
8
Pigou's acquiescence
had been in the
making
for
many years.
In 1913,
he had
recognized
the
importance
of interde-
pendence
between individual
and market
supply (demand)
curves (Pigou
1913, p. 21).
In 1920, he had made explicit
that,
under partial-equilibrium
conditions,each industry
must be small enough
so that an increasein its output
did not
appreciably
affect the price
of factors of production(Pigou
1920, p. 935).
In 1924, in a response
to Allyn Young,
he
had agreed
thatthe
increasein
long-run
costs due to
diminishing
returnscould not be caused
by
technicaldiseconomies.
Such diseconomies
as
rising
factor
prices
were
only pecuniary,involving
income redistributionbut not increased
input
requirements(Pigou
1924, p.
194).
Once Sraffa's article
pulled
all the above
argumentstogether,Pigou agreed
that
increasing
costs and, by implication,upward-slopingsupply
curves were
incompatible
with the Marshallian
partial-
equilibrium,perfect-competition
model.
9
Sraffa had objected
that such economies were not very
common empirically.
For Pigou,
an analyticalstudy merely
exposed
the possibility
of such economies (Pigou
1927, p. 196).
' Pigou (1912, p. 174)
defined
total cost
to include
interestor the
opportunity
cost of
capital.
Scissors or Horizon 523
reaction toward
underproduction.
Actual
long-run
price
and output
would be unstable,
but would oscillate
above and below stable
points
of
equilibriumprice
and
equilibriumoutput.(p.
Viner's discussion is
vague.
He
appears
to
say
that the
industry supply
curve cannot be
defined,
in
light
of firm behavior. However,
the
quote
above addresses the
stability
of market
equilibrium,
not the
shape
of the
supply
curve; industry supply
would be horizontal and there-
fore infinitely
elastic, given
uniform and horizontal
long-run average
cost curves for all firms.
The problem
that
produces
the
instability
is that
perfectly
elastic
industry
and firm
supply
cause
an indeterminate division of industry output among
firms. Therefore,
firm
output
will be
subject
to swings
that economists cannot
accurately predict
and market forces will not attenuate. The
precision
of the
price-quantity
solution
implied by
a horizontal
industry supply
curve and
downward-sloping
demand curve belies the market
volatility
and
indeterminacy
at the firm
level,
of which Viner was acutely
conscious.
U-Shaped
Costs and
Upward-Sloping Industry Supply
Viner's Pecuniary
Diseconomies. Viner
sought
to
remedy
this
indeterminacy by
incor-
porating
externalities, appealing
to external economies and diseconomies of
production
as the
only possible
sources of
increasing
or
decreasing
costs for the
perfectly competitive industry.
External economies and diseconomies were mainly pecuniary14-due
to
changes
in
"prices
of
services and materials"
(Viner
p. 218), although initially
Viner had identified "techno-
logical
conditions" as
determining
the
shape
of the
long-run supply
curve
(p. 206).
In fact,
he
found pecuniary
external diseconomies of scale to be of
"indisputable practical importance"
(p. 220).
For Viner,
external economies and diseconomies of scale shifted a
plant's (short-run)
av-
erage-
and
marginal-cost
curves in a manner that
kept
its
output
constant.16 The increase in the
output
of the
industry
would have to take
place through
the
entry
of new
plants (Viner
p. 217)
or
producers (p. 221).
Viner did not
explicitly identify
firm
supply. Industry supply,
however,
could be
upward-
or
downward-sloping, coinciding
with the
industry's average-cost
curve. Unlike the modern
portrayal
of Viner's cost
curves,
the firm would not have a
U-shaped
long-run average-cost curve,
but the
industry
could.
Viner
presumed
that the
incorporation
of external diseconomies had remedied the indeter-
minacy
of firm
supply.
This, however,
is not
necessarily
the case. While
plants
had
U-shaped
cost curves,
firms with
reproducible
plants
had constant internal returns, suggesting
indeter-
minate size. As Viner himself observed,
"the individual
producers
will not concern themselves
13
Viner's concern
with the
problem
of
instability
seems
to be derived
from his static notion of
the
industry
and the
associated assumption
of identicalfirms.
14
Viner had difficultylocating examples
for technological
externaleconomies and even more so for diseconomies (Viner
1931, pp. 217-218).
One possible example
of technological
external diseconomies "might
be higher
unit highway
transportation
costs when an industry
which provides
its own transportation
for materialsand productsexpands
its
output
and therebybrings
about traffic congestion
on the roads" (p. 221).
'5 Viner (1931, p. 220) suggested
that industryexpansion
caused "increased purchases
of
primary
factorsand materials
which
... must tend
to raise
their unit
prices."
This referenceto nonmanufactured
inputs implicitly suggested
dimin-
ishing
returnsto land as the source of diseconomies. However,
he
provided
no concrete
examples
or
supporting
evi-
dence.
16 Viner did not
explain why
the
output
of the
plant
or the individualconcern (he
used the two terms
interchangeably
[Viner 1931, pp.
218-219])
would remain constant.If the
price
of an
input changed relatively,
factor ratios in each
plant
would be adjusted.
This would modify
the firm's blueprints
and could affect optimal plant output.
For Pigou,
underthe same circumstances,the firm'ssize could easily change.
524 Nahid
Aslanbeigui
and Michele I.
Naples
with the effect on the costs of other
producers
of their own withdrawalfrom or entranceinto
the
industry"(Viner
p. 222).
Therefore,
firmswould decide whetherto
expand
or contract
based on internalcosts and
returns,
not full informationabout
general-equilibriuminput-price
changes.
Firms would face a
prisoner's
dilemma:A
good output
choice for one firm would turn
out to be
unprofitable
if all firms
simultaneously
made the same
output
choice.
Volatility
would
result. Not
only
would
expansion
mean excess
supply
for the
industry,
but
input prices
would
also have risen due to external diseconomies. Both
phenomenaprovide
incentives to contract
output.
Note
that,
since firms in the
aggregate
would
probably
overcontractand the
industry
experience
excess
demand,
this would raise
output prices,
while
input price
was
falling.
Vola-
tility
would be a common state of
affairs,
despite
external diseconomies and
upward-sloping
industry supply.
Austin Robinson's Coordination Failure. In the same
year
that Viner
published
his article,
Austin Robinsonhad discovered a
way
to reduce the theoretical
possibility
of unstable
equilib-
rium
through
internaldiseconomies for the firm
(see
Robinson
Robinson discussed the
supply
curve for the
optimum (representative)
firm,
the
optimum
size
being
determined
by
the
interplay
of five different forces:
technical,
managerial,
financial,
marketing,
and risk and
fluctuation
(p.
The
optimum
firm could
experience
technicaleconomies due to increaseddivision of labor
and
dexterity
as well as to
purchases
of
larger
and more efficient machines.However,Robinson
agreed
with
Pigou
and Viner
that,
beyond
some
point,
the firm would exhausttechnicalinternal
economies and would never
experience
technicalinternaldiseconomies. "If otherconsiderations
require
a scale
larger
than the technical
optimum,
the technical scale of
production
can be
increased
by
mere
multiplication"
(Robinson 1931,
p.
Robinson
believed, however,
that other diseconomies could
emerge.
Economies of scale in
administrationwould be exhausted at some level of
output (Robinson
p. 43).
This co-
ordinationfailure would increase costs for the
firm;
there is such a
thing
as too
big
a firm
(p.
How big
a
firm can successfully grow
will
dependupon
how it solves this problem
of the co-ordinating
of
separateddepartments
and
separatedspecialists.(p. 45)
In this the
managerialoptimum
differs from the technical
optimum.
The latteraffords
only
a minimumscale
below which the
greaterefficiency
cannotbe achieved.Additional
outputmay
be
produced
underconditions
of
approximately
constantcost. But if the
managerialoptimum
is exceeded, costs, throughdeclining
effi-
ciency
and the need for additional co-ordination, begin
to rise. The
managerialoptimum
sets, therefore,
not
only
a lower but also an
upper
limit to the scale of
operations.(p.
Financialfactors could
equally
lead to economies and diseconomies of scale.
Large
firms
could borrowat more
advantageous
interestratesfrom banks and
by issuing
stock to the
general
public.
These economies would
likely,
though
not
necessarily,
come to a halt at some
point,
however,
since the leaders of
larger
firms
would have
to face
the
"cramping
influence
upon
their
enterprise"
from the shareholdersand
others;
"the
autocracy
which is the foundationof
their
efficiency"
would be "confinedand limited"
(Robinson
pp. 59-60).
Marketing
factors related to scale economies included the cost of
buying
(raw material,
etc.)
and the cost of
selling
the
product.Larger
firmscould
buy
at discounted
prices,
hire
expert
17
The optimumfirm,according
to Robinson (1931, pp. 14-15), represented
"the scale of production
which, havingregard
to the circumstancesof the
industry,
was looked
upon
as the best scale of
production
sometimein the recent
past";
it
was the firm with the lowest
long-runaverage
cost.
526 Nahid Aslanbeigui
and Michele I.
Naples
manufacturing processes
were
typically
not
cranky
or
idiosyncratic
in
finding
substitutes for
inputs
in short
supply.
Joan
Robinson's
analysis
of the
elasticity
of substitution as well as
supply
can be
interpreted
as
supporting
Sraffa's views on constant costs. Sraffa had said that if
supply elasticity
for the
input
were
infinite for all factors,
which is true
by assumption
in
partial equilibrium,
constant
returns
would be
implied.
Yet Viner and
Stigler
focused on the first
part
of her
paper
and
rising
supply price
at full
employment, given imperfect supply
elasticities.
Stigler's
Textbook Treatment
By
the
early
1940s, U-shaped
cost and
upward-sloping supply
curves were
part
of the
orthodox economist's tool box.
Stigler's
textbook, The Theory of Competitive Price,
which came to be the basis of microeconomics education for the next several decades (Yordon
1992), synthesized
the theories of
perfect competition emerging
from this literature.
Stigler
introduced the firm first and foremost. He
separated
the short-run
U-shaped
cost
curves from cost curves in the
long
run,
the latter
being
an
envelope
of the
many possible
short-run cost curves. Unlike the short run,
the
long
run was defined as a
period
in which the
size of
the
plant
could
vary.
Like
many
before him, Stigler recognized
that,
under
perfect competition,
internal econ-
omies
would be exhausted21 and the firm would reach a
point
of constant returns to scale.
However, beyond
some size,
the firm would face internal diseconomies of scale:
... the growth
of a firm puts
heavierand heavierburdens
on the
management.Quite
aside from the difficult
problem
of
expansion
itself, large groups
are much harder
to coordinatethan smallerunits. For
management,
and controlin general,inherently
face a problem:
the final authority
to make decisions
cannotbe subdivided
or
delegated.
Large
units are,
in fact,
confronted by
a dilemma.At one extremeall authoritymay
be
delegated.
Then there will be no
unity
of
policy
or uniformity
of performance.
At the otherextreme,
all decisions
may
be made
by
a final center.This system
involves bureaucracy
in its worst form: "red tape," hopeless delay,
decisions based on dilutedmemoranda.Between these two extremesthe large
firm attempts
to steera middle
course,
but it never achieves that compactness,flexibility,
and singleness
of purpose
which are possessed by
every well-managed
medium-sizedfirm. The growing difficulty
of coordinationand decision-making
even-
tually stops
the
growth
of every
firm. (Stigler 1942, p. 138)
As is evident from the above
quotation, Stigler replicated
Austin Robinson's
arguments-
without attribution-with a minor twist. In his hands,
coordination failure translated into a fixed
factor of
production,
the final
authority
or
entrepreneur. Stigler
did not address the
possibilities
that Robinson had identified
whereby competitive
firms could avoid these sources of
rising
costs
(e.g., decentralization). Stigler
assumed that "diseconomies of
large-scale production
set in soon
enough
to insure numerous firms and therefore
competition" (Stigler
p.
implying
that the
long-run envelope
of short-run
average
costs is also
U-shaped,
albeit flatter.
The firm would
operate
at the minimum
point
on this curve or be
competed
out of business.
Stigler
concluded that all firms must have identical minimum costs in the
long
run, although
their sizes and
technologies might
differ;
firms with
some
advantage
due to a
superior
resource
20
Samuelson's text,
Economics: An
IntroductoryAnalysis,
first
published
in 1948, suggested
that perfect competition
was rare, existing only
in a few lines of
agriculture
(Samuelson1948, pp.
491, 509).
Since he focused on monopolistic
or
imperfectcompetition,
we do not include his text in our analysis.
21
Increasing
returnswere possible
due to indivisible machinery.
Otherforms
of
indivisibility
included
marketing,
finan-
cial,
research,and management
indivisibilities (Stigler
1942, pp.
136-138); Stigler
did not mentionthat three of these
four would have no
place
under
perfect competition.
If
increasing
returnswere
ongoing,
the
long-run average
cost
curve for the firm would be downward
sloping,
and
monopoly
or
oligopoly
would
replacecompetition.
Scissors or Horizon
would be forced by competition
to
pay
that resource a rent that would
bring
unit costs in line
with the industry's.
However, upward-sloping
firm
supply
need not mean
upward-sloping
in-
dustry supply. Stigler
realized what Viner had realized,
that
rising
costs due to coordination
failure need not mean rising industry
costs.
Industry expansion
could "take
place primarily
through
an increase in the number of firms"
(Stigler
p. 162).
Hence,
external
pecuniary
factors
proved key
to
generating
a
rising-cost industry. Stigler
(1942, p.
first
argued
that
inputs
were
heterogeneous (although
under
perfect competition
output
is
homogeneous [p. 161]),
and therefore the
expanding industry
attracts
productive
services
less suitable for the
product.
He then
reproduced
J. Robinson's
full-employment
ar-
gument
that factor
prices
would
tend to rise as
any industry expands (p.
Stigler's synthesis
of the debates on
perfect competition provided
a basis for
partial-equi-
librium
analysis
that
produced
determinate results for the firm and the
industry.
It rested on
two crucial
assumptions:
coordination failure
leading
to internal diseconomies and external
diseconomies
reflecting heterogeneous inputs
and
rising input prices.
Like
any synthesis, Stig-
ler's leaves out certain
insights
from the historical literature. In this case,
these
challenged
the
viability
of his
assumptions.
Stigler's Synthesis:
An Assessment
The Irrational
Entrepreneur
Internal diseconomies for the firm are
problematic.
Once a firm has achieved minimum-
cost conditions,
it is
illogical
to assume it does not have the know-how to
duplicate
them.
Austin Robinson's
explanation,
which was later seized
upon by Stigler
and others, begs
the
question:
Why
should coordination fail? Given
perfect foresight,
the
entrepreneur
can find other
ways
to
expand
besides
shooting
him/herself in the foot,
as Austin Robinson himself
empha-
sized. If
managing
a
larger
firm is
dysfunctional,
the
entrepreneur
has the alternative of
setting
up
smaller subsidiaries under a
holding company
or of
decentralizing (consider
the 1995 break-
up
of AT&T into three autonomous entities).
When Stigler
assumed the
entrepreneur puts
his/
her desire for control ahead of the drive for
profits,
he created an
inconsistency
in a model
premised
on
profit-maximization-determinate
firm size rests on irrational behavior
by
the en-
trepreneur.
22
The heterogeneity
of
inputs
is a corollary
of the rising supply price.
In perfect competition,
an industry'soutput
is
homogeneous.
To
say,
as
Stigler
did,
that
expansion requiresusing
inferior
inputs
means substituting
lower quality
inputs
for the preferredinput,
which is only
rational
if the price
of the preferredinput
has risen.
23
Stigler acknowledged
that there might
exist constant-costindustrieswhere the industry
was small relative to its sup-
plying
industriesand changes
in its size did not affect the price
of inputs.
In that case, "[t]he long-runsupply
curve
of the industry
... is of course a horizontalline" (Stigler 1942, p. 163).
He allowedfor the possibility
thatfactor prices
might
fall as the competitive industry expanded,
because it would buy inputs
from a monopoly
or from another
decreasing-costindustry,
but deemed this case rare.
Stigler
also addressedwhat he called "Sraffa's Case," where
industry
demand would not be well defined since
changes
in
prices
of
productive
services
potentially
affect
output
prices
in industrieswhose
products
are
complements
or substitutesfor this
industry'soutput.
He claimed that,
never-
theless,
"it is
always possible
to draw an
unambiguoussupply
curve for a
competitiveindustry"(p.
165).
24 As
Sargant
Florence observed,
"thereis little to prove
the universality
and
inevitability
of any
such law of increasing
costs in the long
runwhen manufacturershave time to get
new equipment
in orderto meet enlarged
ordersor anticipated
ordersand have time to reorganize
and delegate responsibilities"(quoted
in Johnston1960, pp. 23-24).
Scissors or Horizon 529
Those seeking
to construct the microfoundations of macroeconomic fluctuations
might
do well
to consider such a constant-cost model.
Finally,
since factor incomes and
employment
and firm
output
are indeterminate in
theory,
economists will have to use the inductive method and infer generalizations
from
empirical
research,
a method at odds with the
positivist
tradition.
Notably,
such induction dominates case
research in the business disciplines (accounting,
finance, production management, marketing),
which typically
treat constant
production
costs as the referent. Further research on the
theory
of the
firm under constant costs
might fruitfully
cross these
disciplinary
boundaries.
- Conclusion
The cost
controversy
led to the articulation of the
relationship
between in-
dustry supply
and firm costs under
perfect competition.
It shifted the
analytical
focus from the
industry
to the firm and led to the
development
of
U-shaped
cost curves. The Marshallian
attention to the
dynamic
nature of the firm and to historical costs was
supplanted by
a
concep-
tualization of the
long
run as
composed
of static,
identical firms. The mathematical and dia-
grammatic
analyses
of costs became much easier and the
theory
more
precise
and standardized.
However,
we have shown that the standard model is still logically
inconsistent.
To
be
consistent, partial-equilibrium
analysis
under
perfect competition requires
constant
costs. Ironically,
if firm size
is indeterminate but
large
size is known to facilitate
market
power,
constant-cost firms will
arguably
tend
to
grow beyond
the size consistent with
perfect compe-
tition.
Then the
only
cost curves consistent with the
partial-equilibrium analysis
of
perfect
competition
are
likely
to lead to its
opposite:
monopolistic competition
or
oligopoly.
References
Abouchar,
Alan. 1990. From Marshall's cost
analysis
to modern
orthodoxy: Throwing
out the
baby
and
keeping
the bath.
Economie
Appliquee
43:119-43.
Aslanbeigui,
Nahid. 1996. The cost
controversy: Pigouvian
economics in
disequilibrium. European
Journal
of History
of
Economic
Thought
3:275-95.
Beattie,
Bruce R. 1988.
Asymmetric stages, ridgelines
and the economic
region
for the two-variable-factor
production
function model. Southern Economic Journal 54:562-71.
Blaug,
Mark. 1978. Economic
theory
in
retrospect.
3rd edition.
Cambridge: Cambridge University
Press.
Chamberlin,
Edward H. 1933. The
theory
of monopolistic competition. Cambridge,
MA: Harvard
University
Press.
Clapham,
John H. 1922. Of
empty
economic boxes. Economic Journal 32:305-14.
Dean,
Joel. 1976. Statistical cost estimation.
Bloomington,
IN: Indiana
University
Press.
Johnston,
Jack. 1958. Statistical cost function: A
reappraisal.
Review of
Economics and Statistics 40:339-50.
Johnston,
Jack. 1960. Statistical cost analysis.
New York: McGraw-Hill.
Kaufman,
Bruce E. 1993. The
origins
and evolution
of
the
field of
industrialrelations in the United States. Ithaca,
NY:
ILR Press.
Knight,
Frank H. 1924
(reprinted
in 1952).
Of
empty
economic boxes: A
rejoinder.
In
Readings
in
price theory,
edited
by George
J.
Stigler
and Kenneth E.
Boulding. Chicago:
Irwin, pp.
160-79.
Larson,
Bruce. 1991. A dilemma in the theory
of short-run production
and cost. Southern Economic Journal 58:465-74.
Lee,
Fred. 1984. The marginalist controversy
and the demise of full cost pricing.
Journal of
Economic Issues 18:1107-32.
Machlup,
Fritz. 1952. The economics of
sellers' competition:
Model analysis of
sellers' conduct. Baltimore:Johns
Hopkins
Press.
Marcuzzo,
M. Cristina. 1994.
R. F Kahn and
imperfect competition. Cambridge
Journal
of
Economics 18:25-39.
Marshall,
Alfred. 1920
(reprinted
in 1949). Principles of
economics. 8th edition. London: Macmillan.
Maxwell,
W. David. 1965. Short-run returns to scale and the
production
of services. Southern Economic Journal 32:1-14.
530 Nahid
Aslanbeigui
and Michele I.
Naples
Moss, Scott. 1984.
The
history
of the
theory
of the firm from Marshallto Robinson and Chamberlin:The source of
positivism
in economics.
Economica 51:307-18.
Pigou,
ArthurC. 1912. Wealth
and
welfare.
London:Macmillan.
Pigou,
ArthurC. 1913. The interdependence
of different sources of demand and
supply
in a market.The Economics
Journal 23:19-24.
Pigou,
ArthurC. 1920. The economics of welfare.
London:
Macmillan.
Pigou,
ArthurC. 1924. The economics of welfare.
2nd edition. London:Macmillan.
Pigou,
ArthurC. 1927. The laws of diminishing
and
increasing
cost. Economic Journal 37:188-97.
Pigou,
ArthurC. 1928. An analysis
of supply.
Economic Journal
38:238-57.
Robertson,
Dennis H. 1924. Those empty
boxes. Economic Journal 34:16-30.
Robinson,
E. Austin G. 1931 (revised
in
1935).
The structure
of competitiveindustry.Cambridge:CambridgeUniversity
Press.
Robinson,
Joan V. 1933. The economics of imperfectcompetition.
London:Macmillan.
Robinson,
JoanV. 1941 (reprinted
in 1952). Rising supplyprice.
In
Readings
in
price theory,
edited
by George
J.
Stigler
and KennethE. Boulding. Chicago:
RichardD. Irwin, pp.
233-41.
Samuelson,
Paul. 1948. Economics: An introductoryanalysis.
New York:McGraw-Hill.
Sardoni,
Claudio. 1994. The generaltheory
and the critique
of decreasing
returns.Journal of
the
History of
Economic
Thought
16:61-85.
Shackle, George
L. S. 1983. The
years
of high theory. Cambridge:CambridgeUniversity
Press.
Sraffa,
Piero. 1926. The laws of returnsunder
competitive
conditions.Economic Journal 36:535-50.
Stigler,George
J. 1942. The theory of competitiveprice.
New York:Macmillan.
Viner,
Jacob. 1931
(reprinted
in 1952).
Cost curves and
supply
curves. In Readings
in price theory,
edited by George
J.
Stigler
and KennethE.
Boulding. Chicago:
RichardD. Irwin, pp.
198-226.
Viner,
Jacob. 1950
(reprinted
in 1952). Supplementary
note (1950).
In Readings
in price theory,
edited by George
J.
Stigler
and KennethE.
Boulding. Chicago:
RichardD. Irwin, pp.
227-32.
Yordon, Wesley
J. 1992.
Stigler's adaptable
and indivisible plant
and the micro/macroschism. History of
Political
Economy
24:455-70.
Young, Allyn.
Pigou's
wealth and welfare. Quarterly
Journal of
Economics 27:672-86.
