A brief prehistory of the

theory of the firm

Paul Walker

Abstract

The mainstream theory of the firm didnt exist until around 1970.
Before then what we had was the prehistory of the theory of the
firm. For more than two thousand years tools were available that
could have given rise to a theory of the firm or, at least, a theory
of micro level production, but none appeared. During this time the
best that occurred were discussions of macro level or aggregate pro-
duction. Given the long empirical history of and the importance to
the economy of firms one may assume that economists have long been
developing a detailed and sophisticated theoretical understanding of
the firm but it turns out this is not the case. Up until the 1970s
the development of the theory of the firm was a story of neglect and
disinterest.

First Preliminary Draft: comments welcome.

28th March 2017

Thisessay draws material from chapter 2 of Walker (2016).

psw1937@gmail.com. We wish to thank (without implicating) Glenn Boyle and Gavin
Kennedy for comments on previous drafts of the paper.
2
The mainstream theory of the firm didnt exist, in any serious manner, until
around 1970. It was only then that the current theory of the firm literature1
began to emerge based largely upon the work of Ronald Coase - Coase (1937) -
and to a lesser degree Frank Knight - Knight (1921b). It was work by Oliver Wil-
liamson (see, for example, Williamson 1971, 1973, 1975), Alchian and Demsetz
(1972) and Jensen and Meckling (1976), among others, that drove the upswing
in interest in the firm among mainstream economists. Before then there was
no great interest shown in the firm as a significant economic institution by any
school of economic thought. Before this time what we had was in effect the
prehistory of the theory of the firm. For more than two thousand years tools
were available that could have given rise to a theory of the firm or, at least, a
theory of micro level production but none appeared. During this time the best
that occurred were discussions of macro level or aggregate production.

1 Background2
The firm3 is an ancient and important empirical feature of the economic land-
scape. Firms, of some description, have existed around the world for several
thousand years.4
If we look, for example, at ancient India we see a sophisticated form of
business organisation called the Sreni. The Sreni was a complex organisational
entity that shared similarities with companies, guilds, and producers cooper-
atives. According to Khanna (2005) the Sreni was being used as early as 800
BC and was in more or less continuous use from that time until 1000 A.D., at
which time an Islamic invasion of India started. Khanna (2005) explains that
the Sreni were separate legal entities which could hold property separately from
their owners, create their own regulations controlling the behaviour of their
members, contract, sue and be sued in their own name (Khanna 2005: 8-9).
Table 2.1 gives a more detailed summary of characteristics of the Sreni. No-
tice that the Sreni shares a number of these characteristics with the modern
business company. Sreni were utilised in occupations involving workers such as
carpenters, ivory workers, bamboo workers, money-lenders, barbers, jewellers
and weavers (Khanna 2005: 10).
In other regions of the world firms go back ever further. Since at the be-
ginning of the second millennium BC firms were involved in the long-distance
1 For a short overviews of the modern theory of the firm see Walker (2015, 2016: chapter
4).
2 Material for this and some of the following sections is taken from chapter 2 of Walker

(2016).
3 Spulber (2008: 5, footnote 8) gives the origin of the word firm as [t]he word firm

derives from the Latin word firmare referring to a signature that confirmed an agreement by
designating the name of the business.
4 The term firm is applied loosely here. Many of the organisations being called firms would

not meet the requirements of some definitions of a firm. For example many would not be firms
under the definition of Spulber (2009). For Spulber the firm is defined to be a transaction
institution whose objectives differ from those of its owners (Spulber 2009: 63). Spulber
writes Farmers, artisans, and merchants from the earliest times to the eighteenth century are
precursors to the contemporary firms. What distinguishes these economic actors from firms
in that their enterprises tended to be integrated with the personal economic affairs of the
entrepreneur. There was no separation between the owners commercial activities and their
personal consumption activities (Spulber 2009: 103). For a critical discussion of Spulbers
approach to the firm see Hart (2011).

1
trade between the city-state of Assur and Anatolia. Documents from around
1850 BC detail profit-oriented trade in commodities such as textiles and metals
(Jursa 2014: 27). Assyrian merchants traded expensive woolen textiles, tin and
lapis lazuli for silver and gold which they shipped back to Assur.5

Characteristics Present in Ancient Indian Sreni?

Separate Entity Yes
Centralized Management Yes
Transferability of Interest Probably Yes
Limited Liability Probably Not
Agent has power to bind entity? Yes
Management elected? Yes (though at times appears hereditary)
Can management be removed? Yes
Duty of Loyalty Probably Yes
Duty of Care Yes
Liability insulation Yes (though apparently not very detailed)
Screens on shareholder suits and internal Yes (though apparently not very detailed)
cement activity
Internal rules have binding effect Yes
Some reimbursement for legal defense Yes
Formation is easy Yes
Register with state Yes
State approval needed Yes
Use of incentive payments Yes (though apparently not very detailed)
Entry is easy Some conditions, but no caste bars.
Sharing of assets and liabilities Terms of agreement and additional rules
Exit is easy Yes, but with obligations potentially
Board/Committee Independence Probably Yes
Other board qualifications Yes (though apparently not very detailed)
Voting Regulation Yes (though apparently not very detailed)
Open debate in meetings & shareholder Yes, with some limits (though apparently
resolutions not very detailed)
Transparency is valuable and disclosure is Probably Yes (though apparently not very
encouraged detailed)

Table 2.1. (Khanna 2005: Table 1, p. 27; table footnotes removed.)

The institutions of trading have been well documented.

Most of these traders had become more independent by having be-
come managers of a joint-stock fund (called naruqqum, money
bag), usually set up in Assur. This phenomenon appeared for the
first time around 1900 BC and seems to have been an Old Assyrian
invention that went beyond individual partnerships and cooperation
in a joint caravan. The arrangement, rather similar to that of the
early medieval compagnia, meant enlisting a number (usually about
a dozen) of investors (ummi anum, financiers), who supplied cap-
ital rated in gold, usually in all ca. 30 kilos, ideally consisting of
shares of 1 or 2 kilos of gold each. It was entrusted to a trader (the
tractator ), usually for ca. ten years, for the generally formulated
purpose of carrying out trade. The contract contained stipulations
5 For more on the Old Assyrian trade see Veenhof (2010) and Larsen (2015). For a dis-

cussion of business companies in the later, first millennium BC, Babylonian period see Jursa
(2010).

2
 on a final settlement of accounts, on paying dividends, on the divi-
sion of the expected profit, and on fines for premature withdrawal of
capital (meant to secure the duration of the business). Investors or
shareholders mostly lived in Assur, but successful traders in Anato-
lia too invested in funds managed by others, perhaps also as a way
of sharing commercial risks. In such cases a contract would to be
drawn up in Anatolia that obliged the tractator to book in Assur x
gold in his joint-stock fund in the investors name. Among the in-
vestors we find members of the tractators family, but also business
relations and others, probably a kind of merchant-bankers, and
other rich citizens, who aimed at fairly safe, long-term investments
(Veenhof 2010: 55).
Silver (1995: 50) notes that,
[p]rivate firms (bt
atu) were prominent in late-third-millennium
Akkad (the region south of Baghdad), in the Old Assyrian trade
with Cappadocia [ . . . ] and, somewhat later, at Nippur. In the
mid-second millennium the firm of Tehip-tilla played a major role in
the real estate transactions and other business activities at Nuzi. A
list of about the some time from Alalakh in northwest Syria refers
to sixty-four firms participating in leatherworking, jewelry, and car-
pentry.
If we examine the economy of ancient Greece we find that there were some
relatively large firms but they were few in number (Bresson 2014: 45). Most
of the commercial operations that did exist were small and of limited duration.
The (in theory) infinitely lived firm did not exist, partners would agree to co-
operate for just a single business operation. There may be many investors or
several active partners, but their cooperation lasted for only one voyage or one
operation. The development of permanent firms for commerce was unnecessary
since low transaction costs meant that market transactions were sufficient for
business operations.6 Alain Bresson argues,
According to Coases famous definition, a capitalist firm is first
defined as an alternative to coordinating production and distribu-
tion through external markets (Coase 1937). But as a matter of
paradox and by contrast to the medieval world, resorting to market
in antiquity was so easy that it did not seem necessary to build per-
manent firms proprio sensu. Investors could contract in a series of
different business operations, thus both minimizing risks and max-
imizing their profits by making for themselves the best choices of
investment (Bresson 2014: 57-8).
With regard to the size of firms in ancient Greece Bresson (2014: 45) writes
[ . . . ] large [handicraft] workshops, with possibly a few dozen work-
ers (sometimes up to 120 as in the case of the metic Kephalos in
Athens in the fourth century, as mentioned by Lysias 12.19 [Todd
2000]) could indeed exist, but they were rare.
6 For more detail on production in ancient Greece see Bresson (2016) and Davies (2007).

3
 If we turn to the rural economy in ancient Greece we see that when compared
to their urban counterparts, farms were longer lived, hieratical organisations,
often family owned and operated with a slave workforce (Bresson 2014: 57-9).
The use of private firms for the provision of many economic goods and ser-
vices during the Roman Republic has been noted by Sobel (1999: 21),
[t]he republican Senate left virtually all economic activities to priva-
te individuals and companies, known collectively as the publicani.
Tax collection, supplying the army, providing for religious sacrifices
and ceremonies, building construction and repair, mining, and so on
were all contracted out. There was even a contract for summoning
the assembly in session and one for feeding the sacred geese.
Micklethwait and Wooldridge (2003: 4) also note that private companies were
formed for the purpose of collecting taxes, and other commercial activities,7
[t]he societates of Rome, particularly those organized by tax farm-
ing publicani, were slightly more ambitious affairs. To begin with,
tax collecting was entrusted to individual Roman knights; but as the
empire grew, the levies became more than any one noble could guar-
antee, and by the Second Punic War (218-202 b.c.), they began to
form companies societates in which each partner had a share.
These firms also found a role as the commercial arm of conquest,
grinding out shields and swords for the legions. Lower down the so-
cial scale, craftsmen and merchants gathered together to form guilds
(collegia or corpora) that elected their own managers and were sup-
posed to be licensed.
In some cases these ancient firms grew to be of reasonable size. Silver (1995:
66-7) notes,
[w]e may note here that during the Ur III period a new mill at
Girsu required the services of 679 women and 86 men (Maekawa
1980: 98)
and [a] number of cities possessed large workshops employing hundreds
of women in spinning and weaving. For example, a late-third-millen-
nium text from Eshnunna lists 585 female and 105 male employees
in a weaving house (Silver 1995: 143).
Ancient firms also diversified their activities.
Large commercial houses flourished in Babylonia from the seventh
to the fourth century. The House of Egibi, for example, bought and
sold houses, fields, and slaves, took part in domestic and interna-
tional trade, and participated in a wide variety of banking activit-
ies. [ . . . ] Earlier, in the late third-millennium Sumer, the rulers
and governors controlled vertically integrated firms that used wool
of the sheep they raised in their weaving workshops. At the same
time, an Umma businessman (- bureaucrat?) named Ur-e-e busied
7 For a brief discussion of the forms that firms could take in ancient Rome see Hansmann,

Kraakman and Squire (2006: 1356-64) and Fleckner (2015).

4
 himself with manifold operations, including raising livestock; trans-
actions involving cheese, oil, leather, carcasses, wool; the weaving
and finishing of cloth; shipments by boat of fish and grain; and even
the construction of boats (Silver 1995: 67).
In medieval Europe firms were typically small with households being the
normal production unit in agriculture. In cities, the evidence, such as it is,
suggests that workshops were in the main small family based operations with
few employees, apprentices aside, supplying the local market. Such producers
carried out a small number of tasks and bought intermediate goods on the
market. Larger scale production took advantage of the division of labour but
the different stages of production were not integrated into a single workshop or
firm. Economies of scale were exploited, not by creating large scale vertically
integrated firms, but by employing the putting out (or Verlag) system of
production (Persson 2014: 242-3). Under this arrangement merchants would
supply raw material inputs to artisan families or family members in urban and
rural households. These producers were paid per unit of output to help counter
moral hazard problems. Economies of scale were exploited in terms of marketing
the produce and in the purchase of inputs but production was not centralised
as in a modern firm (Persson 2015: 99).
In the modern world the scope of the firm has expanded enormously. In
the contemporary economy the majority of economic activity takes place not
across markets but within firms. One estimate, due to McMillan (2002: 168-
9), suggests that less than a third of all the transactions in the US economy
occur through markets, and instead over 70 percent are made within firms while
Lafontaine and Slade (2007: 629) assert that the [d]ata on value added, for
example, reveal that, in the United States, transactions that occur in firms are
roughly equal in value to those that occur in markets. In terms of the number of
firms and their contribution to employment Otteson (2014: 30), taking the U.S.
as an example, reports that: [i]n 2008, the United States had some 31.6 million
businesses across thousands of industries employing some 120 million people.
The range of sizes of contemporary firms goes from the sole proprietorship to
the global giant:
[...] in 2011, Royal Dutch Shell operated in over 80 countries, had
annual revenue exceeding the GDP of 150 nations, and paid its CEO
35 times more than the president of the United States. In the same
year, the total number of employees at Wal-Mart exceeded the pop-
ulation of all but 4 US cities. In addition to such giants, tens of mil-
lions of smaller firms operate around the world (Kikuchi, Nishimura
and Stachurski 2012: 2).
Given the long empirical history of and the importance to the economy of
firms one may assume that economists have long been developing a detailed
and sophisticated theoretical understanding of the firm but it turns out this
is not the case. As will become evident in what follows up until the 1970s the
development of the theory of the firm has been a story of neglect and disinterest.

5
2 Development of a theory of production or the
firm
To begin our survey of the development of the theory of production and/or the
theory of the firm we briefly look at the history of thought on the division of
labour. As has been demonstrated by work in the twentieth century the division
of labour can act as a catalyst for a theory of the firm but it took more than
two thousand years for such theory to evolve. Until Alfred Marshall at the end
of the nineteenth century many authors, including Adam Smith, wrote on the
division of labour without applying it to the theory of micro-level production
or the firm. Following on from this discussion we will consider other approaches
to production and the firm, noting that before the neoclassical economists no
group of writers developed a theory of micro-level production and only Alfred
Marshall wrote explicitly on the theory of the firm.

2.1 Division of labour8

The division of labor is not a quaint practice of eighteenth-century pin
factories; it is a fundamental principle of economic organization
Stigler (1951: 193).

As suggested by Stigler the division of labour is integral to the theory of

economic organisation, including the theory of the firm. But while the division
of labour is a very old idea in economics it didnt spark discussion about the
firm or about production until relatively recently.
Discussions of the division of labour go back at least as far as the ancient
Greek philosophers such as Democritus, Xenophon, Plato and Aristotle. Start-
ing at about the same time as Plato and Aristotle Chinese philosophers such
as Kuan Chung, Lao Tzu, Confucius and Mencius were also investigating the
division of labour. During the medieval period both Islamic and Christian theo-
logians and philosophers considered its consequences. Works by al-Ghazali,
Nasir al-Din Tusi, Thomas Aquinas and Ibn Khald un all treated aspects of the
division of labour. The pre-classical, classical and neo-classical economists con-
tinued and expanded the enquiry, while in the 20th century theories of the firm
based around the division of labour finally appeared.

2.1.1 Ancient philosophers

The historian of economic thought James Bonar makes the point that for Plato
the division of labour drives the social organisation of production. Bonar writes,
Platos conception of Production is in close connection with this
view of Wealth. It is important not that men should have as many
wants as possible, and satisfy them all, but that they should find out
what their special work is in the world and do it. He illustrates this
doctrine in various passages of the Republic, and especially in the
clearest of his economic analyses, the account of Division of Labour
8 For a more complete analysis of aspects of the history of the division of labour see, for

example, Bonar (1893: book 1), Groenewegen (2008), Hosseini (1998, 2003: 36-7), Lowry
(1987: 68-73), Robbins (1998: 12-14) and Sun (2008, 2012, 2016).

6
 in the Second Book. A State, he there says, is formed because the
individual is not able to supply all his wants by himself, but only
when he makes common cause with other men, and devotes himself
to one single industry for the common good, on the understanding
that the rest are doing the same. Thus arise the separate trades of
farming, building, weaving, and shoemaking ; and this division of
labour is best for the following reasons : Men and women are not
all born alike, but with special powers fitting them for special work.
Second, by attention to one occupation alone men will do much
better work than when attempting several. Third, because time is
saved and opportunities (of season, etc.) are more promptly utilized.
In this way articles are made in greater number, of better quality,
and with greater ease, than when each man is a Jack-of-all-trades
(Bonar 1992: 14-5).
As this quote shows the discussion was about the social division of labour, that
is, the separation of employments and professions within a society, rather than
the division of labour within a factory or within the limits of a single industry,
that is, the manufacturing division of labour. Bonar (1992: 34-5) argues that
while on the issue of production Aristotle was clearer than Plato he did not
deviate much from Platos earlier analysis of the division of labour. Xenophon
saw that larger cities provided a larger market for individual products which
resulted in a greater division of labour and a increased level of skill among
workers (Groenewegen 2008). With regard to Chinese writers in the period of
philosophers, which ran from the time of Confucius (551 BC-479 BC) to about
100 BC, Sun (2016: 103) writes,
Over such a long period of philosophers, a number of Chinese
thinkers wrote extensively on the division of labour. In particu-
lar, Kuan Chung (Kuan Tzu, Guan Zhong, d. 645 BC), Mencius
(c. 372-289 BC) and Hs un Tzu (Xunzi, Xun Kuang, c. 312-238
BC) explicitly discussed the division of labour and indeed carried
out rather sophisticated analyses of the subject. The necessity of
the division of labour and the division of employment into what
is termed the main genera of social production (agriculture, manu-
facturing and services) posited in the classical political economy of
the eighteenth-nineteenth centuries (see, e.g., Marx 1867/1976, p.
471, and the citations therein) had been recognised in the writings
of those authors as has long been known in the literature on the
history of Chinese philosophy and political thought (see, e.g., Fung
1937; Hsiao 1979).
As with the Greeks we see an emphasis on the social division of labour in the
work of the Chinese philosophers.

2.1.2 Medieval period

When discussing the contribution to the literature on the division of labour of
the medieval Muslim scholars in Persia Hosseini (1998: 656-7) writes,
This explicit discussion concerning division of labor is found in the
works of various Persian-speaking medieval Muslim writers including

7
 Farabi (873-950), Kai Kavus (eleventh century), Ibn Miskaway (d.
1030), Ibn Sina (980-1037), Ghazali [1058-1111], Nasir Tusi (1201-
1274), and Asaad Davani (b. 1444).
Sun (2012: 27-35) also examines the treatment of the division of labour in
medieval Islamic thought and he explains that social division of labour was dis-
cussed by al-F
ar
abi, trade and international and interregional division of labour
was discussed by al-F ar
abi, al-Ghazali and Kai Kavus and the sexual division
of labour by Ibn Sina, Nasir al-Din Tui and al-Ghazali. Sun also notes that
al-Ghazali gave an example of the manufacturing division of labour strikingly
similar to Adam Smiths famous pin factory example,
[ . . . ] al Ghazali used needle production as an example, writing that
even the small needle becomes useful only after passing thought
the hands of needle-makers about twenty-five times, each time going
through a different process (Ihya, 4:119, 390) Sun (1998: 29).
In addition Sun (2012: 28) contends that the Islamic scholarship on the division
of labour was incorporated into the Latin Scholasticss system of thought, largely
without acknowledgment, and therefore become available to later writers such
as the mercantilists and the classical economists, including Adam Smith.
Turning to the medieval Latin Scholastics Sun (2012: 35) explains that the
medieval Latin schoolmens contribution does not relate directly to the division
of labour but rather to providing a framework for study of the institutions
underlying the social division of labour. In particular the schoolmen discussed
private property, private ownership and market exchange without which the
division of labour could not develop and function. As Koehler (2016: 59) puts
it,
If private property is the personal responsibility of its owner, then
by implication it brings in its wake division of labour and, moreover,
a rightful claim to the fruits of that labour.
But McGee (1990: 472) notes that Thomas Aquinas recognised the need for,
and the benefits of, the division of labour,
Smith advanced this idea in 1776, based on his observations of Scot-
tish industry on the eve of the industrial revolution. But Aquinas
anticipated Smiths division of labor theory by 500 years.
One man does not suffice to perform all those acts deman-
ded by society, and therefore it is necessary that different
persons be occupied in different pursuits. The diversi-
fication of men for diverse tasks is the result, primarily,
of divine providence, which details the various compart-
ments of mans life in such a way that nothing necessary to
human existence is ever lacking; secondarily, this diversi-
fication proceeds from natural causes which bring it about
that different men are born with aptitudes and tendencies
for the different functions and the various ways of living.
Groenewegen (2008) points out that,

8
 By the end of the Middle Ages, social division of labour was extens-
ively practiced; manufacturing division of labour, generally speaking,
came with the Industrial Revolution.
Importantly for our purposes the practice of the social division of labour did not
generate a need for the development of a theory of production and/or the firm.
The manufacturing division of labour, eventually, did. But such developments
have been largely ignored in the mainstream literature.

2.1.3 Pre-classical economics period

Groenewegen (2008) contends that the English economics literature rediscovered
the division of labour in the late seventeenth century and it was only then that
the manufacturing version of it was development in any detail. The manufac-
turing form was linked to productivity growth, cost reduction and increased
international competitiveness. The relationship between an increased manufac-
turing division of labour and the greater extent of markets that arose due to
urbanisation was also highlighted. Authors such as William Petty saw benefits
from the division of labour in industries as varied as textiles and shipping,
[ . . . ] for as Cloth must be cheaper made, when one Cards, another
Spins, another Weaves, another Draws, another Dresses, another
Presses and Packs; than when all the Operations above-mentioned,
were clumsily performed by the same hand; so those who command
the Trade of Shipping, can build long slight Ships for carrying Masts,
Fir-Timber, Boards, Balks, &c. And short ones for Lead, Iron,
Stones &c. One sort of Vessels to Trade at Ports where they need
never lie a ground, others where they must jump upon the Sand k
twice every twelve hours; One sort of Vessels, and way of manning in
time of Peace, and cheap gross Goods, another for War and precious
Commodities; One sort of Vessels for the turbulent Sea, another for
Inland Waters and Rivers; One sort of Vessels, and Rigging, where
haste is requisite for the Maidenhead of a Market, another where 15
or 14 part of the time makes no matter. One sort of Masting and
Rigging for long Voyages, another for Coasting. One sort of Vessels
for Fishing, another for Trade. One sort for War for this or that
Country, another for Burthen only. Some for Oars, some for Poles,
some for Sails, and some for draught by Men or Horses, some for the
Northern Navigations amongst Ice, and some for the South against
Worms, & c. And this I take to be the chief of several Reasons, why
the Hollanders can go at less Freight than their Neighbours, viz. be-
cause they can afford a particular sort of Vessels for each particular
Trade (Petty 1690: 260-1).
Petty was also aware of the advantages for the division of labour that follow
from the increased size of markets in large cities,
But the Gain which is made by Manufactures, will be greater, as
the Manufacture it self is greater and better. For in so vast k City
Manufactures will beget one another, and each Manufacture will be
divided into as many parts as possible, whereby the Work of each
Artisan will be simple and easie ; As for Example. In the making

9
 of a Watch, If one Man shall make the Wheels, another the Spring,
another shall Engrave the Dial-plate, and another shall make the
Cases, then the Watch will be better and cheaper, than if the whole
Work be put upon any one Man. And we also see that in Towns, and
in the Streets of a great Town, where all the Inhabitants are almost
of one Trade, the Commodity peculiar to those places is made better
and cheaper than elsewhere (Petty 1683: 473).
Groenewegen (2008) notes that during 18th century an increasing number of
authors discussed the advantages of the division of labour:
Practical writers like Patrick Lindsay (1733), Richard Campbell
(1747) and Joseph Harris (1757) tended to concentrate on manufac-
turing division of labour using examples from linen and pin produc-
tion as well as from the familiar watch making. Those writing from
the position of moral or political philosophy, like Mandeville (1729),
Hutcheson (1755), Ferguson (1767) and Josiah Tucker (1755; 1774)
concentrated more on aspects of the social division of labour.
Groenewegen (2008) also points out that consideration of the division of la-
bour took place outside of England as well. The German Ernst Ludwig Carl
saw benefits from an international division of labour driven by differences in
climate, resource availability and locational advantages. The gains from this
international division of labour could be exploited though free trade among
countries. In France Francois Quesnay and A. R. J. Turgot dealt with the
subject. Turgot discusses the division of labour in his Reflections on the Pro-
duction and Distribution of Wealth. Turgot explains the need for a division of
labour since if everyone had to produce whatever he needed, starting from an
equal distribution of natural resources, almost no one would be able to secure
his needs. To develop the division of labour and stages of production, it is neces-
sary to accumulate large sums of capital, and to undertake extensive exchanges,
none of which is possible without money. For Turgot the division of labour
results in inequality, but this is the price of progress. Quesnay, another leading
Physiocrat, examined, albeit only briefly, the social aspects of the division of
labour in a 1765 essay entitled Natural Right. Quesnay wrote,
[ . . . ] the system of co-operation in which each person contributes
to the welfare of the society according to his ability. Everybody
makes his contribution to it in a different way, but the services per-
formed by one lessen the services which have to be performed by
another; as a result of this distribution of services, each person can
perform his own more thoroughly; and as a result of this mutual sup-
plementation each person contributes almost equally to the welfare
of the society (Quesnay 1765: 51).
Two further contributions of note are those of the Italian criminologist and
economist Cesare Beccaria and the French Encyclopedie, ou dictionnaire rais-
onne des sciences, des arts et des metiers (Encyclopaedia, or a Systematic
Dictionary of the Sciences, Arts, and Crafts) edited by Denis Diderot and Jean
le Rond dAlembert. In his work Elementi di economia publica Beccaria shows
that he was aware of the benefits of the division of labour in so far as it results in
greater skills and dexterity of workers. As to the Encyclopedie of Diderot and

10
dAlembert two articles are of note. The article on Art discussed the man-
ufacturing division of labour pointing out its benefits including improvements
in skill, improved quality of products, the saving of time and of materials, and
creating an incentive for the invention of a new machinery or the discovery of
a better ways of working. In the second article on pins (Epingle) a clear
example of the manufacturing division of labour is described in so much as it
is explained how the manufacture of pins is separated into eighteen different
operations.
An additional reference should be made to Bernard de Mandeville (1670-
1733) who also discussed the division of labour and is often accorded the credit
of inventing the phrase.9 For example F. B. Kaye wrote in the introduction to
Maneville (1988),
The celebrated phrase, too-division of labour-was anticipated by
Mandeville, and, apparently, by no one else (Mandeville 1988: v.1.
cxxxv).
Sun (2012: 49) writes,
As is widely known, it was Bernard Mandeville (1714-1729)[sic] who
coined the term division of labour [ . . . ]
and Mandeville did write about dividing and subdividing tasks,
There are many Sets of Hands in the Nation, that, not wanting
proper Materials, would be able in less than half a Year to produce,
fit out, and navigate a First-Rate : yet it is certain, that this Task
would be impracticable, if it was not divided and subdivided into a
great- Variety of different Labours ; and it is as certain, that none
of these Labours require any other, than working Men of ordinary
Capacities (Mandeville 1988: v2, p.142).
and No number of Men, when once they enjoy Quiet, and no Man needs
to fear his Neighbour, will be long without learning to divide and
subdivide their Labour (Mandeville 1988: v2. p. 284)
and By dividing the Employments in a great Office, and subdividing
them into many parts, every Mans Business may be made so plain
and certain, that, when he is a little used to it, it is hardly possible
for him to make Mistakes : (Mandeville 1988: v2, p.325)
For all the increasingly detailed examination of the manufacturing division
of labour there was still no development of a theory of either production or the
firm. Adam Smith and the classical economists also discussed the division of
labour without creating a theory of the firm. The theory of (industry/firm level)
production that did develop did not rely on the division of labour, it followed the
more macro (classical economics) approach of thinking of production as being
the result of inputs being transformed into output via a production function.
More on this below.
9 Backhouse (2002: 130) credits Francis Hutchinson with the phase: The phrase division

of labour was coined by Hutcheson, and the concept was widely understood in Xenophons
day. But he gives no reference.

11
2.1.4 19th century
Following on from Adam Smith a number of writers in the 19th century took up
expanding thinking on the division of labour especially the manufacturing form
of it. Among these authors Charles Babbage was one of the most prominent.
For Babbage [p]erhaps the most important principle on which the economy
of a manufacture depends, is the division of labour amongst the persons who
perform the work (Babbage 1832: 121). Babbage extended the advantages that
Smith had set forth. Babbage (1932: 122-6) listed 4 main advantages that flow
from the manufacturing division of labour: 1) Of the time required for learning.
A fewer the number of operation that have to be learned, the less time it takes
to learn them. 2) Time is always lost from changing from one occupation to
another. The fewer changes a worker has to make, the less time is lost. Also the
fewer changes in occupation the less time is lost in adjusting machinery. 3) Skill
acquired by frequent repetition of the same process. The more a worker carries
out a task the better they become at it. 4) The division of labour suggest the
contrivance of tools and machinery to execute its processes. The more a worker
carries out a task the more likely it is that they will think of improvements
in their machinery or methods of using the machines. Although all these four
advantages are important, Babbage believed that a fifth principle needed to be
added.
That the master manufacturer, by dividing the work to be executed
into different processes, each requiring different degrees of skill and
force, can purchase exactly that precise quantity of both which is ne-
cessary for each process; whereas, if the whole work were executed
by one workman, that person must possess sufficient skill to perform
the most difficult, and sufficient strength to execute the moat labori-
ous, of the operations into which the art is divided (Babbage 1932:
127; emphasis in the original).
Another important discussion of the manufacturing division of labour is
Andrew Ure (1835). Ure saw the expanded use of machinery in manufacturing
as a method for superseding skilled labour. Ure pointed out what he saw as
limitation in Smiths old principle of the division of labour. Ure argued that
under Smiths division of labour the assignment of tasks to workers is done with
reference to the skills of the worker. But the factory system calls for a new
principle of the division of labour one where machine were replacing the most
skilled of workers. It was the most difficult tasks, those which required the most
sophisticated skills in the workmen, that were being taken over by machinery.
The principle of the factory system then is, to substitute mechanical
science for hand skill, and the partition of a process into its essential
constituents, for the division or graduation of labour among artisans.
On the handicraft plan, labour more-or less skilled, was usually the
most expensive element of production - Materiam superabat opus
[The workmanship was better than the subject matter]; but on the
automatic plan, skilled labour gets progressively superseded, and
will, eventually, be replaced by mere overlookers of machines.
By the infirmity of human nature it happens, that the more skil-
ful the workman, the more self-willed and intractable he is apt to

12
 become, and, of course, the less fit a component of a mechanical sys-
tem, in which, by occasional irregularities, he may do great damage
to the whole. The grand object therefore of the modem manufac-
turer is, through: the union of capital and science, to reduce the
task of his work-people to the exercise of vigilance and dexterity,
faculties, when concentred to one process, speedily brought to
perfection in the young. ln the infancy of mechanical engineering, a
machine-factory displayed the division of labour in manifold grada-
tions the file, the drill, the lathe, having each its different workmen
in the order of skill : but the dexterous hands of the filer and driller
are now superseded by the planing, the key-groove cutting, and the
drilling-machines ; and those of the iron and brass turners, by the
self-acting slide-lathe (Ure 1835: 20-1).
For Ure the advantages due to this new system of manufacturing included im-
provements in the quality of products, savings in time and costs workmen would
otherwise pay in terms of an apprenticeship, the creation of new products that
could not be made without the machinery and improvements in the wellbeing
of workers.
It was indeed a subject of regret to observe how frequently the
workmans eminence, in any craft, had to be purchased by the sac-
rifice of his health and comfort. To one unvaried operation, which
required unremitting dexterity and diligence, his hand and eye were
constantly on the strain, or if they were suffered to swerve from their
task for a time, considerable loss ensued, either to the employer, or
the operative, according as the work was done by the day or by the
piece. But on the equalization plan of self-acting machines, the op-
erative needs to call his faculties only into agreeable exercise; he is
seldom harassed with anxiety or fatigue, and may find many leisure
moments for either amusement or meditation, without detriment to
his masters interests or his own. As his business consists in tend-
ing the work of a well regulated mechanism, he can learn it in a
short period; and when he transfers his services from one machine
to another, he varies his task, and enlarges his views, by thinking on
those general combinations which result from his and his compan-
ions labours. Thus, that cramping of the faculties, that narrowing
of the mind, that stunting of the frame, which were ascribed, and
not unjustly, by moral writers, to the division of labour, cannot,
in common circumstances, occur under the equable distribution of
industry (Ure 1835: 22-3).
Senior argued that one advantage of the division of labour omitted by Smith
was situation is which the same exertions which are necessary to produce a
single given result are often sufficient to produce many hundreds or many thou-
sands similar results. Senior used the Post Office as an example:
The same exertions which are necessary to send a single letter from
Falmouth to New York are sufficient to forward fifty, and nearly
the same exertions will forward ten thousand. If every man were
to effect the transmission of his own correspondence, the whole life

13
 of an eminent merchant might be passed in travelling, without his
being able to deliver all the letters which the Post Office forwards
for him in a single evening. The labour of a few individuals, devoted
exclusively to the forwarding of letters, produces results which all
the exertions of all the inhabitants of Europe could not effect, each
person acting independently. (Senior 1836: 74).
Senior goes on to state that Additional Labour when employed in manufactures
is MORE, when employed in Agriculture is LESS, efficient in proportion (Senior
1836: 81-6). Here Senior implicitly sees manufacturing as satisfying increasing
returns to scale while agriculture does not. Senior writes,
The proposition that, in agriculture, additional labour generally
produces a less proportionate result, or, in other words, that the
labour of twenty men employed on the land within a given district,
though it will certainly produce more than that of ten men, will
seldom produce twice as much, [ . . . ] (Senior 1936: 84)
and On the other hand, every increase in the number of manufactur-
ing labourers is accompanied not merely by a corresponding, but by
an increased productive power. If three hundred thousand families
are now employed in Great Britain to manufacture and transport
two hundred and forty millions of pounds of cotton, it is absolutely
certain that six hundred thousand families could manufacture and
transport four hundred and eighty millions of pounds of cotton. It
is, in fact, certain that they could do much more. It is not improb-
able that they could manufacture and transport seven hundred and
twenty millions (Senior 1836: 86)
One reason for this is the division of labour,
Every increase in the quantity manufactured has been accompanied
by improvements in machinery, and an increased division of labour,
and their effects have much more than balanced any increase which
may have taken place in the proportionate labour necessary to pro-
duce the raw material (Senior 1836: 84).
Interestingly in the modern literature Allen and Lueck (1998) argue that because
of seasonal forces farms can not take advantage of specialisation and the division
of labour.
With regard to the work of John Stuart Mill Groenewegen (2008) states,
Mill (1848) treated division of labour as an important aspect of co-
operation, arguing that irrespective of its well known productivity
advantages, without this complex cooperation in the modern divi-
sion of labour few things would be produced at all (Mill, 1848, p.
118) In discussing the productivity advantages, Mill cited the modi-
fication and additional advantages provided by Babbage (1832) and
Rae (1834), adding little to their discussion. However, in Chapter 9
dealing with large scale and small scale production, he highlighted
the point, so ably illustrated by Mr Babbage . . . [that] the larger the
enterprise, the farther the division of labour may be carried . . . as one
of the principal causes of large manufactories (Mill, 1848, p. 131),
thereby bringing the argument firmly into the corpus of economics.

14
Groenewegen also points out that Mills discussion of the division of labour was
largely followed by other late 19th century authors such as Fawcett (1863) and
Nicholson (1893).
We owe the distinction between the social division of labour and the manufac-
turing division of labour to Karl Marx. For Marx (1867) the difference between
the two forms of the division of labour lies in the mechanisms by which the two
versions are coordinated. For the social division of labour it is the decentralised
market exchange of commodities that coordinates while for the manufacturing
division of labour coordination take place through the exploitation of the au-
thority inherent in the employment relationship. Marx also recognised that the
manufacturing division of labour originated from developments in the social di-
vision of labour and that the manufacturing form then exerted an influence on
the social division of labour.
Since the production and the circulation of commodities are the
general prerequisites of the capitalist mode of production, division
of labour in manufacture requires that a division of labour within so-
ciety should have already attained a certain degree of development.
Inversely, the division of labour in manufacture reacts back upon
that society, developing and multiplying it further (Marx 1867:
473).
The manufacturing division of labour is what differentiates capitalism from other
modes of production.
While the division of labour in society at large, whether mediated
through the exchange of commodities or not, can exist in the most
diverse economic formations of society, the division of labour in the
workshop, as practised by the manufacture, is an entirely specific
creation of the capitalist mode of production (Marx 1867: 480).
To understand Alfred Marshalls ideas on the division of labour we will ex-
amine Book IV of his Principles of Economics. In Book IV Marshall sees
four factors as contributing to production: land, labour, capital and organisa-
tion. Marshall regards organisation as increasing the efficiency of labour and
its introduction begins his discussion of the division of labour (Marshall 1920:
240).
Marshall (1920: 241) introduces two new concepts, firstly differentiation
and, second, integration. By the first Marshall means the division of labour,
and the development of specialized skill, knowledge and machinery: , by the
second he refers to a growing intimacy and firmness of the connections between
the separate parts of the industrial organism, shows itself in such forms as
the increase of security of commercial credit, and of the means and habits of
communication by sea and road, by railway and telegraph, by post and printing-
press.
In following chapters Marshall discusses the division of labour four forms:
i) the division of labour among operatives and its relation with the use of ma-
chinery; ii) the reciprocal effects of the division of labour and the localisation of
industry; iii) the advantages of the division of labour in relation to large-scale
production; iv) the emergence of specialised business management (Marshall
1920: 250).

15
 With regard to point i) Marshall (1920: 250) notes that the division of labour
simplifies workers tasks and increases their productivity. But at a certain degree
of simplification labour is replaced by machinery (Marshall 1920: 255). There is,
however, a counterbalancing effect in that the introduction of machinery allows
us to increase the scale of manufactures and to make them more complex; and
therefore to increase the opportunities for division of labour of all kinds, and
especially in the matter of business management (Marshall 1920: 256). As an
example consider printing and watchmaking: In the printing trades, as in the
watch trade, mechanical and scientific appliances attaining results that would
be impossible without them; at the same time that they persistently take over
work that used to require manual skill and dexterity, but not much Judgment;
while they leave of a high for mans hand all those parts which do require the
use of judgment, and open up all sorts of new occupations in which machinery
there is a great demand for it (Marshall 1920: 261).
The advantages of localisation of industry, as in point ii), include the in-
teractions among people following the same skilled trade that are possible only
because they are in near proximity to each other. Also subsidiary trades can
development because there is a market for their outputs. Firms providing ac-
cess to specialised machines, too expense for any single firm to own, can develop
since there are enough small firms to keep the machine is constant use. The
concentration of firms also provides a market for highly skilled and specialised
workers and thus such workers are drawn to this location in way that they would
not be if there was only one, or a few, firms.
As to point iii) the major advantages of large scale production that Marshall
sees are to do with the economising on skill and machinery and savings of mater-
ials. Marshall argues that large scale production allows firms to use specialised
machines since the large firm can keep the machine in constant employment
which a smaller firm can not (Marshall 1920: 280). In addition the large firm is
better able to afford the fixed cost involved with the invention and development
of new machinery.
Marshall divides the economies that arise from an increase in the scale of
production of any kind of goods into two general groups: Points i) and iii) refer
to what Marshall calls internal economies, that is, those dependent on the re-
sources of the individual houses of business engaged in it, on their organization
and the efficiency of their management while point ii) refers to external eco-
nomies, that is, those dependent on the general development of the industry
(Marshall 1920: 266).
Such economies lead Marshall into a discussion of increasing returns: The
general argument of the present Book shows that an increase in the aggregate
volume of production of anything will generally increase the size, and therefore
the internal economies possessed by such a representative firm; that it will al-
ways increase the external economies to which the firm has access; and thus will
enable it to manufacture at a less proportionate cost of labour and sacrifice than
before (Marshall 1920: 318; emphasis added). Or more precisely, The law of
increasing return may be worded thus:-An increase of labour and capital leads
generally to improved organization, which increases the efficiency of the work
of labour and capital (Marshall 1920: 318). A point to note about Marshalls
argument is that it links the division of labour to economies of scale. The larger
a firm is the better able it is to take advantage of specialisation, and thus grow
even larger. This point was also made in an earlier book, Marshall and Marshall

16
(1879):
10. It will be useful to refer to the Law of Division of Labour,
which may be stated thus :
When the demand for a commodity becomes very large,
the process of making it is generally divided among several
distinct classes of workers, each with its proper appliances,
and each aided by Subsidiary industries ; for such a divi-
sion diminishes the difficulty of making the commodity.
Anticipating a term which will be defined later on, we may say:
The Cost of production of a manufactured commodity is diminished
whenever an increase in the demand for it leads to an increased
division of labour in making it.
The Law of Division of Labour implies that an increase in the amount
of capital and labour which is applied to any process of manufacture
is likely to cause a more than proportionately increased return. It is
therefore sometimes called the Law of Increasing Return, so as to
bring out the contrast in which it stands to the Law of Diminishing
Return which applies to agriculture (Marshall and Marshall 1879:
57).
An obvious problem with increasing returns is that it leads to monopoly,
what today we refer to as a natural monopoly. But Marshall argued that pro-
duction under increasing returns may not result in monopolisation because:
(1) Some economies are external to the firm. External economies
may derive from flows of information among firms in the same
sector. This is relevant for localized industries (1910: 266), or
for small firms unable to exploit economies of scale.
(2) Manufacturers mainly produce differentiated goods, the de-
mand for which is negatively sloped, implying a definite limit
to the expansion of the firm. For Marshall it is more common
that commodities produced under increasing returns are spe-
cialized goods, i.e. they may encounter difficulties in selling
(1910: 286).
(3) Firms experience a life cycle: like the trees in a forest, sooner
of later even the most vigorous firms age, and become less dy-
namic, and their established positions are constantly under the
threat of potential new firms. Dynamics in an economic sector
is such that at any one moment some firms [are] in the ascend-
ing phase and others in the descending (1910: 317) (Lavezzi
2003: 92-3).
For our purposes the critical point here is that for the first time in our discus-
sion we see a theory of the firm being offered. In his discussion of increasing
and diminishing returns Marshall introduced his concept of the representat-
ive firm (Marshall 1920: 317). Unfortunately the representative firm was a
nebulous concept, mired in controversy, before being replaced in the econom-
ics literature during the 1920s and 30s. The representative firm has been a

17
much-criticized concept, subject to conflicting interpretations with respect to
both its configuration and its intended role in Marshalls Principles (the Prin-
ciples). The concept found itself a focal point of much of the debate during the
significant cost controversies of the 1920s; however, it has appeared infrequently
in subsequent economic analysis. In its place, the equilibrium firm has taken
centre stage in the microeconomic textbooks, sometimes being mistaken for its
vanquished predecessor (Hart 2003: 158). Importantly, however, the contro-
versy surrounding the representative firm did lead to the development of the
neoclassical theory of micro level production. More on all of this below.

2.1.5 20th century

Up to this point in nearly two thousand years of discussion starting with the
ancient Greeks and Chinese the division of labour has played little part in the
development of a theory of production/the firm but in the 20th century the man-
ufacturing division of labour does, finally, play such a role. Robinson (1931) is
one of the earliest attempts to relate the division of labour to the size of firms.
In The Structure of Competitive Industry Robinson offered an analysis of the
factors that determined the optimum size for a firm. For Robinson the inter-
action of five factors determined the size of the firm: technique, management,
finance, marketing and risk of fluctuations. These various theoretical optima
have then to be reconciled in the size or constitution of a real firm after allowing
for difficulties and anomalies of growth. The division of labour has a role to play
with regard to technique and management. Because of this we will concentrate
on these two factors here.
For Robinson the optimum firm is that firm which in existing conditions of
technique and organising ability produces at the minimum of long-run average
costs. Under the conditions of perfect competition we would expect to see the
optimum firm emerge but under conditions of imperfect competition, Robinson
notes, it may not materialise. Consider, for example, the case of monopolistic
competition in which a firm will be in equilibrium at less than the minimum of
average cost.
The first application of the division of labour to the size of the firm that
Robinson considers is the relationship between the division of labour and the
optimum technical unit. Robinson follows Adam Smith in seeing three different
reasons for the division of labour giving rise to more efficient production. First
is the increase in dexterity of workmen; secondly, the saving of time which is
commonly lost in passing from one type of work to another; and thirdly, the
invention of a great number of machines which facilitate and abridge labour,
and thus enable one person to do the work of many.
With regard to the issue of dexterity, Robinson notes Smiths observation
that a person who works at a given task for some time is likely to develop a
skill or knack for doing that task. In addition the division of labour can allow
those people with a natural skill for carrying out a given task to specialise in
that task.
Adam Smith (and Robinson) also saw an advantage in the division of labour
in that specialisation at a task saved the time that would otherwise be spent
on passing from one task to another. Time could be saved because workers do
not have to move between machines or processes. Also time would be lost if
machines had to be reset to perform a different function. The division of labour

18
saves time by concentrating both workers and machine upon a given function,
and a larger factory enjoys an advantage over a smaller one in so far as it makes
this concentration possible.
The third economy Smith saw is due to the development of specialised equip-
ment to carry out the tasks that the manufacture of an item is divided into.
Separation of a process into its constituent parts makes development of machines
to carry out those parts easier.
It is important to keep in mind when considering the size of a firm that the
principle of the division of labour requires a firm of sufficient size to obtain the
maximum profitable division of labour. This size will differ across industries
depending on the nature of the production process for that industry and how
detailed a division of labour can be implemented for that particular process.
Larger firms will, often, have the capacity to implement a greater division of
labour than a smaller firm, giving the larger firm an advantage in terms of
efficiency.
The next issue discussed by Robinson is what he calls the integration of
process. Robinson explains that often a large firm has fewer rather than more
processes of manufacture. They can utilise a large machine which has been
designed to takeover what would otherwise be a series of manual, or at least less
completely mechanical, operations. A complicated machine can perform two
or three or more consecutive processes and it can thereby eliminate the labour
and time which would be required to up the work on each of the successive
earlier machines. Only large firms can keep such a machine running at its full
capacity and this fact gives the large firm an advantage over the smaller, and
less mechanised, firm. But this difficulty can be overcome by the small firm as
long as the size of the market for the process is large enough. If a given process
requires a scale of production too great for a smaller firm the small firm can
outsource the process to specialist firm. But such outsourcing if only possible
if the extent of the market for a particular process is large enough to allow
the division of labour to develop to the point where a specialist firm is viable.
Robinson refers to this outsoucing as vertical disintegration.
The second of the areas for which Robinson sees the division of labour having
a role to play is with regard to management. A manager in a small firm will
have multiple tasks to preform, some of which he will be good at, others that he
will not be so good at. In a larger firm a division of labour can develop which
allows managers to specialise on those function for which they are best suited.
The larger firm gains in two ways from its division of managerial labour: 1)
special abilities to be used to their fullest extent. Talents are not wasted by
having managers carry out functions which could be better assigned to another
manager with a particular ability at that function. 2) a manger who specialises
in a given task will increase their knowledge of that task.
A potential downside of the managerial division of labour is the problem
of coordination. As the division of labour becomes greater the problems as-
sociated with the coordination of the different parts of the production process
also increases. As new tasks are created by dividing up the production process,
new administrative functions are also created to coordinate the ever more dis-
joint production process. The advantage that a larger firm has over the smaller
depends, in a large part, on how well it solves this coordination problem.
An additional theoretical problem with Robinsons discussion follows from
the implicate assumption in the competitive model that complete contracts can

19
be written. In such a world it is not clear why a firm is needed to carry out
production at all. As Coase (1937) first highlighted in a world of complete
contracts any organisational form can mimic any other meaning that production
could be carried out via the market just as efficiently as within a firm.
Stigler (1951) also utilises the division of labour to explain the functions of
the firm. Stigler begins his argument by saying that the division of labour, and
its limit due to the extent of the market, lies at the core of a theory of the
functions, and thus the boundaries, of a firm. Stigler outlines this theory in the
second section of his paper.
In this theory a firm is seen as engaging in a series of distinct operations
leading to the production of a final product. That is, the firm is partitioned not
among its input markets but among the functions or process that determine the
scope of its activities. And thus determine the firms boundaries.
To allow the graphical representation of the firms costs of production we
will assume that the average costs of each activity depends only on the rate of
output of the firm. In addition, if we assume that there is a constant proportion
between the rate of output of each activity and the rate of output of the final
product then all the cost functions can be drawn on the same diagram and
the vertical sum of these costs will be the conventional average cost curve for
the firm. With reference to the Figure 20.1, to produce q units of final output
requires a given number of units of activity 1, costing C1 (q), a number of unit
of activity 2, costing C2 (q), and a number of units of activity 3, costing C3 (q).
These costs can be summed to give the average cost of production for q units
of output, C1 (q) + C2 (q) + C3 (q).

P
C1 (q)+C2 (q)
+C3 (q)
AC

C1
C3 (q)

C2
C2 (q)
C3
C1 (q)

q Q
Figure 20.1. (Based on Stigler 1951: Fig. I, p. 187)

With respect to the shape of the average cost curves for the various activities,
some are increasing continuously (C1 ), some are falling continuously (C3 ) and
some are conventionally U-shaped, (C2 ).
Now consider the Adam Smiths idea that the division of labour is limited by
the extent of the market. First take the activities for which there are increasing
returns, Why doesnt the firm exploit the returns more fully and in the process

20
become a monopoly in the output market? Because as the firm expands out-
puts other activities also have to be increased and some of these are subject to
diminishing returns and these cost increases are such that they overwhelm the
cost advantages of the increasing returns and increase the average cost of the
final product. So why then does the firm not abandon these C3 -like activities
and let some other firm (and thus industry) specialise in them to exploit the
increasing returns fully? At a given time the market for these activities may be
too small to support specialised firms.10 Given this firms must perform these
activities for themselves.
But with an expansion of the market for the increasing returns activity firms
specialised in that activity would develop. The firms currently carrying out this
activity for its own consumption would forgo this activity and let it be taken
over by a new (monopoly) firm. This monopoly could not fully exploit its
market power however since it has charge a price which is less than the average
cost of production for the firm abandoning the activity. As the market for this
activity grows even larger the number of firms specialising in it grows. That is
the industry becomes increasingly competitive.
The abandonment of this activity by the original firms will change the cost
function for each firm. The cost curve, C1 , will be replaced by a horizontal line
(the black dashed line in the diagram above) in the effective region. This also
changes the average cost curve for the final product with the new curve (black
dashed curve in the diagram above) being lower than the current curve.
What about the increasing cost case? Why not abandon or reduce use of
those activities with increasing cost? Much of the previous discussion carries
over to this case with the exception that as the market and the industry grows
the original firms does not have to stop utilising that activity completely. Part
of the needed use of that activity can still be produced in-house without high
average (and marginal) cost, with the rest being purchased via the market.
A third, and more modern, approach to specialisation and the division of
labour is Becker and Murphy (1992).11 The paper starts with the idea that
productivity is increased by a more extensive division of labour since the returns
to any time spent on a task is normally greater for workers who concentrate on
a narrow range of skills. As seen with Stigler (1951) the stated argument is that
the division of labour is limited by the extent of the market. Becker and Murphy
argue that the degree of specialisation is often determined not by the size of
the market but by other factors such as the costs involved in coordinating
specialised workers who perform complementary tasks, and by the amount of
general knowledge available.
Assume a continuum of tasks, s, along a unit interval which must be per-
formed to produce a good Y . Becker and Murphy model must be performed
by the production function
Y = min Y (s) (21.1)
0s0

The rate of production from the sth task (Y (s)) is the product of the working
10 Becker and Murphy (1992: footnote 3, p. 1149) argue against this: a firm need not

specialize only in these functions. Each firm could be the sole provider of some functions
subject to increasing returns and one of several providers of functions subject to decreasing
returns.
11 This paper is clearly outside our time frame but it is one of the small group of papers

that relate the division of labour to issues to do with the firm and thus is worth considering.

21
time devoted to task s (Tw (s)) and the productivity of each hour (E(s))

Y (s) = E(s)Tw (s) (22.1)

For Becker and Murphy a team is a group of workers who cooperate, by

carrying out different tasks and functions, to produce the good Y . An obvious
interpretation of the team is a firm or a division, factory or group within a firm.
It is assumed that all workers are intrinsically identical and that all tasks are
equally difficult. Each of the intrinsically identical members of an efficient team
concentrates on an equal set of tasks, w = n1 , where the team size is denoted by
n. The output for each task depends on the size of the set and on the general
knowledge (H) available:

Y = Y (H, w) where Yh > 0, Yw < 0 (22.2)

Importantly Yw < 0 implies that there are increasing returns to specialisation.

In some cases, as examined above in the paper by Stigler, the division of
labour is limited by the extent of the market. Under other situations it is
limited by other factors. Conflict among members of the team grows as the
team grows. Principal-agent conflicts, hold-up problems and breakdowns in
supply and communications all tend to incease as the degree of specialisation
grows. Becker and Murphy call such problems part of the costs of coordinating
specialists and assume that the total coordination costs per team member (C)
depends on n (or w).
C = C(n), Cn > 0. (22.3)
Net output per member y is the difference between benefits and costs,

y = B C = B(H, n) C(n), Bn > 0, Cn > 0 (22.4)

Note that if B did not depend on n then one-member teams would be efficient
just as long as C increases in n. If C were independent of n, the division of
labour would be limited only by N , the extend of the market assuming that B
increases in n. In situations where both Bn > 0 and Cn > 0 any efficient team
would have more than one member but less members than all workers in the
market. The efficient amount of specialisation can be found by differentiating
equation 22.4 with respect to n. This gives

Bn Cn (22.5)

The second order condition is Bnn Cnn < 0 and it is assumed that Bn > Cn
for small n.
B, C B()

C()

Cn
Bn

n = N
Figure 22.1.

22
 If Bn > Cn for all n N , then the division of labour is only limited by the
extent of the market. Consider, by way of an illustration, the case where B and
C are linear with Bn > Cn , then the optimal n = N , the extent of the market.
See Figure 22.1.
If this is not the case then the optimal n < N is found where Bn = Cn . In
these situations the division of labour is limited not by the size of the market
but rather by coordination costs. See Figure 23.1.

B, C B()
C()

Bn = Cn

Cn = Bn

n N
Figure 23.1.

Any analysis of the costs of coordinating tasks can offer insights into the
organisation of firms. Team members within a firm are coordinated by the
rules of the firm. Firms are less vertically integrated, they do less in-house,
in situations when it is cheaper to coordinate team members through market
transaction, that is, when transaction costs are lower. So the size of the firm
can be seen as being limited, not by the size of the market, but by coordination
costs.
At this point Becker and Murphy take advantage of a particular example:

E(s) = dH Th (s) (23.1)

where > 0 determines the productivity of Th - the time devoted to acquiring

task-specific skills. H is general knowledge and is assumed to increase the
productivity of time spent on investing in skills, > 0. The total time spent on
the sth skill is denoted T (s) and thus

T (s) = Th (s) + Tw (s) (23.2)

That is, time is either devoted to investing Th or working Tw so that output

is maximised. This implies,12

Y (s) = A()H T (s)1+ (23.3)

12 Using equation 22.1
Y (s) = E(s)Tw (s)
dH Th (s)(T (s) Th (s)) using equations 23.1 and 23.2
dH Th (s)T (s) dH Th+1 (s)
Y (s)
= dH Th1 (s)T (s) (1 + )dH Th (s) = 0
Th (s)
dH Th1 (s)T (s) = (1 + )dH Th (s)

23
 where A = d (1 + )(1+)
Note that if a worker allocates one unit of working time uniformly between
a set w = n1 of tasks, then T (s)w = T (s) n1 = 1, i.e. T (s) = n. Substituting this
into equation 23.3 gives the output on each task as a function of the team size:

Y = AH n1+ (24.1)

Output per worker is given by

Y
y= = B(H, N ) = AH n . (24.2)
n
From this equation it is clear that B rises as the size of the team increases
as long as > 0, that is, just as long as investments in task-specific skills have
a positive marginal productivity.
Becker and Murphy assume that in equation 23.2 an increase in human
capital not only increases the average product per team, but also the marginal
product of a bigger team.
 
B
= Bnh > 0 (24.3)
H n

Remember that the first-order condition given by equation 24.2 gives us the
maximum income per worker, if we differentiate this with respect to H we get13
dn Bnh
= >0 (24.4)
dH Cnn Bnn
where Bnn Cnn < 0 is the second-order condition. Equation 24.4 tells us
that teams get larger and workers become specialised as human capital and
technological knowledge grows.
(1 + )
T (s)Th1 = Th (s)


Th (s) = T (s)
(1 + )
Y (s) = dH Th (s)(T (s) Th (s)) using equations 22.1, 23.1 and 23.2
     

= dH T (s) T (s) T (s) using the Th (s) term
1+ 1+
   
1
= dH T (s) T (s)
1+ 1+
= dH (1 + )(1+) T 1+ (s)
= A()H T 1+ (s) where A() = d (1 + )(1+)

13 Equation 24.2 gives Bn (H, n) Cn (n) 0. Remember that n is a function of H.

Bn Bn dn Cn dn
= + =0
H H n dH n dH
 
Bn dn Cn Bn
=
H dH n n
dn Bnh
=
dH Cnn Bnn

24
 Equation 24.3 determines how workers with different knowledge get allocated
to different areas of a firm. The costs of coordinating specialists differ among
the areas within a firm. An efficient allocation of workers assign those whose
productivity is less effected by coordination costs to high cost areas. This implies
that those workers who have lower human capital would be allocated to the high
cost sectors if greater coordination costs lower the marginal product of human
capital. The first-order condition for n and the envelope theorem gives this
result.14
2y (Bh ) n
= = Bhn <0 (25.1)
H
where is a coordination-cost-raising parameter, with Cn > 0, Bhn > 0 by an

application of Youngs Theorem to equation 24.3 and n < 0.
The paper now turns to the question of the division of labour being limited
by the extent of the market. It is noted that Adam Smith offers the most famous
statement of the idea. While Becker and Murphy largely disagree they do admit
that there are some situations where the result does hold. In their model, for
example, it is true when n , the optimal number of team members, is greater
than or equal to N , the number of workers in the market. In this case, each
worker can specialise in different tasks so that each will have some monopoly
power ex post. However, it is not true when there are many workers with
essentially the same specialised skills and they compete in the same market.
Becker and Murphy claim that the division of labour is not, in these cases,
limited by the extent of the market but by the costs of coordinating workers
with different specialties. Thus the size of a firm may depend on coordination
costs rather than the extent of the market. Becker and Murphy argue that this
can even be seen in Smiths famous pin factory example.15
There even seems to be a problem with Smiths justly famous ex-
ample of a pin factory, where workers specialize in various functions,
including drawing out, straightening, and cutting the wire. Why
14 Using equation 22.5 we get n and using the envelope theorem we know we can write n

as a function of the parameter , i.e. n = n ().

Substituting n () into equation 22.4 gives,
y =B(H, n ()) C(n ())
Differentiate this equation with respect to H (this gives the marginal product of human
capital) and then (this gives the effect of coordination costs on the marginal product of
human capital):
y
=Bh (H, n ())
H
2y Bh
=
H
n
=Bhn <0

where Bhn > 0
n
and < 0.

15 The rest of the Becker and Murphy paper deals with economic growth and the relationship

between workers who produce for current consumption and teachers who, indirectly, contribute
to future consumption by raising the human capital of others. Such issues are outside our
concerns with the division of labour and firms.

25
 didnt the several factories that made pins in Smiths England com-
bine their activities, get a larger scale and market, and specialize
more within each factory? If the answer is that the cost of combin-
ing these factories exceeded the gain from a greater division of labor,
then specialization was limited by these costs of coordination, not
by the extent of the market (Becker and Murphy 1992: 1147-8).
For all this, belated, work on a division of labour based theory of the firm, it
has been work to no avail, as will be seen below, the economic mainstream has
largely ignored such an approach when developing a theory of production/the
firm.

2.2 Pre-classical economists

While consideration of the division of labour did not give rise to a theory of the
firm or production before the 20th century, nor, it must be said, did anything
else. At different times, different authors, made observations on economic mat-
ters, including production, but none of these lead to an ongoing literature on
production. Whittaker (1940: 361) argues that the ancient Greek philosophers,
and later writers such as those of the early Christian Church, did not place
much importance on wealth and thus on production. But Whittaker (1940:
362) does note that more practically originated authors did consider aspects of
production. He gives the example of the Roman Marcus Terentius Varro:
Varro (116-27 B.C.) is an example of a Roman writer who has made
comments on production. In his Rerum rusticarum, he referred to
the way in which market forces influenced the selection of product:
These estates which have a suitable market in the neigh-
bourhood for the sale of their products, and can thence
obtain what is needed for the farm, are so far profitable
. . . Thus close to a city it pays to cultivate gardens on a
large scale, fields of violets and roses, for instance, and
many other things which a city welcomes . . . [Varro 1912:
47]
In the same book reference was made to the usefulness of rewards
in stimulating output:
You should quicken the interest of the overseas in their
work by means of rewards . . . [Varro 1912: 51]
Other matters of like nature were discussed.
Whittaker also explains that in the early Christian period attention was fo-
cused on the form of production. The question of which occupations should be
pursued, and thus what good and services should be produced, was emphasised.
In production, as in everything else, the Christian man was to be
a servant of God, occupying himself only in those activities that
received divine favor. But occasionally, especially as time went
by, comments appeared on the efficiency of production (Whittaker
1940: 362).

26
But such speculations did not evolve into an analysis of production or the firm.
In fact much before the 16th century there was little in the way of any schools
of economic thought and analysis in general.
It could be argued that the beginnings of the development of economics
proper had to wait for the writings of the Mercantilists and the Physiocrats.
These groups were were the two leading pre-classical groups of authors interested
in economic issues. But neither group offered much in terms of an analysis of
the firm.
For the Mercantilists16 there was a times much discussion of firms but it
was a limited discussion. Limited in the sense that it dealt not with issues
to do with firms per se but with effects of firms on more macro issues such
as the balance of trade. It was also limited in that it largely dealt only with
the regulated companies17 and their monopolies. When discussing the period
1649-1690, Magnusson (1994) argues that several mercantilist writes attacked
the regulated companies. Some authors argued for the adoption of measures
to end the monopoly position that regulated companies such as the Merchant
Adventurers, the Russian Company, the Levant Company and the East India
Company held. There were also debates about the effects of companies like the
East India Company on the balance of trade. Gerrad Malynes, for example,
argued that the East India Company was exporting money beyond the seas
and thus hurting Englands balance of trade. But other voices where added
to the chorus against the regulated companies as the seventeenth century pro-
gressed. In 1645, for example, an anonymous writer, in a pamphlet entitled A
Discourse Consisting of Motives for the Enlargement and Freedome of Trade,
attacked the Merchant Adventurers. The author argued that there is nothing
more . . . pernicious and destructive to any Kingdom or Common-wealth than
Monopolies . Bur regulated companies also had their defenders. In 1602 John
Wheeler defended the Merchant Adventurers saying that its traffic in cloth led
to that . . . a number of laboring men are set to work and gain much monie,
besides that which the Merchant gaineth. That is, whats good for the Adven-
turers is good for the country! In 1641 Lewell Roberts recommended that more
regulated companies should be set up. He was of the opinion that joyn one
with another in a corporation and Company, and not to kase their Traffike by
themselves asunder, or apart would lead to increased strength and maximum
benefits for a trading nation. In addition, Thomas Mun, Edward Misselden
and Sir Josiah Child had all defended the East India Company from attack at
different times.
It should be noted, however, that much of these debates were partisan rent
seeking with each side just dressing up their position in terms of the public
good. Importantly for our purposes such attacks are more policy relevant than
economics revenant. One point is that although such arguments involve firms
they do not require a theory of the firm. Just accepting that the firms do exist
is enough for policy evaluation, there is no need for an explanation of what a
firm is, what its boundaries are or what its internal organization is. So what
we see here is much like the situation with the classical economists, a largely
macroeconomic originated outlook with no need for a serious theory of the firm.
16 For a detailed discussion of mercantilism see Heckscher (1934), Viner (1937), Beer (1938:

Chapter VI), Magnusson (1994, 2003, 2015) and Ekelund and Tollison (1997).
17 See Cawston and Keane (1896), Griffiths (1974) and Ekelund and Tollison (1997: chapters

6 and 7) for a general history of the regulated companies.

27
 The aim of the Physiocrats18 was to analyse the determinants of the general
level of activity. For the Physiocrats the key variable affecting the general level
of economic activity was the capacity of agriculture to yield a net product.
The Physiocrats saw the wealth of a nation as being determined by the size of
any surplus of agricultural production over and above that needed to support
agriculture (by feeding farm labourers etc). They argued that it was only when
labour was applied to land did it create a surplus over and above what was
required for its maintenance. It was out of this surplus that all other classes in
society were supported. Agriculture alone was productive, it alone produces the
net product. Other classes in society were stipendiary, sterile or unproductive.
To see the difference between productive and unproductive note that the
Physiocrats believed that the artisan sold his output for a payment that covered
1) his production costs plus 2) subsistence wages for himself, while the cultivator
received an amount that covered 1) his production costs plus 2) his subsistence
wages plus 3) a surplus, which would be paid to the landowner as a rent. Thus
productive means productive of a surplus (Whittaker 1940: 369-70).
Johnson (1966: 617) describes the Physiocrats approach to production brief-
ly as,
The physiocratie theory of production and the associated theory of
commodity circulation formed the basis of Francois Quesnays fam-
ous Tableau economique. This table summarized reproduction and
distribution in an extensive agricultural kingdom with a population
of thirty million similar to France. The population was divided into
three classes: the productive comprising one half the population
who were engaged in agriculture, fishing, and mining; the sterile,
the quarter of the population which included manufacturers, artis-
ans, distributors, artists, professionals, and domestic servants; and
the proprietary, the quarter who owned the lands or those, such as
crown officials and church personnel, who got their support immedi-
ately from proprietor revenue. The Quesnay analysis suggested that
the economy was in a state of self-perpetuating equilibrium with
the ratios of its components remaining always the same. Economic
growth operated with equal force in all directions without altering
the proportions. The reproduction process once underway was thus
essentially circular.
For our purposes the relevant point in all of this is that this is a description
of the aggregate economy. It is a macro-level analysis with no real attention
being paid to microeconomic level agents such as the firm (or the consumer). A
theory of micro-level production or the firm is unnecessary for such analysis.

2.3 The classical economics period

As noted above the discussion of the division of labour that took place in clas-
sical economics did not result in the creation of a theory of the firm, but nether
did anything else. The classical economists did develop a theory of production
but it was, largely, a theory of macro production aimed at explaining production
18 For discussions of Physiocracy see Beer (1939), Higgs (1897) and Meek (1962)

28
of an entire economy rather than being a microeconomic theory of firm produc-
tion. OBrien (2003: 112) remarks that [c]lassical economics ruled economic
thought for about 100 years [roughly 1770-1870]. It focused on macroeconomic
issues and economic growth. Because the growth was taking place in an open
economy, with a currency that (except during 1797-1819) was convertible into
gold, the classical writers were necessarily concerned with the balance of pay-
ments, the money supply, and the price level. Monetary theory occupied a
central place, and their achievements in this area were substantial and - with
their trade theory - are still with us today. Foss and Klein (2006: 7-8) note
that from at least the time of the mercantilists and carrying on into the clas-
sical economics period economics was largely carried out at the aggregate level
with microeconomic analysis acting as little more than a handmaiden to the
macro-level investigation,
[e]conomics began to a large extent in an aggregative mode, as wit-
ness, for example, the Political Arithmetick of Sir William Petty,
and the dominant interest of most of the classical economists in
distribution issues. Analysis of pricing, that is to say, analysis of a
phenomenon on a lower level of analysis than distributional analysis,
was to a large extent only a means to an end, namely to analyze the
functional income distribution.
Lionel Robbins remarked that the classical theories of production and distribu-
tion were about determining the total wealth, or total product, of the nation:
[t]he traditional approach to Economics, at any rate among English-speaking
economists, has been by way of an enquiry into the causes determining the pro-
duction and distribution of wealth. Economics has been divided into two main
divisions, the theory of production and the theory of distribution, and the task
of these theories has been to explain the causes determining the size of the total
product and the causes determining the proportions in which it is distributed
between different factors of production and different persons (Robbins 1935:
64).
This emphasis on macro-level analysis could help explain why the classical
economists missed the opportunities they had to develop a theory of the firm.
As an example of such a missed opportunity consider Adam Smith who opens
his magnum opus, An Inquiry into the Nature and Causes of The Wealth of
Nations, with a discussion of the division of labour at the microeconomic level,
the famous pin factory example,19 but quickly moves the analysis to the market
level. When discussing Smiths approach to the division of labour McNulty
(1984: 237-8) comments,
[h]aving conceptualized division of labor in terms of the organiz-
ation of work within the enterprise, however, Smith subsequently
failed to develop or even to pursue systematically that line of ana-
lysis. His ideas on the division of labor could, for example, have led
him toward an analysis of task assignment, management, or organ-
ization. Such an intra-firm approach would have foreshadowed the
much laterindeed, quite recentefforts in this direction by Herbert
Simon, Oliver Williamson, Harvey Leibenstein, and others, a body
19 For a discussion of the origins of Smiths pin making example see Peaucelle (2006) and

Peaucelle and Guthrie (2011).

29
 of work which Leibenstein calls micro-microeconomics. [ . . . ] But,
instead, Smith quickly turned his attention away from the internal
organization of the enterprise, and outward toward the market and
the realm of exchange, perhaps because he found therein both the
source of division of labor, in the propensity in human nature . . . to
truck, barter and exchange and its effective limits.

On the other hand Zouboulakis (2015) argues that Smiths discussion of the divi-
sion of labour does offer an elementary explanation for the existence of firms. In
Zouboulakis view of Smith the existence of firms is explained through division of
labour dynamics. As the market grows more firms are created and they become
larger thereby employing more labour and capital and thus there is an increase
in specialisation and the division of labour. This in turn increases efficiency and
productivity which increases general economic wellbeing. One possible issue20
with Zouboulakiss argument is that while large firms which were able to take
advantage of the division of labour did exist in Smiths time, they were not the
norm.21 Most firms were small, with many larger firms being partnerships and
thus restricted in their ability expand, and they would cooperate in production
via long supply chains in which each firm would specialise in making a small
contribution to the overall production process. An example of such a process
in given by Smith in this description of the making of the woollen coat for a
day-labourer:
Observe the accommodation of the most common artificer or day-
labourer in a civilized and thriving country, and you will perceive
that the number of people of whose industry a part, though but a
small part, has been employed in procuring him this accommoda-
tion, exceeds all computation. The woollen coat, for example, which
covers the day-labourer, as coarse and rough as it may appear, is the
produce of the joint labour of a great multitude of workmen. The
shepherd, the sorter of the wool, the wool-comber or carder, the
dyer, the scribbler, the spinner, the weaver, the fuller, the dresser,
with many others, must all join their different arts in order to com-
plete even this homely production (Smith 1776: Book 1, Chapter
1, p. 13).
20 Iam grateful to Gavin Kennedy for making this point to me.
21 Mokyr (2002: 1223) summarises manufacturing in the UK before the Industrial Revolu-
tion by noting that
[...] large plants were not entirely unknown before the Industrial Revolution.
For instance, Pollard (1968) in his classic work on the rise of the factory, mentions
three large British plants, each employing more than 500 employees before 1750.
Perhaps the most modern of all industries was silk throwing. The silk mills in
Derby built by Thomas Lombe in 1718 employed 300 workers and were located
in a five-story building. After Lombes patent expired, large mills patterned after
his were built in other places as well. Equally famous was the Crowley ironworks,
established in 1682 in Stourbridge in the Midlands (not far from Birmingham),
which at its peak employed 800 employees. [...] In textiles, supervised workshops
production could be found before 1770 in the Devon woollen industry and in
calico printing (Chapman 1974).
The development of factories and firms during the industrial revolution is discussed in
Mokyr (2009: Chapter 15). Chartered companies were also well known, as witnessed by Adam
Smiths negative assessment of chartered companies in general and the East India Company
in particular, contained in the Wealth of Nations.

30
 Another such missed opportunity is when, from the third edition on, Smith
discusses joint-shock companies. When considering the internal organisation
of such firms Smith raises, but does not develop a theory of, what we would call
today, the principal-agent problems that arise from the separation of ownership
from control. Perhaps his most famous remark is,
[t]he directors of such companies, however, being the managers
rather of other peoples money than of their own, it cannot well
be expected, that they should watch over it with the same anxious
vigilance with which the partners in a private copartnery frequently
watch over their own. Like the stewards of a rich man, they are
apt to consider attention to small matters as not for their masters
honour, and very easily give themselves a dispensation from having
it. Negligence and profusion, therefore, must always prevail, more
or less, in the management of the affairs of such a company (Smith
1776: Book V, Chapter 1, Part III, p. 741).
But [ . . . ] Smith neither used the modern terms, agency or corporate
governance, nor developed a general theorya fact that is often overlooked
(Fleckner 2016: 22).
When writing about Adam Smiths approach to the firm Williams (1978: 11)
says, [t]he firm was disembodied and became a unit in which resources congeal
in the productive process. When we come to examine the equilibrium/value
theory of The Wealth of Nations it will be shown that, in that context, the firm
is little more than a passive conduit which assists in the movement of resources
between alternative activities. Best (2012: 29) states simply that Adam Smith
did not elaborate a theory of the firm. Fleckner (2016: 8) writes [n]ot once
does he [Adam Smith] speak of a firm, nor does he develop anything that
would resemble a theory of the firm. This lack of a theory of the firm is, in
Fleckners view, one of the weaker points of An Inquiry into the Nature and
Causes of the Wealth of Nations.
Historian of economic thought Mark Blaug summed up the classical eco-
nomics approach to the firm by arguing that the classical economists simply [
. . . ] had no theory of the firm (Blaug 1958: 226). Bowen (1955: 5-6) argues
in a similar fashion: [ . . . ] economists of the classical tradition had usually as-
sumed that the level and distribution of income and the allocation of resources
were determined by forces that could be understood without a detailed theory
of the firm. [ . . . ] Everything else would be settled by the impersonal forces of
the market, and there would be no need to consider in detail the decisions and
actions of the individual firm.

2.4 The neoclassical era

As noted previously, the classical economists neglected micro-microeconomics
in favour of a more macro based approach. OBrien (2004: 63) contends that
one important difference between classical and neoclassical economics is the
change from macro to micro level analysis:
[t]he core of neo-Classical economics is the theory of microeconomic
allocation, to which students are introduced in their first year in an

31
 elementary and largely intuitive form, and which receives increas-
ingly sophisticated statements during succeeding years of study. On
top of this, as a sort of icing on the cake, comes the macroeconomics
theory of income determination, with, in little attached boxes so to
speak, theories of growth and trade appended. But the approach of
the Classical economists was the very reverse of this. For them the
central propositions of economics concerned macroeconomic prob-
lems. Their focus above all was on the problem of growth, and the
macroeconomic distribution conclusions which followed from their
view of growth. On the one hand, international trade, at least for
Smith, was inextricably bound up with all this: on the other, the
microeconomic problems of value and microdistribution took their
place as subsets of the greater whole.
But even so, in the period following the classical economists, with the pos-
sible exception of Alfred Marshall22 , few economists wrote anything substantial
or penetrating on the theory of the firm. When reviewing the contribution of
the old institutionalists to the theory of the firm Hodgson (2012: 55) writes,
[ . . . ] we search in vain for a well-defined theory of the firm within the old
institutional economics. Carl M. Guelzo argues that one of the leading old
institutionalists, John R. Commons, [ . . . ] did not construct a rigorous theory
of the firm since this was never his purpose (Guelzo 1976: 45). With refer-
ence to the German historical school Le Texier (2013: 80) writes [m]embers of
the German historical school such as Gustav von Schmoller analysed at length
the birth and growth of the business enterprise, but they were more histori-
ans than economists. None of these thinkers proposed a theory of the business
firm. When writing about the work of Joseph Schumpeter, Hanappi (2012: 62)
says [a] well-defined theory of the firm thus cannot be found in Schumpeters
oeuvres. As to Austrian economics Per Bylund writes, [b]ut despite the focus
in Austrian economics on [ . . . ] mundane economics, and the fact that the
Austrians [have] so many necessary ingredients for a theory of the firm [ . . . ],
there is no Austrian theory of the firm (Bylund 2011: 191) and [w]hereas
the theory of the firm has been a neglected area of study in mainstream eco-
nomics, it has been missing from the Austrian economics literature (Bylund
2011: 191). Hutchison (1953: 308) comments [t]he Austrian School, with the
exception of Auspitz and Lieben, did not concern themselves much with the
analysis of markets and firms, except in respect to their general principle of
imputation. Hutchison also summarised the early neoclassical contributions to
the theory of the firm, and markets, as Jevons has little on the firm. [ . . . ]
Walrass assumptions of perfect competition (maintained virtually throughout)
and of fixed technical coefficients, limited his contribution to the analysis of
22 OBrien (1984: 25) writes, [s]erious discussion of the history of the theory of the firm

has to start with Alfred Marshall. OBriens argument is based, in the main, on Marshall
(1920a). OBrien also argues that developments subsequent to Marshall have resulted in many
of Marshalls insights being lost to succeeding generations of economists. We would therefore
argue that Marshall has left little in the way of a legacy in terms of the mainstream theory
of the firm. In addition to his views on Marshalls work and later developments OBrien also
argues that any attempt to construct a pre-Marshallian theory from the materials available
is likely to be unsuccessful. See, however, Williams (1978) for such an attempt. On the
neglect of Marshalls Industry and Trade (Marshall 1920a) see also Liebhafsky (1955). The
development of the theory of the firm from Marshall to Robinson and Chamberlin is also
dealt with in Moss (1984a).

32
firms and markets, [ . . . ]. Paretos contribution to the theory of firms and mar-
kets were not rounded off, and of very varying value, [. . . ] (Hutchison 1953:
307).
As has been pointed out by Harold Demsetz (e.g. Demsetz 1982, 1988a,
1995) before the contemporary approach to the firm, which is based around
Knight (1921b) and/or Coase (1937), the fundamental preoccupation of (micro-
) economists was with the market and the price system and hence little, or no,
attention was paid to either the firm or the consumer as separate, significant,
economic entities. Firms (and consumers) existed as handmaidens to the price
system.
This interest in the price system had its intellectual origins in the eighteenth-
century debate between the mercantilists and the free traders. This debate, to
a large degree, was about the proper role of government in the economy23 and
the model it eventually gave rise to reflects this. The (neoclassical) perfect
competition model that arose was, in a sense, the end point of a controversy
that, implicitly, began with the question, Is central planning necessary to avoid
the problems of a chaotic economic system? Adam Smiths famous answer was
no.24 Smith
23 Mercantilism requires a dominate state, not just to provide and enforce monopolies but
also to regulate and control both domestic and international trade and to direct the economy
in general. Beer (1939: 13, footnote 1) lists the characteristics of mercantilism as: (i)
Conception of money (coin and bullion or treasure) as the essence of wealth. (This conception
prevailed from the end of the Middle Ages up to the end of the seventeenth century.) (ii)
Regulating foreign trade with a view to bringing in money by the balance of trade. (iii)
Making the balance of trade the criterion of national prosperity or decline. (iv) Promotion
of manufacture by supplying it with cheap raw materials and cheap labour. (v) Protective
customs duties on, or prohibition of, import of manufactured commodities. (vi) The view
that the economic interests of nations are mutually antagonistic. Higgs (1897: 16) explains
[t]he Mercantilists seem always to have propounded to themselves the problem, How can
Government make this nation prosperous? Nationalism, state-regulation, and particularism
are the essence of their policy while Backhouse (2002: 58) notes [m]ercantilist policies
include the use of state power to build up industry, to obtain and increase the surplus of exports
over imports, and to accumulate stocks of precious metals. In France during this period
[mid-1700s] the concept [mercantilism] was utilized in order to describe an economic policy
regime characterized by direct state intervention, intended to protect domestic merchants and
manufacturers (Magnusson 2003: 46). When discussing the general economic background to
the development of the mercantile chartered companies in England, Griffiths (1974) explains
that [t]he rightand the dutyof the Crown to control the economy was taken for granted and
according to Coke the royal prerogative had an ancient and special force in the government
of trade (p. ix) and [t]he underlying concepts were those of monopolies, collective trading
or regulation of trade and the right of the Crown to control the economy (p. 3). In a
comment on Eli Heckschers view of mercantilism Deepak Lal writes that Heckscher had
argued that the mercantilist system arose as the Renaissance princes sought to consolidate
the weak states they had inherited or acquired from the ruins of the Roman Empire. These
were states encompassing numerous feuding and disorderly groups which the new Renaissance
princes sought to curb to create a nation. The purpose was to achieve unification and power,
making the States purposes decisive in a uniform economic sphere and to make all economic
activity subservient to considerations corresponding to the requirements of the State. The
mercantilist policies-with their industrial regulations, state-created monopolies, import and
export restrictions, price controls-were partly motivated by the objective of granting royal
favors in exchange for revenue to meet the chronic fiscal crisis of the state [ . . . ]. Another
objective was to extend the span of government control over the economy to facilitate its
integration (Lal 2006: 307).
24 According to Smith the government has three duties:

[t]he first duty of the sovereign, that of protecting the society from the violence and invasion
of other independent societies [ . . . ] (Smith 1776: Book V, Chapter 1, Part First, p. 689).
The second duty of the sovereign, that of protecting, as far as possible, every member of

33
 [ . . . ] realised that social harmony would emerge naturally as hu-
man beings struggled to find ways to live and work with each other.
Freedom and self-interest need not lead to chaos, but as if guided
by an invisible hand would produce order and concord. They
would also bring about the most efficient possible use of resources.
As free people struck bargains with others solely in order to better
their own condition the nations land, capital, skills, knowledge,
time, enterprise and inventiveness would be drawn automatically
and inevitably to the ends and purposes that people valued most
highly. Thus the maintenance of a prospering social order did not
require the continued supervision of kings and ministers. It would
grow organically as a product of human nature (Butler 2007: 27-8).
For Smith competitive markets were the most effective institution for co-
ordinating and motivating people to maximise the grains that result from in-
creasing specialisation and an expanded division of labour. As historian of
economic thought Mark Blaug has noted [ . . . ] a decentralised competitive
price system was held to be desirable because of its dynamic effects in widening
the scope of the market and extending the advantages of the division of labour
- in short, because it was a powerful engine for promoting the accumulation of
capital and the growth of income (Blaug 1997: 60-1). Well functioning market
institutions leave individuals free to pursue self-interested behaviour, but guide
their choices by the prices they pay and receive. For theoretical economists,
the 200 years following Smith involved a search for conditions, no matter how
abstract, under which the price system would not descend into chaos, conditions
under which markets forces would result in an equilibrium.
The formal model that arose from this search is one which eliminates all
forms of centralised or institutional control in the economy.25 It is a model, as
Harold Demsetz has explained, delineated by perfect decentralisation (Demsetz
1982, 1988a).
The power of the perfect competition model rests on a remarkable
conceptualization of an important limiting case of the coordination
problem the complete absence of conscious control by any-
one over the plans of others. Government authority is kept to the
background by the assumption of laissez-faire. The formalization
the society from injustice or oppression of every other member of it, or the duty of establishing
an exact administration of justice, [ . . . ] (Smith 1776: Book V, Chapter 1, Part II, p. 709).
The third and last duty of the sovereign or commonwealth is that of erecting and main-
taining those publick institutions and those publick works, which, though they may be in the
highest degree advantageous to a great society, are, however, of such a nature that the profit
could never repay the expense to any individual or small number of individuals, and which
it therefore cannot be expected that any individual or small number of individuals should
erect or maintain (Smith 1776: Book V, Chapter 1, Part III, p. 723). For book length
discussions of Smiths thought see, for example, Evensky (2005), Kennedy (2005, 2010) and
Otteson (2002, 2011).
25 For Adam Smith this would be an abstraction too far. Smith knew of the importance

of institutions to the proper functioning of the market economy. Mark Blaug points out that
[ . . . ] Smiths faith in the benefits of the invisible hand has absolutely nothing whatever
to do with allocative efficiency in circumstances where competition is perfect ` a la Walras
and Pareto; the effort in modern textbooks to enlist Adam Smith in support of what is now
known as the fundamental theorems of welfare economics is a historical travesty of major
proportions. For one thing, Smiths conception of competition was, as we have seen, a process
conception, not an end-state conception (Blaug 1997: 60).

34
 of institutions does the same for firms and households. The coordin-
ating function of the price system is thus isolated and highlighted
by imagining it to operate in a perfectly decentralized setting in
which no element of authority is present. The formulation of perfect
decentralization is the accomplishment of the perfect competition
model (Demsetz 1982: 8; emphasis in the original).

One reason that the emphasis on the firm diminished as the perfect com-
petition model developed was simply that the neoclassicals placed a growing
emphases on market processes, or the price system, and thus less emphases was
given to the firm (or the government or the household). As McNulty (1984:
240) explains, with regard to the partial equilibrium version of the model,
[t]he perfection of the concept of competition, beginning with the
work of A. A. Cournot and ending with that of Frank Knight, which
was at the heart of the development of economics as a science during
the nineteenth and early twentieth centuries, led on the one hand
to an increasingly rigorous analytical treatment of market processes
and on the other hand to an increasingly passive role for the firm.
Frank Knights ending point being:
[p]erfect competition is conditioned by the existence of a set of as-
sumptions, the most important of which are the following: (1) a
perfect market for productive services [ . . . ], that is, uniform prices
over the whole field (1921[a], 316); (2) complete rationality and
perfect knowledge by free and independent individuals; (3) perfect
mobility in all economic adjustments, no cost involved in movements
or changes (1921[b], 77); (4) virtually instantaneous and costless
exchange of commodities (1921[b],78); (5) perfect, continuous, cost-
less intercommunication between all individual members of the so-
ciety (1921[b], 78); (6) perfect divisibility of commodities; and (7)
an indefinitely large number of competing organizations, each of
the most efficient size (1921[a], 316) (Marchionatti 2003: 58).
Within such a framework, or the general equilibrium framework, authority,
be it in the form of a government or a firm or a household, plays no role in
coordinating resources.26 The only parameters guiding decision making are
26 The household in the neoclassical model is as lacking in substance as the firm. Kenneth

Boulding made the point that

[t]his type of analysis [the theory of the firm] is exactly analogous to the ana-
lysis of the reactions of a consumer by means of indifference curves. Indeed, a
consumer is merely a firm whose product is utility. The indifference curves
are analogous to the isoquants, or product contours, the only difference being
that they cannot be assigned definite quantities of utility. The utility surface,
whose contours form the system of indifference curves, is a mountain whose
shape we theoretically know, but whose height at any point probably cannot be
known; by contrast, we can assume that both shape and height of the production
surface are known. The substitution effect and the scale effect are likewise
known in consumption theory, where the scale effect is usually called the income
effect. Thus, a rise in the price of a single object of consumption will have a
substitution effect tending to reduce the consumption of that object as cheaper
alternatives are substituted for it. There will also be an income effect tending
to reduce all consumption, as the higher price makes the consumer poorer. The

35
those given within the model tastes and technologies and those determined
impersonally on markets prices. All parameters are outside the control of
any of the economic agents and this effectively deprives all forms of authority
a role in resource allocation. Thus we have, as noted by Demsetz, perfect
decentralisation.
The firm receives no serious consideration in terms of being a real decision
making, problem solving, hierarchical organisation in either the general equilib-
rium version of the neoclassical model, characterised by Walrass tat onnement
process, or the partial equilibrium version, characterised by Pigous equilibrium
firm.
The high water mark for neoclassical general equilibrium approach is argu-
ably Debreu (1959). For Debreu there are no firms, in the normal sense of the
word, there are just producers,
[ . . . ] when one abstracts from legal forms of organization (corpor-
ations, sole proprietorships, partnerships, . . . ) and types of activity
(Agriculture, Mining, Construction, Manufacturing, Transportation,
Services, . . . ) one obtains the concept of a producer, i.e., an eco-
nomic agent whose role is to choose (and carry out) a production
plan (Debreu 1959: 37).
It is also clear from the context that the agent referred to is a person. The only
role for the agent is to pick the profit maximising production plan from the set
of available plans. Langlois (1981: 5) explains that
[ . . . ] the interesting feature of the general-equilibrium formula-
tion is not so much that it takes as given the mix of market and
internal transactions; rather, it is that the assumptions of general-
equilibrium theory themselves actually suggest that there need be no
internal activity whatsoever. If all commodities are predetermined
for all time and the techniques for producing them are given and
fully known in all details, then one could easily conceive of a situ-
ation where every separate part of the production process would be
in the nature of a market transaction.
Like so much of modern microeconomics, when we turn to the partial equi-
librium version of the neoclassical model, it was Alfred Marshall who began the
developments that have resulted in the standard textbook theory of the firm. It
effect of a given rise in price, therefore-i.e., the elasticity of demand-depends
first on the substitutability of the commodity concerned, and, secondly, on its
importance in the total expenditure. This is true either of a consumption good
or of a factor of production (Boulding 1942: 799).
Fritz Machlup argues that the household is not the subject of study in the theory of the
consumer:
[t]he household in price theory is not an object of study; it serves only as a
theoretical link between changes in prices and changes in labor services supplied
and in consumer goods demanded. The hypothetical reactions of an imaginary
decision-maker on the basis of assumed, internally consistent preference functions
serve as the simplest and heuristically satisfactory explanation of empirical re-
lationships between changes in prices and changes in quantities. In other words,
the household in price theory is not an object of study (Machlup 1967, footnote
4, p. 9).

36
was Marshalls notion of the representative firm that began a controversy that
led to the development of the now common textbook theory of the firm. For
Marshall firms were dynamic, heterogeneous, in disequilibrium; they progressed
through a life cycle in much the same way as people. They began young and
vigorous, but after a period of maturity they became old and were displaced by
newer more efficient firms (Backhouse 2002: 179). Marshall gave us the famous
metaphor of an industry being like a forestwhile it might appear unchanged
if considered as a whole, the individual trees that make it up are constantly
changing. To reconcile his dynamic view of individual firms with the static view
of industries Marshall introduced his (somewhat imprecise) idea of a represent-
ative firm The representative firm is composed of the salient characteristics
of all firms in the industry (Moss 1984: 308). It would need to be in some
sense representative both of the cost and of the sales position of other firms
within the industry. For this to be true it would need to be representative with
respect to its business ability, age, luck, size and its access to net external eco-
nomies (Williams 1978: 102). Importantly Williams (1978: 101) argues that
the output of the representative firm will change if and only if the output of
the industry changes and any alteration in output will be in the same direction
for the representative firms as it is for the industry. Thus if the output of the
industry increases this has to mean that the price exceeds the representative
firms unit normal expenses of production.27
D. H. Macgregors interpretation of the representative firm is explained in
the following paragraph,
[t]he firm which is to be regarded as our unit is the representative
firm, the structure which is typical of a period of economic develop-
ment, which has access to all the normal economies of that period,
and is of the size which is suited to their most efficient use. It has
had a fairly long life, and fair success, is managed with normal
ability, while its size takes account of the class of goods produced,
the conditions of marketing them, and the economic environment
generally (Macgregor 1906: 9).
Macgregor (1949) uses a version of Figure 38.1 to illustrate the properties of
the representative firm. In the diagram the supply is denoted by 0 M and the
price by M P . The line Q R28 shows the average costs at which different firms,
placed from left to right in the order of their efficiencies, are able to produce.
27 Williams (1978: 101) also notes that a firms long-period supply price (average of normal

expenses) includes income forgone by investing capital in this particular enterprise. So when
industry output is stable, the representative firm must be earning its opportunity income on
capital. This opportunity income is the definition of the normal rate of interest or, if earnings
of management are counted in, of profit and the representative firm will have a supply curve
found by the vertical summation of the supply schedules of the factors it uses, when the factor
supply schedules plot the amount of the factor needed to produce a unit of a given quantity
of output against the supply price of that quantity of the factor. This exercise does not give
any indication of the supply schedules of firms within the industry in question, but it possibly
helps to illustrate the meaning of the costs of particular factors per unit of output.
28 Q R is what Marshall called the particular expenses curve. Viner defines the curve

as To a curve representing the array of actual average costs of the different producers in
an industry when the total output of the industry was a given amount, these individual
costs being arranged in increasing order of size from left to right, Marshall gave the name of
particular expenses curve, [ . . . ] (Viner 1932: 44; footnotes removed). For more detail see
Appendix 1, Walker (2016).

37
The firms in the interval between S and R are, currently, high-cost producers.
They are unable to cover their full costs at the price M P . Some of these firms
will be driven from the market while others will adjust their production methods
so that they work their way to a lower cost position on the line Q R. Other
firms, some of those currently between Q and S, will, for different reasons, suffer
from lower efficiency, and thus find themselves between S and R. In Figure 38.1
the price-determining cost29 will be M P and the representative conditions are
those around S, or in other words, the representative firm is, roughly, the firm
which has average total costs equal to the market price.
Price/Cost R

S
P

Q
0 M Output
Figure 38.1 (A modified version of the diagram from Macgregor 1949: 44).

For Marshall his analysis of the firm sort to rationalise his studies of real
world firms while the idea of the industry was an abstract concept under the
umbrella of which the various producers of goods and services could be grouped
to facilitate the analysis of the matter under investigation. The role of the
representative firm was to link the dynamic view of the firm with the abstract
view of the industry.
N. Hart (2003: 1140) argues that the representative firm allowed Marshall
to have, simultaneously, the market in equilibrium while the firm is in disequi-
librium,
[i]t [the representative firm] was an avenue through which Mar-
shall conjectured a notion of equilibrium at a point in time for the
industry as a whole, while at the same time individual firms were
in disequilibrium, being subject to an organic process of change.
The representative firm therefore meets at the junction of Marshalls
biological and mechanical notions of opposed forces described in the
introductory comments in book 4 of Principles.
The representative firm has been seen as a forerunner of the representative
agent, the role of which is to stand in for the behaviour of the group, meaning
the industry for Marshall and the economy for those utilising the representative
agent (Blankenburg and Harcourt 2007: 46, Hartley 1996).
29 Recent economists, while adhering to the doctrine of marginal cost as a price-

determinant in the case of commodities subject to the law of diminishing returns, have been
disposed to accept Marshalls theory of the representative firm in the case of commodities
subject to the law of constant or decreasing cost. It is not towards the cost of production to
the least efficient producer that price gravitates, they say, but to that of the well-established,
solid business man - the man doing a conservative, prosperous, but not phenomenally brilliant
business (Wright 1919: 560).

38
 Moss (1984a,b) argues that there were three crucial steps in the movement
from the Marshallian to the now textbook view of the firm. The first step began
with the publication of Wealth and Welfare by A. C. Pigou. Pigou utilised a
formalisation of Marshalls industrial taxonomy, that is the distinction between
constant, increasing and decreasing returns to scale industries, to study the
effect of these industries on the national dividend. Pigou applied this taxonomy
in both Wealth and Welfare, published in 1912, and the first, 1920, edition
of The Economics of Welfare in a wholly abstract manner and this abstract
analysis was the target of economic historian Sir John Harold Claphams attack
on Marshall and Pigou in his famous 1922 essay On Empty Economic Boxes.
Clapham argued that it was not in general possible to assign actual industries
to any one of the three categories. If we were to open these conceptual boxes,
Clapham asked, would we find anything real inside? Claphams aim was to
show that these three categories could not, in their current form at least, be
usefully used by an applied researcher.
Sraffa (1926) also raised two objections to Marshalls theory.30 Robinson
(1971: 19) explains,
[w]e have already seen that supply-and-demand analysis, despite its
status as the textbook introduction to all price situations, if taken
literally, really applies only to the special case of pure competition
(that being the only case in which the back-ground of the supply
curve can be explained). In brief, Sraffas argument was this: 1) this
supply-and-demand, pure-competition package relies excessively on
the law of diminishing returns, while at the same time it is blind to
the observed fact of increasing returns; 2) the resulting analysis is
based on such restrictive assumptions as to have little application to
real-life situations.
With regard to increasing returns Sraffa argued that Marshall failed to show
that increasing returns to scale are compatible with perfect competition. The
problem for the static theory of the industry is that a firm that faces a given
price and produces under (internal) increasing returns to scale will increase its
output without limit. If one firm expands to the point that it captures the whole
market, What are we to make of perfect competition? In todays terminology
this is the problem of natural monopoly.31 Assuming external increasing returns
to scale meant reliance on a class of returns that were seldom to be met with
(Sraffa 1926: 540).
For the case of decreasing returns to scale Sraffa (1926: 538-9) argued that
such returns, and thus rising marginal cost and supply curves, are incompatible
with partial equilibrium analysis.32 First, he noted that for perfect competition
we require that it must be possible to draw each of the demand and supply
curves in such a manner that both the shape and position of the curves are
unaffected by movements along the other curve. But this mutual independence
of curves can not be assumed if the production of a given commodity employs
a considerable part of a input that is fixed in quantity. For any increase in
30 Sraffa began this assault on Marshalls theory with a paper in Italian, Sraffa (1925). The

first few pages of Sraffas 1926 paper are a summary of the arguments made in the 1925 paper.
An English translation of Sraffa (1925) appeared as Sraffa (1998).
31 For an introduction to the theory of natural monopoly see Sharkey (1982).
32 See Robinson (1969: 116-9) and Shackle (1967: 13-21) for more detailed discussion.

39
the production of the commodity there will be a corresponding increase the
unit price of the fixed factor due competition for that input from other goods
that utilise it. Thus the prices of these other goods, be they substitutes or
complements, will increase and this will alter the conditions of demand for the
original good.
If, on the other hand, the first commodity employs just a small fraction of the
available amount of the fixed factor, any increase in its use will have little effect
on the factors price or the average cost of production. This means that under
perfect competition it is difficult to account for either an increasing average cost
curve or an increasing marginal cost curve. This in turn implies that upward
sloping supply curves are difficult to rationalise under perfect competition. A
more modern way of saying this is to note that perfect competition applies to
the input markets as well as the output markets and thus an industry is able to
purchase its inputs at the market price which is independent of that industrys
output. If all industries expand output then we get decreasing returns but this
assumption violates the ceteris paribus assumption because one thing being kept
constant is the output of other industries.33
Robinson (1971: 20) counters Sraffas argument by noting two points. First
he notes that in the short-run at least, a firms stock of plant and equipment is
fixed and is fact is enough to ensure, assuming a sufficient increase in the firms
level of output, higher per-unit costs, that is, diminishing returns. Second,
Robinson explains that Sraffas argument to do with the long-run amounts to
little more than the claim that the long-run pure-competition supply curve is
(approximately) flat.
As part of a response to Claphams claim of empirical irrelevance and Sraffas
claim of logical incoherence Pigou argued that for carrying out comparative
static analysis Marshalls highly complex analytical starting point in a popu-
lation of heterogenous disequilibrium firms was, strictly speaking, unnecessary.
Pigou insisted on the possibility-and, indeed, desirability-of eliminating this
complexity (Foss 1994a: 1121). Pigous response is the second of Mosss three
steps and, importantly, involved Pigou introducing, as a way to help eliminate
complexity, the equilibrium firm. In Pigou (1928: 239-40) he describes the
equilibrium firm, at some length, as
[m]ost industries are made up of a number of firms, of which at
any moment some are expanding, while others are declining. Mar-
shall, it will be remembered, likens them to trees in a forest. Thus,
even when the conditions of demand are constant and the output
of an industry as a whole is correspondingly constant, the output
of many individual firms will not be constant. The industry as a
whole will be in a state of equilibrium; the tendencies to expand and
contract on the part of the individual firms will cancel out; but it
is certain that many individual firms will not themselves be in equi-
librium and possible that none will be. When conditions of demand
33 Shackle (1967: 19) explains,

[i]f we allow ourselves to speak in modern terms of perfect competition, and mean by this
that prices of both product and factors to the individual firm are independent of its output,
then the conclusion of Mr Sraffas argument at this stage is the failure of perfectly competitive
assumptions to show any equilibrium of the individual firm. For both the demand curve for
its product and the curve relating unit cost to output would be horizontal straight lines. This
indictment of the perfectly competitive assumptions is Mr Sraffas first objective.

40
 have changed and the necessary adjustments have been made, the
industry as a whole will, we may suppose, once more be in equi-
librium, with a different output and, perhaps, a different normal
supply price; but, again, many, perhaps all, the firms contained in
it, though their tendencies to expand and contract must cancel one
another, will, as individuals, be out of equilibrium. This is evidently
a state of things the direct study of which would be highly complic-
ated. Fortunately, however, there is a way round. Since, when the
output of the industry as a whole is adjusted to any given state of
demand, the tendencies to expansion and contraction on the part
of individual firms cancel out, they may properly be regarded as ir-
relevant so far as the supply schedule of the industry as a whole is
concerned. When the conditions of demand change, the output and
the supply price of the industry as a whole must change in exactly
the same way as they would do if, both in the original and in the
new state of demand, all the firms contained in it were individually
in equilibrium. This fact gives warrant for the conception of what I
shall call the equilibrium firm. It implies 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 regular output y in response to
a normal supply price p, will itself also individually be in equilib-
rium with a regular output xr . The conditions of the industry are
compatible with the existence of such a firm; and the implications
about these conditions, which, whether it in fact exists or not, would
hold good if it did exist, must be valid. For the purpose of studying
these conditions, therefore, it is legitimate to speak of it as actually
existing. For any given output, then, of the industry as a whole,
the supply price of the industry as a whole must be equal to the
price, which, with the then output of the industry as a whole, leaves
the equilibrium firm in equilibrium. The industry, therefore, con-
forms to the law of increasing, constant or decreasing supply prices
according as the price which leaves the equilibrium firm in equilib-
rium increases, remains constant, or decreases with increases in the
output of the industry as a whole.
It is worth noting at this point, that for Marshall industry equilibrium had no
need for an equilibrium firm. Long run equilibrium was achieved when market
demand equaled market supply, there was no pretense that individual firms were
in equilibrium.

P rice F (y) F (y)

y

q Quantity
Figure 41.1.

41
 The construct of the equilibrium firm allowed Pigou to utilise marginal and
average cost curve diagrams to develop the idea of industries producing under
increasing returns but where any economies of scale are external to the firm
but internal to the industry. When outlining the conditions for a firm being in
equilibrium Pigou notes that these requirements involve the significant condition
that all the internal economies of scale are exhausted so that all economies have
to be external. Pigou maintained that the equilibrium firm produced at its
minimum efficient scale so that the output level of the equilibrium firm, for a
many-firm industry, would occur where the marginal cost curve cuts the average
cost curve (i.e. p = F (y)
y = F (y)) (Pigou 1928: 254). See Figure 41.1.
It was not Pigou alone who expunged Marshalls representative firm from
the economic record, Robbins (1928) was his willing accomplice.34 In Marshalls
approach to the firm there is variety among firms in terms of their products, age,
internal organization, innovation capabilities, etc. This variety causes problems.
Foss (1994a) writes,
It is this element of variety among firms that explains the introduc-
tion of the representative firm-that firm that has cost of production
equal to the industry average in long run equilibrium, is of aver-
age size, and earns normal profit. The representative firm is a
heuristic fiction, not to be found in any given industry; however,
what exactly is its analytical significance? Is it merely a statistical
summary measure? Or, does it have analytical significance, as for
example a device for comparative static analysis, knowledge of the
cost structure of the representative firm allowing qualitative predic-
tions about the average industry response if, for example, demand
changes? All this is, as Lionel Robbins [1928] pointed out, unclear
(Foss 1994a: 1120).
and then adds,
Robbins pointed out the unclear analytical status of the represent-
ative firm. But more fundamentally, he made clear that (general)
equilibrium was not inconsistent with variety among firms(Foss
1994a: 1120-1).
The last of the three steps was to assume that industries are comprised en-
tirely of equilibrium firms with identical cost curves, and to assume that firms,
as production functions,35 faced household preference (demand) functions. This
34 And their criticism was effective, years later J. N. Wolfe commented that [i]t is now

more than twenty-five years since Professor Robbinss famous article on the representative
firm finally drove that concept from the pages of economic text-books (Wolfe 1954: 337).
But as Qu er
e (2006) makes clear there was something of a revival of interest in Marshalls
analysis in the 1950s.
35 Moss (1984a: 313) notes that Pigous analysis of the equilibrium firms gave us the firm

as a production function.
Whatever its relationship to the representative firm, Pigous introduction of the equilibrium
firm gave us the firm as production function. In An Analysis of Supply, Pigou demonstrated
diagrammatically the various possible relationships between marginal and average curves on
the assumption, made clear in his algebraic analysis, that factor prices were either unchanged
or compensated. All changes in average and marginal costs were due to technological factors
alone, and since the equilibrium firm was characterized by given average and marginal cost
curves which did not shift as a result of any activity of the firm, those technological factors
were considered to be entirely exogenous to the firm. Pigous technique here was analytically
equivalent to the derivation of a cost curve from the expansion path of a production function.

42
task was carried out by Robinson (1933) and Chamberlin (1933)36 in their devel-
opment of imperfect competition and monopolistic competition, respectively.37
But as Moss (1984a: 314) points out,
[b]y assuming that every firm in the industry has an identical cost
curve, Robinson and Chamberlin stood Pigous construction of the
equilibrium firm on its head. Where Pigou argued that an equilib-
rium firm could be derived from the laws of returns obeyed by any
particular industry, Robinson and Chanberlin defined the industry
on the basis of a population of equilibrium firms.
Thus, by the 1930s the neoclassical approach38 to the firm had developed.39
But many economists would argue that the neoclassical model isnt a theory of
the firm in any meaningful sense. The output side of the standard neoclassical
model is a theory of supply or production rather than a true theory of the firm.
In neoclassical theory, the firm is a black box there to explain how changes in
inputs lead to changes in outputs. It is a black box in the sense that inputs go
in and outputs come out, without any explanation of how one gets turned into
the other. The firm is taken as given; no attention is paid to how it came into
existence, the nature of its internal organisation, where the boundary between
one firm and another is or between a firm and the market; or whether anything
would change if two firms merged and called themselves a single firm. The firm is
a conceptualisation that represents, formally, the actions of the owners of inputs
who place their inputs in the highest value uses, and makes sure that production
36 For a brief but enlightening essay on Chamberlins work see Robinson (1971).
37 Backhouse (2003: 315) writes that Robinsons Economics of Imperfect Competition [
. . . ] virtually created the modern geometry of the theory of the firm, analyzing perfect and
imperfect competition, monopoly, monopsony, and even the kinked demand curve (conven-
tionally attributed to Sweezy, 1939).
Shackle (1967: 61-2) writes,[t]he two books [Robinsons and Chamberlins] are very dif-
ferent in scope. Mrs Robinsons central concern is with the effect of supposing the demand
for firms output to be less than perfectly elastic, so that each firm, though only one among
a multitude of firms producing substitutes of varying closeness for each others products, can
exploit the essential position, powers and policies of a monopolist. She eschews discussion of
those markets where each firm reckons on other firms active retaliation to its moves, and of
expenditure on selling effort. [ . . . ] Professor Chamberlin includes selling expenditure in his
analysis with a most ingenious formal precision. He also duplicates many arguments about
price behaviour in order to point out that the entrepreneur should consider the profit pos-
sibilities of all products and choose in the end that output of that product which, with the
optimal selling expenditure, yields the biggest total profit. With these ostensibly large exten-
sions of the field, compared with Mrs Robinsons; with different emphases and a chief reliance
on different diagrammatic tools; and especially with a personal interpretation of such words
as supply and with impalpable distinctions between his own and Mrs Robinsons use of the
expressions monopoly, imperfect competition and others, Professor Chamberlin is at great
pains to insist that the two approaches are essentially different. Almost all other students of
the matter have agreed with each other that in describing the structure and mechanism of
equilibrium in firms and groups of firms when oligopoly and selling expenditure are absent,
the two books present identical theories. Chamberlin (1937) discuss the differences between
monopolistic and imperfect competition.
38 Or the marginalist theory of the firm as it was often referred to at this time. Mongin

(1997: 558) notes that marginalist was the commonly used term in the 1940s and 50s, as
the term neo-classical was not yet popular. The term marginalist was still being used in
the 1960s as witnessed by the title of Fritz Machlups 1967 American Economic Association
Presidential Address, Theories of the Firm: Marginalist, Behavioral, Managerial (Machlup
1967).
39 Puu (1970: 230), for example, writes, [ . . . ] the theory of the firm had, in substance,

been developed to its present state by 1940.

43
is separated from consumption. The firm produces only for outsiders, there is
no on-the-job or internal consumption, no self-sufficiency. In fact there are no
managers or employees to indulge in on the job consumption and as production
is separated from consumption, no self-sufficiency. Production for outsiders is,
according to Demsetz (1995), the definition of a firm in the neoclassical model:
[w]hat is needed is a concept of the firm in which production is
exclusively for sale to those formally outside the firm. This require-
ment defines the firm (for neoclassical theory), but it has little to do
with the management of some by others. The firm in neoclassical
theory is no more or less than a specialized unit of production, but
it can be a one-person unit (Demsetz 1995: 9).
As inputs are combined in the optimal fashion by the actions of independent in-
put owners motivated solely by market prices, there is no need for management
of some by others, there is no role for managers or employees. Also note that as
competition assures the absence of profits and losses in (long-run) equilibrium,
there is no need to have a residual claimant. This means that, in one sense at
least, there are no owners of the firm.40 As there are no physical assets controlled
by the firm, there are no (residual) control rights over these assets to allocate.
This implies there are no owners of the firm in the Grossman-Hart-Moore sense.
About the partial equilibrium approach to the firm Klein (1996: 5) writes,
[i]n neoclassical economic theory, the firm as such does not exist at
all. The firm is a production function or production possibilities
set, a means of transforming inputs into outputs. Given the avail-
able technology, a vector of input prices, and a demand schedule,
the firm maximizes money profits subject to the constraint that its
production plans must be technologically feasible. That is all there
is to it. The firm is modeled as a single actor, facing a series of relat-
ively uncomplicated decisions: what level of output to produce, how
much of each factor to hire, and so on. These decisions, of course,
are not really decisions at all; they are trivial mathematical calcu-
lations, implicit in the underlying data. In the long run, the firm
may also choose an optimal size and output mix, but even these are
determined by the characteristics of the production function (eco-
nomies of scale, scope, and sequence). In short: the firm is a set of
cost curves, and the theory of the firm is a calculus problem.
The neoclassical production function is a way of representing the black box
conversion of inputs into outputs but tells us little about the inner workings
of the black box. The production function is independent of the institutional
framework of output creation. Thus it represents the firm without explaining
the firm.
That the theory cannot explain the boundaries of the firm has been noted
by several authors. Williamson (1993: 4), for example, asks,
40 Hansmann (1996), for example, states [a] firms owners, as the term is conventionally

used and as it will be used here, are those persons who share two formal rights: the right
to control the firm and the right to appropriate the firms profits, or residual earnings (that
is, the net earnings that remain with the firm after it has made all payments to which it is
contractually committed, such as wages, interest payments, and prices for supplies) (p. 11).
He later adds [n]ot all firms have owners. In nonprofit firms, in particular, the persons who
have control are barred from receiving residual earnings (p. 12).

44
 [w]hat determines which activities a firm chooses to do for itself
and which it procures from others?
A simple answer to that question is that the natural boundaries
of the firm are defined technology-economies of scale, technological
nonseparabilities, and the like. The firm-as-production function is
in this tradition. [ . . . ] In mundane terms, the issue is that of make-
or-buy. What is it that determines which transactions are executed
how?
That posed a deep puzzle for which the firm-as-production function
approach had little to contribute.
In a discussion of the neoclassical model with regard to its application to the
modelling of global production Antr`as (2016: 13-4) adds that,
[ . . . ] if firms could foresee all possible future contingencies, and
if they could costlessly write contracts that specify in an enforce-
able manner the course of action to be taken in all of these pos-
sible contingencies, then firms would no longer need to worry about
controlling the workers, the internal divisions, or the supplying
firms with whom they interact in production. The complete con-
tract would in fact confer full control to the firm regardless of the
ownership structure that governs the transactions between all these
producers. In other words, and as Coase (1937) anticipated more
than seventy-five years ago, firm boundaries are indeterminate in a
world of complete contracts.
Hart (1995: 17) criticises the neoclassical model based on three character-
istics of the theory. First, he notes that the theory completely ignores incentive
problems within the firm. The firm is a perfectly efficient black box. Second,
the theory has nothing to say about the internal organisation of the firm. Noth-
ing is said about the hierarchical structure, how decisions are made, who has
authority within a firm. Third, the theory tells us nothing about how to pin
down the boundaries of the firm. The theory is as much a theory of plant or
division size as firm size. As Hart points out,
[t]o put it in stark terms [ . . . ] neoclassical theory is consistent
with there being one huge firm in the world, with every existing
firm [ . . . ] being a division of this firm. It is also consistent with
every plant and division of an existing firm becoming a separate and
independent firm (Hart 1995: 17).
But while the neoclassical model is certainly consistent with the two interpreta-
tions that Hart delineates, Foss (2000) points out that it is also consistent with
a third possibility, that there are no firms, since consumers can do it all!
With perfect and costless contracting, it is hard to see room for
anything resembling firms (even one-person firms), since consumers
could contract directly with owners of factor services and wouldnt
need the services of the intermediaries known as firms (Foss 2000:
xxiv).

45
 Nearly 30 years before Foss wrote Cyert and Hedrick (1972) had addressed
similar points. They argued that in the neoclassical system the firm doesnt
exist, that the theory doesnt address any of the real world problems that firms
face and that there are no organisational problems or any internal decision-
making process at all.
In one sense the controversy over the theory of the firm has arisen
over a non-existent entity. The crux of microeconomics is the com-
petitive system. Within the competitive model there is a hypothet-
ical construct called the firm. This construct consists of a single
decision criterion and an ability to get information from an external
world, called the market [8, Cyert and March, 1963, pp. 4-16].
The information received from the market enables the firm to apply
its decision criterion, and the competitive system then proceeds to
allocate resources and produce output. The market information de-
termines the behavior of the so called firm. None of the problems of
real firms can find a home within this special construct. There are
no organizational problems nor is there any room for analysis of the
internal decision-making process (Cyert and Hedrick 1972: 398).
Loasby (2015: 246) makes clear that the neoclassical model of the firm can not
explain why firms exist:
What was ironically called the theory of the firm could give no
theoretical reason for the existence of firms, because it relied en-
tirely on market transactions to explain the prices and quantities
of all goods and services. This theory simply required consumers
and producers, all conceived as individual agents: in the goods mar-
ket consumers provided the demand curve and producers the supply
curve, and in the labour market the roles were reversed. Demand
curves were conceived to be directly derived from individual pref-
erences, which were subjective but well-ordered, and supply curves
from costs, which were determined by technology and resources; and
preferences, technology and resources were all presumed to be ob-
jective data. The intersection of these curves, properly defined, was
then sufficient to determine outcomes; there was no need to explore
market processes.
Thus within the neoclassical model of the price system, the firms only role
is to allow input owners to convert inputs into outputs in response to market
prices. Firms have no internal organisation since they have no need of one, they
have no owners since there is nothing to own. Questions about the definition,
existence, internal structure and boundaries of the firm are to a large degree
meaningless within this framework since firms, by any meaningful definition of
that term, do not exist. As Foss, Lando and Thomsen (2000: 632) summarise
it:
[t]he pure analysis of the market institution leaves almost no room
for the firm (Debreu 1959). Under the assumption of a perfect set
of contingent markets, as well as certain other restrictive assump-
tions, the model describes how markets may produce efficient out-
comes. The question how organizations should be structured does

46
 not arise, because market-contracting perfectly solves all incentive
and coordination issues. By assumption, firm behaviour (profit max-
imization) is invariant to institutional form (e.g. ownership struc-
ture). The whole economy can operate efficiently as one great system
of markets, in which autonomous agents enter into very elaborate
contracts with each other. However, by treating the firm itself as a
black box, where internal structure, contracts, etc. disappear from
the picture, there are many other issues that the theory cannot ad-
dress. For example, the theory does not tell us why firms exist.
Despite the fact that by the 1930s the neoclassical approach was the domin-
ant theory of the firm, in its early years the basic tenants of this new orthodoxy
were the subject of a number of controversies leading to several protracted de-
bates in both the U.K. and the U.S.A.41 The most famous of these debates were
the full cost controversy42 in the U.K. and the related marginalist controversy
in the U.S. (Mongin 1992, 1998). The full cost controversy was started by the
publication in 1939 of a paper by R. L. Hall and C. J. Hitch which looked at
pricing policies of firms (Hall and Hitch 1939). On the basis of questionnaire
data Hall and Hitch argued that firms set prices in a full-cost way by estim-
ating an average-cost amount at a reference level of output and adding to it a
fixed percentage. Full-cost pricing came to be seen as a challenge to the usual
marginalist (neoclassical) profit-maximising view of the firm. Long-run profit
maximisation would only be achieved if the mark-up bore the correct relation-
ship to the firms perceived elasticities of demand. The most famous defense of
the marginalist theory came from Machlup (1946). Machlups response, how-
ever, was not solely directed towards the full-cost arguments, he also attacked
a paper by labour economist R. A. Lester which argued that the theoretical
predicts regarding the relationship between wages and employment could not
be found in the data (Lester 1946). Lester argued that [ . . . ] his empirical
research raised grave doubts as to the validity of conventional marginal theory
and the assumptions on which it rests in the following ways: (1) market de-
mand was more important in determining a firms volume of employment than
wage rates; (2) the firms cost structure was not that suggested by conventional
marginalism and its capital-labor ratio was not tied to its wage rate structure;
and (3) the practical problems involved in applying marginal analysis to the
multi-process operations of a modern plant seem insuperable, and business ex-
ecutives rightly consider marginalism impractical as an operating principle in
such manufacturing establishments [Lester 1946, pp. 81-82] (Lee 1984: 1114).
Lesters conclusion was that businessman did not adjust their employment levels
in relationship to changes in wages and productivity in a manner consistent with
the marginal theory.
At this point it is difficult to separate out the full cost controversy from
the Lester initiated marginalist controversy. In reply to both sets of arguments
Machlup [ . . . ] managed to dispute the quality and relevance of the evidence,
41 In addition to the controversies discussed below other disputes occurred throughout the

1940s and 1950s with regard to the applicability of the traditional (marginalist) theory of
the firm. See Bodenhorn (1959), Cooper (1949a, 1949b, 1951), Drucker (1958), Earley (1955,
1956), Gordon (1948), Margolis (1958, 1959) and Means (1958), and their references, for some
contributions to these debates.
42 Notes on the full cost controversy by G. B. Richardson appear as Appendix 12 of Arena

(2011).

47
and at the same time, to claim that data on price-setting were compatible with
several of the available models of imperfect competition; he also sketched a
general decision-theoretic argument to the effect that rules of thumb (the
expression in Hall and Hitch) often reflect an underlying optimizing process.
Most of the later neoclassical arguments are already in Machlups proteistic
plea. His general conclusion was that the current theory of the firm hardly
needed revising even if the allegedly damaging findings were taken at face value
(Mongin 1992: 314-5).
Effectively these controversies ended when Richard B. Heflebower presen-
ted a paper at the Conference on Business Concentration and Price Policy in
June 1952 (Heflebower 1955). Heflebower showed that full-cost pricing could
be viewed in marginalist terms. He argued that profit maximisation should
be understood in a long-run sense and that oligopoly should became the main
theoretical focus for economists. He added that the full-cost doctrine did not
constitute a well developed body of price theory and that the empirical work
on which it was based was spotty in quality and in its representation of situ-
ations (Heflebower 1955: 391). Heflebower was not alone in questioning the
quality of the data that the full-cost doctine relied upon. Earlier Haley (1948:
13) had also questioned the full-cost conclusions given the nature of the survey
data these studies utilised, [t]hose responsible for the studies have relied so
heavily upon the answers of their respondents alone, however, that it probably
would be unwise to give too much weight to their conclusions until these studies
have been supplemented by further research in the behavior and motivation of
entrepreneurs with respect to price policy.
Importantly little changed because of these controversies. As Mongin (1998:
280) notes, for the majority of economists [ . . . ] drastic adjustments in the
theory of the firm were not needed to resolve the marginalist controversy.
Overall, [a]lthough no contribution to the AER controversy [the marginalist
controversy] can be said to be decisive, it can be conjectured that it influenced
American economists into thinking that Robinsons and Chamberlins initial
models had to be refined, but that the profit-maximizing framework was flexible
enough to accommodate the available evidence (Mongin 1998: 279) and [i]t
is clear from Heflebowers masterly survey that many of the arguments used
by supporters of the fullcost principle are in no way inconsistent with orthodox
economic theory (Coase 1955: 393). In other words, these controversies had
little impact on mainstream thinking about the theory of the firm.
In summary, it took until the mid-1930s for the neoclassical (textbook) model
of the firm to develop. Before then none of the major schools of economic
thought dedicated much effort to the modelling of micro level production or
the firm. The neoclassical model, however, can be interpreted as a model with
production but without firms. Given that transaction costs are zero, consumers
do not need firms to carry out production. A number of controversies to do
with the marginalist theory arose in the period up until the mid-1950s, but
none of these had any lasting effect on the mainstream theory of the firm.
Further challenges to the neoclassical model arose in the 1950s and 1960s from
economists who developed the managerial and behavioural theories of the firm.43
In terms of the history of the theory of the firm these two sets of models are
43 A good intermediate level discussion of these models is given in Sections E and F of

Koutsoyiannis (1979).

48
particularly significant since they represent some of the first attempts to look
inside the black box of the neoclassical firm, even if their ultimate impact on
mainstream economics has also been limited.
In the next section we will briefly review each of these two approaches to
the firm in turn.

2.4.1 Behavioural and managerial models

In general terms the managerial models share the same basic assumption, they
see the firm as being manager controlled rather than owner controlled, while
the behavioural theories see the firm as a coalition of self-interested groups,
the conflicting demands of which have to be resolved via an ongoing bargaining
process within he firm. Like the neoclassical model, the managerial models work
within a maximising framework, the difference being that they maximise some
form of management utility function rather than profit. The behavioural models,
unlike both the neoclassical and managerial models, seek a satisfactory level
of their objective rather than a maximum.

2.4.1.1 Behavioural models

Economists have been developing behavioural models of the firm since the
1950s. In these models it is assumed that there is a separation between own-
ership and control. Behavioural theorists consider the consequences of conflict
between self-interested groups within firms for the way in which firms make
decisions on price, output etc. The emphasis in these models is on the internal
relations of the firm with little attention being paid to the external relations
between firms.
Although some of the seminal work on the behavioural theories can be traced
back to Simon (1955), the theory has largely been developed by Cyert and
March, with whose names it has been connected right up to today.44
In behavioural theory the company has a multiplicity of different goals. Ul-
timately these goals are set by top management via a continual process of bar-
gaining between the groups within the firm. An important point here is that
the goals take the form of aspiration levels rather than strict maximisation con-
straints. Attainment of the aspiration level satisfices the firm: the behavioural
firms behaviour is satisficing in contrast to the maximising behaviour of the
traditional firm. The firm seeks levels of profits, sales, rate of growth etc that
are satisfactory, not those that are maxima. Satisficing is seen as rational
behaviour given the limited information, time and computational skills of the
firms management. The behavioural theory redefines rationality, rationality is
now that of bounded rationality.45
Cyert and March argue that there are two sources of uncertainty that a firm
has to deal with. The first is uncertainty that arises from changes in market
conditions, that is, from changes in tastes, products and methods of production.
The second is uncertainty arising from the behaviour of competitors. According
44 The major reference for the behavioural model of the firm is Cyert and March (1963). For

a review of Cyert and March, after 50 years, from the perspective of organisational economics
see Gibbons (2013).
45 This refers to behavior that is intendedly rational but only limitedly so; it is a con-

dition of limited cognitive competence to receive, store, retrieve, and process information
(Williamson 1996a: 377).

49
to the behavioural theory the first form of uncertainty is avoided, as much as it
can be, by search activity, by spending on R&D and by concentrating on short-
term planning. A difference between the traditional and behavioural theories is
the importance given in the behavioural theory to the short-run, at the expense
of the long-run. To avoid competitor-originated uncertainty, Cyert and March
argue that firms operate within a negotiated environment, that is, firms act
collusively with their competitors.
The instruments the behavioural firm uses in decision-making are the same
as in the traditional theories. Both theories consider output, price and sales
strategy as the major instruments.46 The difference between the theories lies in
the way firm choose the values of these instruments. In the neoclassical theory
such values are selected so as to maximise long-run profits. In the behavioural
theory the choice is made so that the outcome is the satisficing level of sales,
profits, growth etc.
The behavioural theory also assumes that the firm learns from its experience.
In the beginning a firm isnt a rational institution in the neoclassical sense of
global rationality. In the long run the firm may tend towards global rationality
but in the short run there is an important adaptive process of learning. Firms
make mistakes, there is trial and error from which the firm learns. In a sense
the firm has memory and learns via its past experience. An aspect of the firm
neglected by the traditional theory is the allocation of resources within the firm
and the decision-making process that leads to that allocation. In the neoclassical
theory the firm reacts to its environment, the market, while the behavioural
theory assumes that firms have some discretion and do not take the constraints
of the market as definite and impossible to change. The important point here
is that the behavioural theory looks at the mechanisms for the allocation of
resources within the firm, while the neoclassical theory examines the role of the
market, or price, mechanism for the allocation of resources between the different
sectors of the economy.
The concept of of slack is used by Cyert and March to refer to payments
made to groups within organisation over and above that needed to keep that
group in the organisation. Slack is therefore the same as economic rent accruing
to a factor of production in the traditional theory of the firm. What is significant
about the behavioural school is their analysis of the stabilising role of slack
on the activities of the firm. Changes in slack payments in periods of good and
bad business means that the firm can maintain its aspiration levels despite the
changes to its environment.

2.4.1.2 Managerial models

Another group of models, from outside the mainstream, which have been de-
veloped mainly since the 1960s in an effort to overcome some of the shortcomings
of the neoclassical model are the managerial models of the firm.47 These models
are also based on the idea that there is a difference between ownership and con-
trol of the firm. It is argued that the managers of the firm have taken control of
the firm away from the owners. The common theme running through this lit-
erature is that the managers of the firm pursue non-profit objectives, generally
46 Sales strategy here includes all activities of non-price competition, such as, advertising,

salesmanship, service, quality etc.

47 For a full treatment of dynamic models of the managerial firm see Ekman (1978).

50
subject to a performance constraint involving a profit related variable.
One of the earliest and most influential works in the managerial revolution
was Berle and Means (1932). Berle and Means began their book by arguing
that during the previous few decades there had been developing in the U.S. an
increased concentration of capital, a situation which had resulted in a small
number of firms acquiring a vast amount of power. As these firm grew in size
the ability of shareholders to control them was being weakened. Shareholdings
were being dispersed among an increasing number of small shareholders which
effectively handed control of the firm to its management. This created a moral
hazard problem since the managers interests were not fully aligned with those
of the owners. Owners preferred profits to be maximised and paid out as di-
vidends while managers favoured reinvesting them or using them to fund utility
enhancing privileges such as perks and/or higher salaries. Berle and Means
argued that this separation of ownership and control had lead to the creation
of a self-perpetuating oligarchy of managers, increasingly unaccountable to the
very people they were expected to represent, the firms owners.
De Scitovsky (1943) was a proto-managerial model of the firm. He modelled
an entrepreneur whose utility depends on income and leisure and who faces an
income/leisure trade-off given by the firms profit function. The entrepreneur
maximises his utility at a point involving more leisure, and less profit, than
the profit maximising point. More recent work, explicitly developing the mana-
gerial approach, are Baumol (1959, 1962), Williamson (1964, 1970) and Marris
(1964).48
The standard theory of the firm can be interpreted as assuming that the
managers of the firm act purely for the good of the owners. Owners can control
what the managers do and thus the managers maximise profits. There are no
principal-agent problems. Managerial models, on the other hand, start from the
twin ideas that ownership and control are separated and that managers, just
like other economic agents, act in ways that promote their own interests. But
within these models maximising assumptions are still maintained. The obvious
question this gives rise to is, What is maximised?
This question has been addressed by Baumol (1959) in which it is assumed
that managers maximise sales subject to a profit constraint and by Baumol
(1962) in which he develops a dynamic model in which the firms objective
is to maximise the growth rate of sales. Marris (1964) also assumes growth
maximisation subject to a rate of return constraint. In the Marris model a
manager has an incentive to grow a firm past its profit maximising size since
managers salaries are higher in larger firms.
While it may seem likely that profit maximisation and growth maximisation
will led to behavioural differences between the two, work by Robert Solow (Solow
1971) argues that each type of firm would react in qualitatively similar ways to
parameter changes such as changes in factor prices, excise taxes or a profit tax.
Williamson (1964, 1970) assumes a more general managerial utility func-
tion. His managerial discretion models let managers make a trade-off between
slack and profits. In the static version, slack can be taken either as excessive
administrative staff or as managerial emoluments (corporate personal consump-
tion). In the dynamic-stochastic version of Williamsons model, slack comes as
in the form of internal inefficiency, which has much in common with Leiben-
48 Alchian (1965a) gives a brief critique of the Marris (1964) and Williamson (1964) models.

51
steins (1966) notion of X-inefficiency. Williamson claims that behaviour in his
discretionary models is qualitatively different from that under profit, sales or
growth maximisation, although Rees (1974), for example, disputes aspects of
this claim.
Fritz Machlup famously attempted to repel the managerial and behavioural
attacks on the neoclassical model in his 1967 Presidential Address to the Amer-
ican Economics Association. He firstly argued that there was confusion as the
role of the firm is price theory.
My charge that there is widespread confusion regarding the pur-
poses of the theory of the firm as used in traditional price theory
refers to this: The model of the firm in that theory is not, as so
many writers believe, designed to serve to explain and predict the
behavior of real firms; instead, it is designed to explain and pre-
dict changes in observed prices (quoted, paid, received) as effects of
particular changes in conditions (wage rates, interest rates, import
duties, excise taxes, technology, etc.). In this causal connection the
firm is only a theoretical link, a mental construct helping to explain
how one gets from the cause to the effect. This is altogether dif-
ferent from explaining the behavior of a firm. As the philosopher
of science warns, we ought not to confuse the explanans with the
explanandum (Machlup 1967: 9).
He then went on to argue that those behavioural and managerial theorists
who were attacking the neoclassical model were doing so erroneously since they
were working at a different level of analysis relative to that of the neoclassical
model. The behavioural and managerial theories are aimed at the level of the
individual firm whereas the neoclassical model concerns the industry level and
thus, Machlup argued, the former are not genuine theoretical rivals to the lat-
ter.49
Lee (1984: 1122) argues that there is a connection between the behavioural
and managerial models of the firm and the marginalist controversy of the
1940s and 50s. He argues that economists, such as Baumol (1959), Cyert and
March (1963) and Marris (1964), were acquainted with and influenced by the
marginalist controversy.
While developing models of firm behaviour based on nonprofit maximising
objectives these authors showed that an augmented neoclassical framework was
compatible and consistent with full cost pricing. Thus by generalising the neo-
classical theory, albeit in different ways, these authors were able to show that it
was possible to both make price theory look more realistic and reconcile it with
the full cost pricing theory.
But the connection resulted in no real change within the mainstream since
as Mongin (1998: 280) notes [ . . . ] it would be a mistake to believe that these
writers [the behaviourists/managerialsist] were representative of the majority of
the economics profession.
49 A related level of analysis attack has been made on the present theories of the firm as

has been noted by Foss and Klein (2008: 429):

[ . . . ] the critics are protesting the application of concepts designed for analysis of markets
exchange to the study of firm organization.
That is, concepts appropriate at the market level are not appropriate at the firm level.

52
 The behavioural and managerial theories can be seen as an early attempt
to develop a theory of the firm at the level of the individual firm, a theory
which, as Oliver Williamson has said of the Cyert and March (1963) book, was
an attempt to pry open what had been a black box, thereupon to examine
the business firm in more operationally engaging ways (Williamson 1996b:
150).50 But the success of this attempt was limited. Williamsons interaction
with people such as Herbert Simon, Richard Cyert and James March while
he was at Carnegie-Mellon University did play a role in the development of
the transaction cost theory of the firm (Williamson 1996b) but outside of this
the behavioural/managerial theories have had little effect on the mainstream
economic theories of the firm.51 In fact the impact of these works may have
been greater in management than economics. Argote and Greve (2007: 337),
for example, claim that A Behavioral Theory of the Firm continues to be one
of the most influential management books of all time.

2.4.2 Coase versus Demsetz on the interpretation of the neoclassical

model
As has been explained above in the Coaseian view of the neoclassical model there
are no firms. As the model is, implicitly, a zero transaction costs model there
is no need for firms since all production can take place via market transactions.
In effect consumers can contract directly with owners of factor services to bring
about production and thus there is no need for firms to act as intermediates.
Harold Demsetz take a very different view. For Demsetz (1995: First com-
mentary) as transaction costs fall, the costs to specialisation52 fall as the use of
the market becomes cheaper and more specialisation takes place, and thus more
firms are created. As transaction costs increase, the use of the market becomes
more expensive and thus it is used less, self-sufficiency becomes more common
and the number of firms falls. Demsetz sums up the specialisation theory of
the firm as,
The bottom line of specialization theory is that firms exist because
producing for others, as compared to self-sufficiency, is efficient; this
efficiency is due to economies of scale, to specialized activity, and to
the prevalence of low, not high, transaction costs (Demsetz 1995:
11; emphasis in the original).
So in the Coaseian literature markets and firms are seen as substitutes, in
that as transaction costs fall, the market is used more and firms do less. In the
limit, as transaction costs go to zero the firm ceases to exist and all activities
50 For retrospective look at A Behavioral Theory of the Firm after 45 years see Augier and

March (2008).
51 If you look at the standard microeconomics textbooks, both undergraduate and graduate,

it is difficult to find a discussion of either behavioural or managerial models.

Koutsoyiannis (1979: section E and section F) is one of the few (even relatively recent)
undergraduate microeconomics textbooks that gives serious attention to these models, and it
is now more than 30 years old. Undergraduate textbooks on the theory of the firm from the
1970s, such as Hawkins (1973: chapter 5), Crew (1975: chapter 5) and Sawyer (1979: chapter
7 and chapter 8), did offer some discussion of these theories but there is little in the way of
more recent consideration of them.
52 For Demsetz specialisation means production being exclusively for consumption by those

outside the firm. This implies there is no self-sufficiency, where self-sufficiency means produc-
tion for ones own consumption.

53
take place via markets. In the Demsetz framework, the relationship between
firms and markets is complementary. The number of firms grows as the cost
of using the market falls and the number of firms falls as use of the market
becomes more expensive.

2.4.3 Profit maximisation

One interesting implication of the Demsetzs specialisation theory is that it
guarantees profit maximisation. Given that firms only produce for sale to those
outside the firm, there can be no on-the-job consumption and thus the owner
maximises utility by having the firm maximise profit and then saving or consum-
ing this profit in his role as a consumer. This implies an additional important
feature of the neoclassical model in that production is separated from consump-
tion. Firms produce, but do not consume while households consume but do not
produce.
But the importance of the separation between consumption and production
for profit maximisation in the neoclassical model was noted some years earlier
by Sir John Hicks. In Hicks (1946: 79) he writes:
[...] the enterprise (the conversion of factors into products) may
be regarded as a separate economic unit, detached from the private
account of the entrepreneur. It acquires factors, and sells products;
its aim is to maximize the difference between their value.
Spulber (2009: 125) calls this separation of the firms objectives and the
consumers objectives the neoclassical separation theorem, which he says makes
three assertions:
(1) firms maximise profits, (2) firms generate gains from trade com-
pared to autarky, and (3) firm decisions are separate from consumer
decisions.
Spulber argues that the neoclassical separation theorem relies on several critical
assumptions, in particular; transaction costs are absent, the firm is defined as
a producer of goods, firms take prices as given, and the Walrasian auctioneer
chooses market-clearing prices (Spulber 2009: 125). He then goes on to argue
that the theorem can be extended to include markets where firms chooses prices.
Spulber writes,
First, with price-setting firms, the chapter shows that the consumer
-owner wants the firm to maximize profit because the consumer-
owner receives the firms profits as income. Competition between
many firms selling differentiated products plays an important role in
extending the separation theorem. The consumer-owner spends the
profit earned from the firm on many goods and thus does not benefit
much from cutting prices on the products produced by the firm that
he owns. The returns to profit maximization thus outweigh changes
in the prices and production decisions of the firm that would benefit
the owner as a consumer. The analysis justifies profit maximization
by firms in industrial organization models of imperfect competition
and in market microstructure models with market-making firms
(Spulber 2009: 125-6).

54
 Mas-Colell et al. (1995: 153) observe that
[i]f prices may depend on the production of the firm, the objective
of the owners may depend on their tastes as consumers.
But if the firms owners differ in their tastes as consumers, it is possible that
the owners will not agree on what the objective of the firm should be. It may
well be that while some owners will want to maximise profits others will not
and thus the objective of profit maximisation by the firm will be lost.
The lack of separation between the role of consumer and producer can give
rise to empirical problems such as the interpretation of the results of production
function estimates. Take the example of the McDonald and Snooks (1986)
analysis of manorial production in Domesday England:
In the absence of trade, all the goods produced on the manor will
he consumed on the manor. Such a situation makes it difficult to
interpret the results of our production function estimates, because it
becomes difficult to draw a distinction between the manor as a unit
of production and as a unit of consumption: manorial production
behaviour will be inextricably combined with the lords consump-
tion preferences (or utility function). The underlying reason is that
the implicit prices of output reflect both the production and the
consumption behaviour of the manor, rather than just the costs of
manorial production (McDonald and Snooks 1986: 99).

2.5 Malmgren (1961)

Another unique, insightful, but mostly ignored pre-1970 contribution to the
theory of the firm was Harald B. Malmgrens 1961 paper Information, expect-
ations, and the theory of the firm. This paper provided the first extensive
elaboration of Coases 1937 analysis in The nature of the firm (Foss 1996:
349). Foss suggests three reason why Malmgrens paper was neglected,
First, like Coases paper it was published at a time when economists
were simply still not very interested in the theory of the firm (except
as a part of the theory of markets). Secondly, Malmgrens paper is
far from coherent and purely verbal. Thirdly, it bristles with (too)
new and provocative ideas, and draws on too many too different
writers (e.g. Jacob Marschak, Ludwig Lachmann, Tjalling Koop-
mans, George Richardson, George Stigler, Edith Penrose, Maurice
Dobb, Philip Andrews, Thomas Schelling, Sune Carlsson, etc.)
(Foss 1996: 350).
To this day Malmgrens paper is little, if ever, cited.
Inspiration for the analysis in Malmgrens paper comes from Coase (1937),
Hayek (1937) and the work of Edith Penrose and George Richardson.
From Coase (1937) comes the somewhat unusual question: why do
multi-person, multi-process firms exist in a competitive economy?
(Malmgren, 1961, p. 399); from Hayek (1937) comes the inspiration
for an accompanying question: Given a state of limited informa-
tion in which that information is not equally distributed among all

55
 possible productive units, what can be said about the nature of the
firm? (Hayek, 1937, p. 411); from Penrose comes the idea of the
managerial learning process and of excess resources as the resource
basis for related diversification; and from Richardson comes insights
in the coordination of complementary investment projects and in
how real-world markets accomplish this coordination (Richardson,
1960) (Foss 1996: 354).
Malmgren begins his analysis by noting that there must be firms in a market
economy since without them consumers could not alone establish market prices.
Foss interprets Malmgren as saying that without firms, there would not be
product markets as we ordinarily think of them. Without firms consumers would
have contract directly with the owners of factors to arrange for production to
take place. In such a world there would in fact be factor markets either and the
factor and product markets would ne one and the same. This is the situation
in a world of zero transaction costs (Foss 1996: 354). So to amke room for
firms in the model we need positive transaction costs. Malmgren takes the
view that uncertainty about the current and future events underlies positive
transaction cots, since if all events were predictable the price mechanism would
render its service at no costs. One source of transaction costs noted by Coase
is discovering what the relevant prices are (Coase 1937: 21) and Malmgren
follows this line of enquiry.
One question that arise when dealing with a decentralised economy is how
equilibrium, in the sense that all agents plans are compatible, could ever arise.
Hayek (1945) argued that the main instution that brings about the coordination
of peoples plans is the price system. Coase (1937) noted that the price system
dos not work costlessly. In addition Richardson (1960) points out that left to
its own devices the price system would not always bring about the coordination
of all plans.
The book [Information and Investment] proposes that with per-
fect competition between autonomous, interchangeable agents and
with perfect information, entrepreneurs would in fact not know what
to do. News of an increase in demand would be available to all and,
in perfect conditions, all would be capable of responding. If all were
to respond, however, there would be gross overinvestment, no-one
would gain, and the market would collapse. Consequently, such en-
trepreneurs would either fight to destruction or do nothing at all. [
. . . ] The ideally rational decision maker would face a curious version
of the tragedy of the commons, wherein no one grazes unless they
know that there are constraintslimited awareness, limited capabil-
ity, limited competitive understandingon others grazing (Duguid
n.d.).
In other words even under perfect competition producers may not be able to
make decisions. Each producers decision will depend the decisions of all other
producers, and since each producer individually does not have access to inform-
ation about other producers decisions there is no rational basis for making any
decision at all. This amounts to saying the long-run supply curve does not exist
in perfect competition. What is worse, prices can not disseminate information
as Hayek would suggest since nobody will be able to make any decisions that
would influence prices.

56
 Thus,
The role played in real economies by industry specific norms, by
contracts, by various kinds of stickiness in the system, by cartels,
etc., is to make it possible to make rational investment decisions.
Sluggish adjustment, ignorance, capital market rationing, etc means
that not every producer can immediately react to a new profit oppor-
tunity (for example, when demand changes); vertical contracts and
integration ease the coordination of complementary investments. As
Richardson describes it, this is the institutional substratum for the
working of the price mechanism. The imperfections he, like Mar-
shall, does so much to highlight are necessary imperfections, as it
were (Foss 1996: 356).
By combining Penrose (1959) with Hayeks and Richardsons work Malmgren is
able to see the firm as a information-controlling, knowledge-bearing and creating
entity. Its knowledge base allows the firm to develop a competitive advantage
over other firms.
Internal plan consistency is the basis for equilibrium for a firm: Equilibrium
for the individual firm exists only for actions which make up a consistent plan
(Malmgren 1961: 405). This notion is part of what Malmgren takes from his
interpretation of Hayek (1937). Now turning to the question of the existence of
firms in a market economy Malmgren writes,
I wish now to suggest that multi-person and multi-process firms
arise in a market economy for a number of reasons which are corol-
lary to the existence of transaction, or information, costs. To begin
with, they gain an advantage over the market in certain areas of pro-
duction and sales by reducing or eliminating the costs of finding out
relevant prices. As a result a number of long-term contracts, which
may well cost no more than each individual short-term contract
which might instead be made, are arranged. This in turn provides
a large amount of controlled information: Not only are a number of
events predictable over the duration of the entire production plan,
but also less information is required to describe that set of events
for control purposes. A number of events which must fall within a
prescribed range can be identified by a single observation when it
is known in advance that there is a necessary link between them
(in this case as a result of contracting ahead for a series of related
activities). The principle of exceptions and other operating rules
of quite simple nature replace a more thorough analysis of every pos-
sible transaction which might arise in market-determined allocation
of resources over the set of activities which make up the firm. This
stabilization of events extends into the market transactions which
remain, for the firm reduces its customers costs of waiting or going
without, and thus its own costs (the costs of reclaiming buyers who
go elsewhere) by acting as an integrated market, holding inventories
and buying and selling in a way which tends to reduce fluctuations,
and thus costs. (This function could, of course, be provided by an
organized futures market, but such markets can operate only un-
der very special conditions.) Finally, the firm predicts the costs of

57
 production of its commodities better than the market could over its
set of activities by eliminating the divergence of expectations which
may arise when interdependent decisions are taken by independent
decision-makers (Malmgren 1961: 404-5).
That is, a firm may have advantages relative to the market in terms of achiev-
ing plan consistency. In addition it also relates to the idea of a firm being a
knowledge-bearing entity. Firms develop norms in an effect to maintain internal
plan consistency. All this reduces the cost of an intra-firm transaction relative
to the costs of a market or inter-firm contract.
Another reason Malmgren gives for firms having an advantage over mar-
kets is the division of knowledge. It will be worthwhile now to look into this
informational aspect in order to see how the firm exploits its informational ad-
vantages in relation to the market. Like our old friend the division of labor,
this is clearly a problem of the division of knowledge in the market (Malmgren
1961: 411). Consider the example of external economies that arise when invest-
ments made by two or more firms carried out simultaneously are profitable for
all but no single firm can gain this profit by acting independently, without the
knowledge that the other firms would follow. If all the firms were to become
divisions of a single large firm then the external economies would be internal-
ised and profits maximised by all divisions making the required investments.
Malmgren writes,
I wish to suggest, moreover, that where output and profitability
of various production units are closely interdependent, the firm is
formed to undertake decisions concerning all or some of the produc-
tion units simultaneously, so as to maximize the joint profit and total
output. What would have been external economies and therefore
not registered in the price system become internalized (Malmgren
1961: 412).
From Penrose (1959) Malmgren gains the idea that firms are heterogeneous
entities that differ from one another because each firm has its own strongly
path-dependent processes of knowledge accumulation and this can give a firm a
competitive advantage.
Since acquiring information takes time and is a costly activity some
firms will have a relative advantage in their knowledge of internal
transactions or techniques of production, market requirements, or
the types of administrative activity necessary to adapt to partic-
ular kinds of fluctuations or uncertainties in demand or supply
(Malmgren 1961: ).
According to Foss (1996) Malmgren makes three different uses of this idea that
acquiring knowledge is costly and time-consuming and therefore can give a firm
an advantage in accumulating valuable knowledge.

First, it provides the theoretical foundation for Penroses (1959)

argument that related diversification is, in general, more profitable
than unrelated diversification because it draws on existing knowledge
bases: efficient diversification reflects economizing with information
costs. Secondly, it provides a more general version of the first point

58
 by suggesting a link between the boundaries of the firm and the
notion of the firm as a knowledge creating entity.
[ ...]
Thus, firms will generally avoid integrating dissimilar activities
(Richardson, 1972).
Finally the idea of costly and time-consuming accumulation of know-
ledge allows Malmgren to identify the sources of the competitive
advantages of firms as hard-to-imitate stocks of knowledge (Foss
1996: 358-9).
An important point to note about Malmgrens approach to the firm is that
it has little or nothing to do with dealing with the misalignment of incentives.
That is, Malmgren makes no use of ideas such as opportunism and moral hazard
which are so central to the contemporary theories of the firm. Malmgren takes
more of a coordination view of the firm. [ . . . ] the precedents and customs
of firms provide coordinating Schelling points, help to stabilize the expectations
of input-owners, and therefore reduce transaction costs (Foss 1996: 361). For
Malmgren the firm coordinates complementary activities under conditions that
mean acquiring knowledge is costly, and thus knowledge is imperfect, and he
agues that the firm may achieve this coordination more efficiently than markets.

3 Discussion
The division of labour offers a framework for the development of a theory of
production/the firm but it took more than two thousand years before it was
used to do so. And even when it was used it was largely ignored. There has
been no real development of a division of labour line of inquiry in the contem-
porary theory of the firm. For the most part the pre-classical and the classical
economists utilised a theory of aggregate production, and distribution, not a
theory of micro level production or a theory of the firm. Although the neo-
classical economists sought to develop a theory of micro level production, their
theory can be seen as one without firms. Given that transaction costs are zero,
consumers do not need firms to carry out production. Consumers can contract
directly with the owners of the factors of production to arrange for the provision
of goods and services.
Questions about the definition, existence, internal structure and boundaries
of the firm are to a large degree meaningless within this framework since firms
do not exist. While there was clearly early dissatisfaction with the neoclassical
model, it was not until the 1970s that this dissatisfaction reached the point
where mainstream economists started to challenge the neoclassical model as
the standard theory of the firm. It was only then that the pioneering efforts
of Coase (1937), and to a lesser degree Knight (1921b), were recognised and
developed.53 Before 1970 The Nature of the Firm was, in Coases own words,
an article much cited and little used (Coase 1972: 63). Coase argues that
this changed both with the publication of The Problem of Social Cost (Coase
53 The majority of authors would date the beginning of the theory of the firm at Coase

(1937), with some, e.g. D. P. OBrien, starting with Alfred Marshall while others argue
for Frank Knight. Demsetz (1988b: 244), for example, writes [...] it can be said without
hesitation that Knight launched the modern theory of the firm in 1921.

59
1960) which helped rekindle interest in The Nature of the Firm via the greater
appreciation of transaction costs it brought about and with the writings of Oliver
Williamson (see, for example, Williamson 1971, 1973, 1975) who incorporated
transaction costs into the analysis of the distinction between hierarchy and
markets (Coase 1988b: 34-5). Throughout the 1970s the upsurge in interest in
the firm was maintained by work such as Alchian and Demsetz (1972), Jensen
and Meckling (1976) and Klein, Crawford and Alchian (1978) and these post-
1970 developments gave rise to the contemporary theories of the firm.
The prehistory of the theory of the firm is one of opportunities not taken.
There were frameworks, like the division of labour, available to writers for over
two thousand years but these were not used to create a theory of micro level
production, or a theory of the firm proper. This does raise the obvious question
as to why. Foss and Klein (2006: 67) argue that there is the possibility of an
empirical reason for the firm being overlooked; the relative unimportance of the
firm. Until relatively recently firms were simply not a large part of the economy.
But they also point out that such an explanation is not wholly convincing.
As noted in the Background section some ancient firms were of a reasonable
size and authors would have been aware of this. A more precise, and more
defensible, version of the argument would be that the large, vertically integrated
and diversified firm was not empirically important until recently. Thus analysing
anonymous firms, at best, may not have been a bad approximation to the
empirical realities of the time.
As an example, consider the case of rifle manufacture in Birmingham, Eng-
land, as late as the 1860s,
Of the 5800 people engaged in this manufacture within the bor-
oughs boundaries in 1861 the majority worked within a small dis-
trict round St Marys Church. . . . The reason for the high degree of
localization is not difficult to discover. The manufacture of guns,
as of jewellery, was carried on by a large number of makers who
specialized on particular processes, and this method of organization
involved the frequent transport of parts from one workshop to an-
other. The master gun-maker-the entrepreneur-seldom possessed a
factory or workshop. . . . Usually he owned merely a warehouse in
the gun quarter, and his function was to acquire semi-finished parts
and to give these out to specialized craftsmen, who undertook the
assembly and finishing of the gun. He purchased materials from the
barrel-makers, lock-makers, sight-stampers, trigger-makers, ramrod-
forgers, gun-furniture makers, and, if he were engaged in the milit-
ary branch, from bayonet-forgers. All of these were independent
manufacturers executing the orders of several master gun-makers.
. . . Once the parts had been purchased from the material-makers,
as they were called, the next task was to hand them out to a long
succession of setters-up, each of whom performed a specific oper-
ation in connection with the assembly and finishing of the gun. To
name only a few, there were those who pre-pared the front sight and
lump end of the barrels; the jiggers, who attended to the breech end;
the stockers, who let in the barrel and lock and shaped the stock;
the barrel-strippers, who prepared the gun for rifling and proof; the
hardeners, polishers, borers and riflers, engravers, browners, and fi-

60
 nally the lock-freers, who adjusted the working parts (Allen (1929:
56-7 and 116-7), quoted in Stigler (1951: 192-3)).
This is a real world approximation to anonymous firm production - that is,
fully price-decentralised, market, production.
During the neoclassical period firms were ignored simply because many eco-
nomists did not see economic theory as being relevant to business or saw the
internal workings of the firm to be outside the competence of economists. Edwin
Cannan saw the usefulness of economics as being in politics rather than busi-
ness, [t]he practical usefulness of economic theory is not in private business
but in politics, and I for one regret the disappearance of the old name political
economy, in which that truth was recognised (Cannan 1902: 60). With regard
to the relationship between economic theory and business Cannan wrote,
I do not mean to argue that a knowledge of economic theory will
enable a man to conduct his private business with success. Doubtless
many of the particular subjects of study which come under the head
of economics are useful in the conduct of business, but I doubt if
economic theory itself is. [. . . ] economic theory does not tell a man
the exact moment to leave off the production of one thing and begin
that of another; it does not tell him the precise moment when prices
have reached the bottom or the top. It is, perhaps, rather likely to
make him expect the inevitable to arrive far sooner than it actually
does, and to make him underrate, not the foresight, but the want of
foresight of the rest of the world (Cannan 1902: 459-60).
Cannan was not alone in making this type of argument, the Cambridge econom-
ist Arthur Pigou wrote:
[. . . ] it is not the business of economists to teach woollen manufac-
turers to make and sell wool, or brewers how to make and sell beer,
or any other business men how to do their job. If that was what
we were out for, we should, I imagine, immediately quit our desks
and get somebody doubtless at a heavy premium,for we should be
thoroughly inefficient to take us into his woollen mill or his brewery
(Pigou 1922: 463-4).
Lionel Robbins argued similarly, in that

The technical arts of production are simply to be grouped among

the given factors influencing the relative scarcity of different eco-
nomic goods. The technique of cotton manufacture [...] is no part
of the subject-matter of Economics [...] (Robbins 1935: 33)
There is also the question of whether before the genesis of the transaction
cost framework economists development of a theory of the firm was limited
by the lack the tools to deal with the task. Backhouse (2002: 318) has put
forward this possibility: Coase, instead, saw the firm as an organization or,
as Williamson put it, as a governance structure. This is, of course, obvious.
However, it was not until Coase introduced the idea of transaction costs that
economists had any way in which to analyse this. Many economists had studied
the organization of industry (a classic example is Marshalls Industry and Trade,

61
1919), but such work was largely descriptive. Economists had not found a
theoretical framework that could explain why industries were organized as they
were. Of course even if there is some truth to Backhouses argument there is
still the question, to what degree is this endogenous. Where tools not developed
because economists had no interest in questions about the firm?
For what ever the reason, it must be conceded that it took a long time for
prehistory to be turned into history.

62
4 References
Alchian, Armen and Harold Demsetz (1972). Production, Information
Costs and Economic Organization, American Economic Review, 62(5)
December: 777-95.
Arena, Lise (2011). From Economics of the Firm to Business Studies
at Oxford: AnvIntellectual History (1890s1990s). Unpublished doctoral
dissertation, University of Oxford, Oxford, England.
Backhouse, Roger E. (2002). The Ordinary Business of Life: A History
of Economics from the Ancient World to the Twenty-First Century, Prin-
ceton, NJ: Princeton University Press.
Becker, Gary S. and Kevin M. Murphy (1992). The Division of Labor,
Coordination Costs, and Knowledge, Quarterly Journal of Economics,
107(4) November: 1137-60.
Berle, Adolf Augustus Jr. and Gardiner C. Means (1932). The Modern
Corporation and Private Property, New York: Macmillan.
Blaug, Mark(1958). The Classical Economists and the Factory ActsA
Re-Examination, Quarterly Journal of Economics, 72(2) May: 21126.
Blaug, Mark (1997). Economic Theory in Retrospect, 5th edn., Cam-
bridge: Cambridge University Press.
Bodenhorn, Diran (1959). A Note on the Theory of the Firm, Journal
of Business, 32(2) April: 16474.
Bonar, James (1893). Philosophy and Political Economy in Some of Their
Historical Relations, London: Swan, Sonnenschein & Co.
Bresson, Alain (2014). Capitalism and the Ancient Greek Economy. In
Larry Neal and Jeffrey G. Williamson (eds.), The Cambridge History of
Capitalism (vol. I The Rise of Capitalism: From Ancient Origins to 1848)
(43-74), Cambridge: Cambridge University Press.
Bresson, Alain (2016). The Making of the Ancient Greek Economy: Insti-
tutions, Markets, and Growth in the City-States, Princeton, NJ: Princeton
University Press.
Cannan, Edwin (1902). The Practical Utility of Economic Science, The
Economic Journal, 12(8) December: 459-71.
Cawston, George and A. H. Keane (1896). The Early Chartered Compan-
ies (A.D. 1296-1858), London: Edward Arnold.
Clapham, J. H. (1922). Of Empty Economic Boxes, Economic Journal,
32(127) September: 3054.
Coase, R. H. (1937). The Nature of the Firm, Economica, New Series 4
No. 16 November: 386-405.

63
 Coase, Ronald Harry (1955). Comment (on Heflebower (1955)). In
Universities-National Bureau Committee for Economic Research, Busi-
ness Concentration and Price Policy (3924), Princeton, NJ: Princeton
University Press.
Cooper, W. W. (1949a). Theory of the Firm: Some Suggestions for
Revision, American Economic Review, 39(6) December: 120222.
Cooper, W. W. (1949b). Theory of the Firm: Some Suggestions for
Revision, II, Commission Discussion Papers: Economics: No. 255, March
29.
Cooper, W. W. (1951). A Proposal for Extending the Theory of the
Firm, Quarterly Journal of Economics, 65(1) February: 87109.
Cyert, Richard M. and Charles L. Hedrick (1972). Theory of the Firm:
Past, Present, and Future; An Interpretation, Journal of Economic Lit-
erature, 10(2) June: 398412.
Cyert, Richard M. and James G. March (1963). A Behavioral Theory of
the Firm, Englewood Cliffs, NJ: Prentice-Hall, Inc.
Davies, John K. (2007). Classical Greece: Production. In Walter Schei-
del, Ian Morris and Richard P. Saller (eds.), The Cambridge Economic
History of the Greco-Roman World (33161), Cambridge: Cambridge Uni-
versity Press.
Demsetz, Harold (1982). Economic, Legal, and Political Dimensions of
Competition, Amsterdam: North-Holland Publishing Company.
Demsetz, Harold (1988a). The Theory of the Firm Revisited, Journal of
Law, Economics, & Organization, 4(1) Spring: 14161.
Demsetz, Harold (1988b). Profit as a Functional Return: Reconsidering
Knights Views. In Harold Demsetz, Ownership, Control, and the Firm:
The Organization of Economic Activity (vol. I, 23647), Oxford: Basil
Blackwell.
Demsetz, Harold (1995). The Economics of the Business Firm: Seven
Critical Commentaries, Cambridge: Cambridge University Press.
Duguid, Paul (n.d.). Untidy or Untractable? G.B. Richardsons view of
economics, Working Paper.
Drucker, Peter (1958). Business Objectives and Survival Needs: Notes
on a Discipline of Business Enterprise, Journal of Business, 31(2) April:
8190.
Earley, James S. (1955). Recent Developments in Cost Accounting and
the Marginal Analysis , Journal of Political Economy, 63(3) June:
22742.
Earley, James S. (1956). Marginal Policies of Excellently Managed Com-
panies, American Economic Review, 46(1) March: 4470.

64
 Ekelund, Robert Jr. and Robert D. Tollison (1997). Politicized Econom-
ies: Monarchy, Monopoly, and Mercantilism, College Station: Texas A&M
University Press.
Ekman, Elan V. (1978). Some Dynamic Economic Models of the Firm:
A Microeconomic Analysis with Emphasis on Firms that Maximize Other
Goals Than Profit Alone, Stockholm: Economic Research Institute at the
Stockholm School of Economics.
Evensky, J. (2005). Adam Smiths Moral Philosophy: A Historical and
Contemporary Perspective on Markets, Law, Ethics, and Culture, Cam-
bridge: Cambridge University Press.
Foss, Nicolai J. (2000). The Theory of the Firm: An Introduction to
Themes and Contributions. In Nicolai Foss (ed.), The Theory of the
Firm: Critical Perspectives on Business and Management (xvlxi), Lon-
don: Routledge.
Foss, Nicolai J. (1996). Harald B. Malmgrens Analysis of the Firm:
Lessons for Modern Theorists?, Review of Political Economy, 8(4): 349-
66.
Foss, Nicolai J. and Peter G. Klein (2006). The Emergence of the Modern
Theory of the Firm, Center for Strategic Management and Globalization,
Copenhagen Business School, SMG Working Paper 1/2006, January.
Foss, Nicolai J. and Peter G. Klein (2008). The Theory of the Firm and
Its Critics: a Stocktaking and an Assessment. In Eric Brousseau and
Jean-Michel Glachant (eds.), New Institutional Economics: A Guidebook
(42542), Cambridge: Cambridge University Press.
Foss, Nicolai J., Henrik Lando and Steen Thomsen (2000). The Theory of
the Firm. In Boudewijn Bouckaert and Gerrit De Geest (eds.), Encyclo-
pedia of Law and Economics (vol. III, 63158), Cheltenham, UK: Edward
Elgar Publishing Ltd.
Gordon, R. A. (1948). ShortPeriod Price Determination in Theory and
Practice, American Economic Review, 38(3) June: 26588.
Griffiths, Percival (1974). A Licence to Trade: A History of the English
Chartered Companies, London: Ernest Benn Limited.
Groenewegen, Peter (2008). division of labour. In Steven N. Durlauf and
Lawrence E. Blume (eds.), The New Palgrave Dictionary of Economics,
2nd edn., Basingstoke, Hampshire New York: Palgrave Macmillan. The
New Palgrave Dictionary of Economics Online. Palgrave Macmillan. 30
December 2016 <http:www.dictionaryofeconomics.com/article?id=pde
2008 D000176>
Haley, Bernard F. (1948). Value and Distribution. In Howard S. Ellis
(ed.), A Survey of Contemporary Economics (vol. I, 148), Homewood, IL:
Richard D. Irwin.

65
 Hart, Neil (2003). From the Represntative to the Equilibrium Firm: Why
Marshall was not a Marshallian. In Richard Arena and Michel Quere
(eds.), The Economics of Alfred Marshall: Revisiting Marshalls Legacy
(158-81), Houndmills, Basingstoke, Hampshire: Palgrave Macmillan.
Hart, Oliver D. (1995). Firms, Contracts, and Financial Structure, Ox-
ford: Oxford University Press.
Hart, Oliver D. (2011). Thinking about the Firm: A Review of Daniel
Spulbers The Theory of the Firm, Journal of Economic Literature, 49(1)
March: 10113.
Hayek, F. A. von (1937). Economics and Knowledge, Economica, New
Series 4 No. 13 February: 33-54.
Hayek, F. A. von (1944). The Road to Serfdom, London: Routledge &
Kegan Paul.
Heflebower, Richard B. (1955). Full Costs, Cost Changes, and Prices. In
Universities-National Bureau, Business Concentration and Price Policy
(35992), Princeton, NJ: Princeton University Press.

Higgs, Henry (1897). The Physiocrats: Six Lectures on The French Econo-
mistes of the 18th Century, London: Macmillan and Co., Limited.

Hosseini, Hamid (1998). Seeking the Roots of Adam Smiths Division of

Labor in Medieval Persia, History of Political Economy, 30(4) Winter:
653-81.
Hosseini, Hamid (2003). Contributions of Medieval Muslim Scholars to
the History of Economics and their Impact: A Refutation of the Schum-
peterian Great Gap. In Warren J. Samuels, Jeff E. Biddle and John B.
Davis (eds.), A Companion to the History of Economic Thought (28-45),
Oxford: Blackwell Publishing Ltd.
Hutchison, T. W. (1953). A Review of Economic Doctrines 18701929,
Oxford: Oxford University Press.
Jensen, Michael C. and William H. Meckling (1976). Theory of the Firm:
Managerial Behaviour, Agency Costs, and Ownership Structure, Journal
of Financial Economics, 3(4) October: 305-60.
Jursa, Michael (2010). Business Companies in Babylonia in the First Mil-
lennium BC: Structure, Economic Strategies, Social Setting. In Myriam
Wissa (ed.), The Knowledge Economy and Technological Capabilities Egypt,
the Near East and the Mediterranean 2nd Millennium B.C.1st Millen-
nium A.D. (53-68), Sabadell: Editorial AUSA.
Kennedy, Gavin (2005). Adam Smiths Lost Legacy, Houndmills, Basing-
stoke, Hampshire: Palgrave Macmillan.
Kennedy, Gavin (2010). Adam Smith: A Moral Philosopher and His Polit-
ical Economy, 2nd edn., Houndmills, Basingstoke, Hampshire: Palgrave
Macmillan.

66
 Khanna, Vikramaditya S. (2005). The Economic History of Organiza-
tional Entities in Ancient India, University of Michigan Law School Pro-
gram in Law & Economics Working Paper No. 14 (updated version Feb-
ruary 2006).
Kikuchi, Tomoo, Kazuo Nishimura and John Stachurski (2012). Coase
meets Tarski: New Insights from Coases Theory of the Firm, Working
Paper, November 1.
Knight, Frank H. (1921a). Cost of Production and Price over Long and
Short Periods, Journal of Political Economy, 29 April: 304-35.
Knight, Frank H. (1921b). Risk, Uncertainty and Profit, Boston: Hough-
ton Mifflin Company.
Koutsoyiannis, A. (1979). Modern Microeconomics, 2nd edn., London:
The Macmillan Press Ltd.
Lafontaine, Francine and Margaret Slade (2007). Vertical Integration and
Firm Boundaries: The Evidence, Journal of Economic Literature, 45(3)
September: 62985.
Larsen, Mogens Trolle (2015). Ancient Kanesh: A Merchant Colony in
Bronze Age Anatolia, Cambridge: Cambridge University Press.

Lavezzi, Andrea (2003). Smith, Marshall and Young on division of labour

and economic growth, The European Journal of the History of Economic
Thought, 10(1) Spring: 81-108.
Lee, F. S. (1984). The Marginalist Controversy and the Demise of Full-
Cost Pricing, Journal of Economic Issues, 18(4) December: 110732.

Liebhafsky, H. H. (1955). A Curious Case of Neglect: Marshalls Industry

and Trade, Canadian Journal of Economics and Political Science, 21(3)
August: 33953.
Loasby, Brian J. (2015). Ronald Coases Theory of the Firm and the
Scope of Economics, Journal of Institutional Economics, 11(2) (Ronald
H. Coase Memorial Issue) June: 24564.
Machlup, Fritz (1967). Theories of the Firm: Marginalist, Behavioral,
Managerial, American Economic Review, 57(1) March: 133.
Magnusson, Lars G. (1994). Mercantilism: The Shaping of Economic
Language, London: Routledge.
Magnusson, Lars G. (2003). Mercantilism. In Warren J. Samuels, Jeff E.
Biddle and John B. Davis (eds.), A Companion to the History of Economic
Thought (4660), Oxford: Blackwell Publishing Ltd.
Magnusson, Lars G. (2015). The Political Economy of Mercantilism, Lon-
don: Routledge.
Margolis, Julius (1958). The Analysis of the Firm: Rationalism, Conven-
tionalism, and Behaviorism, Journal of Business, 31(3) July: 18799.

67
 Margolis, Julius (1959). Traditional and Revisionist Theories of the Firm:
A Comment, Journal of Business, 32(2) April: 17882.
Marshall, Alfred (1920a). Industry and Trade: A Study of Industrial Tech-
nique and Business Organization; and of Their Influences on the Condi-
tions of Various Classes and Nations, 2 vols., Honolulu, HI: University
Press of the Pacific, 2003.
Marshall, Alfred (1920b). Principles of Economics: An Introductory Vol-
ume, 8th edn., London: Macmillan and Co., Limited. (First published:
1890).
Marshall, Alfred and Mary Paley Marshall (1879). The Economics of
Industry, London: Macmillan and Co., Limited.
McMillan, John (2002). Reinventing the Bazaar: A Natural history of
Markets, New York: W.W. Norton and Company.
Micklethwait, John and Adrian Wooldridge (2003). The Company: A
Short History of a Revolutionary Idea, New York: The Modern Library.
Means, G. C. (1958). Looking Around, Harvard Business Review, 36(3)
MayJune: 27180.
Moss, Scott(1984a). The History of the Theory of the Firm from Marshall
to Robinson and Chamberlin: the Source of Positivism in Economics,
Economica, n.s. 51(203) August: 30718.
Moss, Scott (1984b). OBriens The Evolution of the Theory of the Firm:
a Discussion. In Frank H. Stephen (ed.), Firms, Organization and Labour:
Approaches to the Economics of Work Organization (638), London: Mac-
millan Press.
OBrien, Denis P. (1984). The Evolution of the Theory of the Firm. In
Frank H. Stephen (ed.), Firms, Organization and Labour: Approaches to
the Economics of Work Organization (2562), London: Macmillan Press.
OBrien, Denis P. (2003). Classical Economics. In Warren J. Samuels,
Jeff E. Biddle and John B. Davis (eds.), A Companion to the History of
Economic Thought (11229), Oxford: Blackwell Publishing Ltd.
OBrien, Denis P. (2004). The Classical Economists Revisited, Princeton,
NJ: Princeton University Press.
Otteson, J. R. (2002). Adam Smiths Market Place of Life, Cambridge:
Cambridge University Press.
Otteson, J. R. (2011). Adam Smith (Major Conservative and Libertarian
Thinkers vol. 16), John Meadowcroft (series ed.), New York: Continuum.
Otteson, J. R. (2014). The End of Socialism, Cambridge: Cambridge
University Press.
Penrose, E.T. (1959). The Theory of the Growth of the Firm, New York:
Wiley.

68
 Persson, Karl Gunnar (2014). Markets and coercion in medieval Europe.
In Larry Neal and Jeffrey G. Williamson (eds.), The Cambridge History
of Capitalism (vol. I The Rise of Capitalism: From Ancient Origins to
1848) (225-66), Cambridge: Cambridge University Press.
Persson, Karl Gunnar (2015). An Economic History of Europe: Know-
ledge, Institutions and Growth, 600 to the Present, 2nd edn., Cambridge
University Press.
Pigou, A. C. (1922). Empty Economic Boxes: A Reply, Economic
Journal, 32(128) December: 458-65.
Pigou, A. C. (1928). An Analysis of Supply, Economic Journal, 38(150)
June: 238-57.
Quere, Michel (2006). The Representative Firm. In Tiziano Raffaelli, Gi-
acomo Becattini and Marco Dardi (eds.), The Elgar Companion to Alfred
Marshall (4127), Cheltenham, UK: Edward Elgar.
Rees, Ray. (1974). A Reconsideration of the Expense Preference Theory
of the Firm, Economica, 41(163) August: 295307.
Richardson, G. B. (1972). The Organization of Industry, Economic
Journal, 82(327) September: 88396.
Robbins, Lionel (1928). The Representative Firm, Economic Journal,
38(151) September: 387-404.
Robbins, Lionel (1935). An Essay on the Nature & Significance of Eco-
nomic Science, 2nd edn., Revised and Extended, London: Macmillan and
Co. Limited.
Robinson, E A G. (1931). The Structure of Competitive Industry, Cam-
bridge: Cambridge University Press.
Robinson, Joan (1933). The Economics of Imperfect Competition, Lon-
don: Macmillan and Co., Ltd.
Robinson, Joan (1969). The Economics of Imperfect Competition, 2nd
edn., London: The Macmillan Press Ltd.
Robinson, Romney (1971). Edward H. Chamberlin, New York: Columbia
University Press.
Shackle, G. L. S. (1967). The Years of High Theory: Invention and Tradi-
tion in Economic Thought 19261939, Cambridge: Cambridge University
Press.
Silver, Morris(1995). Economic Structures of Antiquity, Westport Con-
necticut: Greenwood Press.
Simon, Herbert A. (1955). A Behavioral Model of Rational Choice,
Quarterly Journal of Economics, 69(1) February: 99118.
Sobel, Robert (1999). The Pursuit of Wealth: The Incredible Story of
Money Throughout the Ages, New York: McGraw-Hill

69
 Solow, Robert (1971). Some Implications of Alternative Criteria for the
Firm.In R. Marris and A. Wood, (eds.), The Corporate Economy: Growth,
Competition, and Innovative Potential (31842), London: Macmillan.
Spulber, Daniel F. (2008). Discovering the Role of the Firm: The Sep-
aration Criterion and Corporate Law, Northwestern Law & Economics
Research Paper No. 0823, December 6.
Spulber, Daniel F. (2009). The Theory of the Firm: Microeconomics with
Endogenous Entrepreneurs, Firms, Markets, and Organizations, Cam-
bridge: Cambridge University Press.
Sraffa, Piero (1925). Sulle relazioni fra costo e quantit prodotta, Annali
di Economia, 2: 277328.
Sraffa, Piero (1926). The Laws of Returns under Competitive Conditions,
Economic Journal, 36(144) December: 53550.
Sraffa, Piero (1998), On the Relations Between Cost and Quantity Pro-
duced. English translation by John Eatwell and Alessandro Roncaglia.
In L.L. Pasinetti (ed.), Italian Economic Papers (vol. 3, 32363), Bologna:
Il Mulino and Oxford: Oxford University Press.
Stigler, George J. (1951). The Division of Labor is Limited by the Extent
of the Market, Journal of Political Economy, 59(3) June: 185-93.
Sun, Guang-Zhen (2008). Nasir ad-Din Tusi on Social Cooperation and
the Division of Labor: Fragment from The Nasirean Ethics, Journal of
Institutional Economics, 4(3) December: 403-13.
Sun, Guang-Zhen (2012). The Division of Labour in Economics: A His-
tory, London: Routledge.
Sun, Guang-Zhen (2016). The economics of the division of labour in
early Chinese literature: With particular comparison to the ancient Greek
thought, The European Journal of the History of Economic Thought,
23(1) February: 102-26.
Varro, Marcus Terentius (1912). Varro On Farming: M. Terenti Varronis
Rerum Rusticarum Libri Tres, Translated, with introduction, comment-
ary, and excursus by Lloyd Storr-Best, London: G. Bell and Sons, Ltd.
Veenhof, K. R. (2010). Ancient Assur: The City, its Traders, and its
Commercial Network, Journal of the Economic and Social History of the
Orient, 53(1/2): 3982.
Walker, Paul (2015). Contracts, Entrepreneurs, Market Creation and
Judgement: The Contemporary Mainstream Theory of the Firm in Per-
spective, Journal of Economic Surveys, 29(2) April: 317-38.
Walker, Paul (2016). The Theory of the Firm: An overview of the eco-
nomic mainstream, London: Routledge.
Whittaker, Edmund (1940). A History of Economic Ideas, New York:
Longmans, Green and Co.

70
 Williamson, Oliver E. (1964). The Economics of Discretionary Behavior:
Managerial Objectives in a Theory of the Firm, Englewoods Cliffs, N.J.:
Prentice-Hall.
Williamson, Oliver E. (1970). Corporate Control and Business Behavior,
Englewoods Cliffs, N.J.: Prentice-Hall.
Williamson, Oliver E. (1971). The Vertical Integration of Production:
Market Failure Considerations, American Economic Review, 61(2) May:
112-23.
Williamson, Oliver E. (1973). Markets and Hierarchies: Some Elementary
Considerations, American Economic Review, 63(2) May: 316-25.
Williamson, Oliver E. (1975). Markets and Hierarchies: Analysis and
Antitrust Implications, New York: The Free Press.
Williamson, Oliver E. (1996a). The Mechanisms of Governance, Oxford:
Oxford University Press.
Wolfe, J. N. (1954). The Representative Firm, Economic Journal, 64
(254) June: 337-49.

71