Anarchist-Literature/books/ACriticalExaminationOfTheEconomyOfScale/Chapter2.txt

Chapter Two

A Literature Survey on Economy of Scale
This chapter must begin with a caveat. Most of the empirical data cited below on the
ideal size for economy of scale reflect the comparative performance of plants and firms in
the existing economy, with the given structure of costs and returns. They ignore the
extent to which the existing economic environment is itself the product of state subsidies
and other interventions. The ideal size for efficiency in the existing economy refers to the
size needed for maximizing profit given subsidized inputs, and given protected monopoly
prices for outputs. The optimally sized firm, in other words, is optimally sized for
maximizing profits in the existing economic environment.
This is indicated, in most cases, by the very methods used to determine the ideal size
for economy of scale. According to F.M. Scherer,1 the methods used to determine
minimum efficient scale (MES) are the following:
1) Analyzing profitability as a function of size. This is problematic because it is hard
to distinguish profitability resulting from internal efficiency from profitability resulting
from monopoly or monopsony power. For example, a Johnson administration study
found the average rate of profit to be 50% higher in concentrated industries.2 Even in
recessions, losses from the late fifties through the early seventies were relatively rare
among the largest corporations. Only one of the top 200 industrials operated at a loss in
the recession of 1957; and only seven and 34 of the Fortune 500 lost money, respectively,
in the recessions of 1964 and 1970.3 One of the forms taken by oligopoly market power
is administered pricing: in the 1960s, for example, General Motors targeted its prices to
provide a 15-20% return after taxes, with costs estimated on the assumption that plants
operated at only 60-70% capacity. And U.S. Steel, likewise, set prices to allow for a
profit even when operating only two days a week. Bethlehem Steel's Chairman
complained in 1971 that the company had to operate at 70% capacity to make a profit,
compared to only 50% in 1966. By the way, this complicates the engineering approach
described below, since that approach estimates peak efficiency on the assumption that the
different size plants being compared operate at 100% of capacity. The comparative
"efficiency" estimates would differ somewhat if it were taken into account that the
smaller plant can operate at full capacity, while the larger one cannot.4 An FTC study
cited by the Nader Group estimated that oligopoly markup amounted to 25% of existing
prices, where the four largest firms controlled 40% or more of an industry's sales.5 So it's

1

F.M. Scherer and David Ross, Industrial Market Structure and Economic Performance. 3rd ed (Boston:
Houghton Mifflin Company, 1990) pp. 111-15.
2
Barry Stein, Size, Efficiency, and Community Enterprise, p. 54
3
Ibid., p. 55.
4
Ibid., p. 56.
5
Mark J. Green, with Beverly C. Moore, Jr., and Bruce Wasserstein, The Closed Enterprise System: Ralph
Nader's Study Group Report on Antitrust Enforcement (New York: Grossman Publishers, 1972), p. 14.

hard to control for the extent to which internal inefficiency costs of large scale are offset
by the increased market power also resulting from large size.
2) Statistical cost analysis, relating costs to volume of output. This method takes into
account such complex variables as capacity utilization, age of capital stock, etc. The
sheer amount of numbers crunching involved makes this approach quite intensive. The
results are also potentially misleading, because detailed cost data are available
disproportionately from regulated monopolies, whose rates are determined by a cost-plus
markup.
3) The "survivor test," associated in particular with George Stigler. "...[F]irm or
plant sizes that survive and contribute increasing fractions of an industry's output over
time are assumed to be efficient; those that supply a declining share of output are deemed
too large or too small." This approach measures "efficiency" in terms of the ability to
thrive under a given set of conditions; it does not distinguish environmental conditions
resulting from monopoly power or discriminatory legislation from others.
4) The engineering approach, based on engineers' technical knowledge of "alternative
equipment and plant designs and the associated investment and operating costs," relies
heavily on a complicated and labor-intensive series of interviews and questionnaires.
So when empirical studies of economy of scale find that the dominant plant or firm is
far larger than the ideal for maximum efficiency, it is something of an a fortiori
argument: the dominant plant or firm size is above the maximum size for ideal efficiency
even in an economy where subsidies make large size artificially profitable, and even
where cartelization enables large firms to escape many of the competitive penalties for
their large size. So even the "ideal" size for plant or firm, as determined by the empirical
studies cited below, is itself artificially large; in an economy without government
subsidies and protections, the most efficient firm or plant would be considerably smaller
even than what is described below by Walter Adams or Joe S. Bain.
A. Economies of Firm Size.
Assessments of economy of scale must also distinguish between economies of plant
size and economies of firm size. Economies of plant size result from purely technical
considerations; as Barry Stein put it,
some of the factors required for production are "lumpier" (that is, less divisible) than
others. In principle, capital can be subdivided as finely as desired, but the same
cannot be said for tools or people. In consequence, these resources can only be used
efficiently when the scale of activity is large enough to employ them fully.6

6

Size, Efficiency, and Community Enterprise, p. 1.

If the smallest available widget producing machine costs $100,000 and turns out a
thousand widgets a day at full capacity, a small firm cannot spend $10,000 for a machine
to produce a hundred widgets a day. And if the widget machine must be used along with
other machines of different capacities, in order to minimize unit costs it is necessary to
purchase the proper ratios of different kinds of machinery, and to maintain sufficient
output that no individual machine has idle capacity.
Plant economies also reflect, to a lesser extent, the geometries of building
construction. The larger the building, within reasonable limits, the less the cost of
materials compared to the usable volume.
...the volume or capacity of physical objects (containers, buildings, vehicles) increases
with the third power of length or radius, and thus faster than the surface area, which
only increases as the second power. Since the costs associated with material needs
and construction tend to be related to the surface area, large units have greater
capacity or volume per unit cost.7
The "within reasonable limits" qualifier is added because, eventually, when a building
gets large enough, the space required for support infrastructure will grow faster than the
space available for productive use; Leopold Kohr gave the example of a skyscraper, in
which the taller the building the larger the percentage of floorspace on each story devoted
to unproductive space (elevator shafts, heating ducts, load-bearing structures, etc.).8
Economies of plant size are real, at least, however much controversy there may be as
to the precise point at which they level off. On the alleged economy of firm size, there is
less agreement.
It rests upon alleged efficiencies of management rather than technology. Efficiency, it
is said, is enhanced by spreading administrative expertise and expenses over
multiplant operations; by eliminating duplication of officials, services, and records
systems; by providing sophisticated statistical, research, and other staff services that
smaller firms cannot afford; by circumventing "transaction costs" by performing
support activities in house rather than purchasing them from outsiders; by obtaining
credit on more advantageous terms; by attracting more competent executives and
mounting more effective marketing campaigns; and so forth.9
The savings from spreading administrative costs over more than one plant are

7

Ibid., p. 2.
The Overdeveloped Nations, pp. 106-07.
9
Walter Adams and James W. Brock. The Bigness Complex: Industry, Labor and Government in the
American Economy. 2nd ed. (Stanford, Cal.: Stanford University Press, 2004)., pp. 30-31.
8

doubtless true, ceteris paribus. But as usual, ceteris in this case is not paribus. Whatever
savings result from administrative rationalization are probably offset, or more than offset,
by the bureaucratic inefficiencies resulting from added layers of administration, and from
increased Hayekian problems of aggregating distributed information.
The advantages resulting from superior bargaining power in the credit market, from
the power of a large-scale buyer to negotiate lower prices, and so forth, are also real. But
as Adams and Brock point out, such advantages of superior bargaining power are not real
operating efficiencies: unlike internal efficiencies, which result in real cost savings
overall, they are zero-sum transactions that merely shift a portion of costs to those with
less bargaining power.10 Barry Stein made same distinction in Size, Efficiency, and
Community Enterprise:
It is necessary... to distinguish between true social efficiency and simple power.
Efficiency has been defined... as a measure of the extent to which social and individual needs
are met for a given set of available resources. But large and well-established firms also have
power, the ability to control the environment toward their own ends. To a considerable
degree, organizations with power can be less efficient; at least, they can change the nature of
the contest so that others, even if more truly efficient, are less able to compete. Thus, many
of the gross measures of the relative efficiency of firms of different scale (such as overall
profit, sales growth, or survival), may be indicative of the power of size, rather than the
economic effect of scale....
It is a well established fact that large firms have a degree of power, simply as a result of
size, which is ordinarily used to acquire more or surer profits on operations over time, to
raise barriers against the entry of new competition, to gain access to funds, or to control a
share of the market greater than may be accounted for by conventional models of a fully
competitive economy....
And, in fact, there is evidence that, in concentrated industries, profits are higher than
they would be otherwise.11

It is quite plausible that most of the "efficiencies" of the large firm fall into this
category: the ability to excercise power outside the firm, especially insofar as large size
creates a power center for the control of state-granted privileges like "intellectual
property," or creates a financial base to lobby the state for special privileges.
Joseph Schumpeter suggested, as we saw in the previous chapter, that large firm size,
by insulating the corporation from risk, put it in a superior position to undertake
expensive and long-term innovations. But as we shall see below, in the real world the
large firm is far less innovative.
William G. Roy. Socializing Capital: The Rise of the Large Industrial Corporation

10
11

Adams and Brock, 2nd ed., p. 31.
Barry Stein, Size, Efficiency, and Community Enterprise, pp. 52-54.

in America (Princeton, N.J.: Princeton University Press, 1997). Oliver Zunz. Making
America Corporate, 1870-1920 (Chicago: University of Chicago Press, 1990).
Economies of firm size are relatively insignificant compared to economies of plant
size. Honda's main operating plants in Japan are about three times the average plant size
for the American Big Three. But Honda as a firm, with only two major plants in Japan, is
far smaller than either GM (28 plants) or Ford (23 plants). Not only does GM's larger
size fail to provide any cost efficiencies compared to Honda; it is riddled with
inefficiency. GM is significantly less efficient than either Ford or Chrysler, while all
three American producers are far (24-38%) less efficient than Honda's North American
operations.12
A 1956 study by Joe S. Bain found that the efficiencies of multiplant firms were
"either negligible or totally absent" in six of twenty industries. In another six, unit cost
economies accruing to multiplant firms were small but measurable, ranging from "slight"
in cigarettes to 2-5% in steel. In the remaining eight, no estimates of multiplant firms'
advantages were available.13
The alleged efficiencies of large firm size are even more dubious in the case of the
conglomerate, a steroidal parody of the M-form corporation. In the heyday of the
conglomerate, its advocates saw it as a sort of private planned economy:
Defenders of conglomerate bigness argued that the capital markets are an inferior
instrument for optimizing society's investment decisions and for planning its output
patterns of goods and services. Decentralized decision making by myriads of
borrowers, lenders, and investors, they claimed, sufferes from inadequate information
and unnecessary and wasteful "transactions costs." Allocation of investment, they
said, would be more effectively achieved through centralized generation, control, and
allocation of capital within the giant conglomerate firm.
Thus, conglomerate giants were portrayed as superior to decentralized capital
markets in allocating financial funds among alternative uses and in ensuring that these
funds would flow to their most socially desirable uses.14
As we shall see in Part III of this book, those at the top of the pyramid even within
the large unitary firm fall victim to "MBA disease," operating on the basis of finance and
marketing considerations while viewing the production process largely as a black box.
They strip departments of productive assets, defer maintenance, and the like, all in order
to inflate apparent short-term profits. Imagine these very same types attempting to make
rational decisions on the shuffling of finance between divisions of a conglomerate:

12

Adams and Brock, 2nd ed., pp. 31-32.
Joe S. Bain. Barriers to New Competition: Their Character and Consequences in Manufacturing
Industries. Third printing (Cambridge, MA: Harvard University Press, 1965), pp. 86-87.
14
Adams and Brock, 2nd. ed., p. 71.
13

holding the ultimate power not only of the purse, but of hiring and firing, over those in
charge of the production process, even though they themselves understand the
conglomerate firm only as a glorified investment portfolio. If the MBAs in charge of
traditional large firms are prone to milk them for short-term profit, imagine the
opportunities for those in charge of a conglomerate to treat entire divisions as cash cows
for asset stripping! And as Duggar pointed out, the management of the old conglomerates
engaged in just that kind of cross-subsidization.15
In the end, the conglomerate movement was largely a failure--even within a state
capitalist economy where the rules are stacked in favor of bigness. The conglomerates
were even less efficient than the general run of subsidized and protected giant
corporations. The conglomerate fad of the '60s and '70s had passed by the '80s, and
existing conglomerates subsequently were largely divested of their non-core holdings.16
In an economy where the average large firm survived entirely through government
welfare, for the conglomerate corporate welfare was not enough; it was in need of a heartlung machine.
B. Economies of Plant Size.
Cross-industry studies have found little evidence to back up the alleged efficiencies of
large plant size. For example, a study by T.R. Saving covering the 1947-54 period found
that in 64 of 91 manufacturing industries, the minimum efficient plant created 1% or less
of industry value added.17
A 1956 study by Joe S. Bain found that in eleven of twenty industries, the plants with
the lowest unit production costs operated on average with an output of 2.5% or less of
total national sales (with the individual outputs ranging from 0.02% to 2.5%); in fifteen
industries, less than 7.5%; and in seventeen out of twenty, less than 10%.18
A 1975 study of 12 industries in seven industrialized nations, based on the
engineering survey method, found that--with the exception of the refrigerator-freezer
industry--the least-cost plant sizes were "quite small relative to the national market." The
same study found a remarkably shallow cost curve for plants below optimal size: in half
of the industries surveyed, a plant operating at one-third the optimal output suffered an
increase in unit costs of under 5%.19
According to F. M. Scherer, the statistical cost analysis method of investigation

15

William M. Dugger, Corporate Hegemony, pp. 34-35.
Adams and Brock, 2nd ed., p. 72.
17
Scherer and Ross, p. 114.
18
Joe S. Bain. Barriers to New Competition: Their Character and Consequences in Manufacturing
Industries. Third printing (Cambridge, MA: Harvard University Press, 1965), pp. 72-73.
19
Scherer and Ross, pp. 114-15.
16

typically shows that, "[w]ith few exceptions, the minimum efficient scale revealed in
studies of U.S. manufacturing industries has been small relative to industry size." The
most common finding has been "distinct economies of scale at relatively small plant
sizes, a range of intermediate sizes over which unit costs did not differ appreciably, and
(in a minority of cases) diseconomies of scale for very large plants."20
In the steel industry, for example, minimills have been cleaning the clocks of the old
steel giants. According to Adams and Brock, minimills operating at infinitesimal
fractions of the output of U.S. Steel and Bethlehem Steel had by 1998 achieved a 45%
share of the U.S. market. They used electric furnaces to process scrap metal, and oriented
their output toward local markets. Minimills produced wire rod and cold-rolled steel
sheets 28% and 29% cheaper, respectively, than U.S. Steel. A minimill could produce
steel bars with only thirty employees on average, compared to 130 even in a single plant
of U.S. Steel.21
C. The Comparative Significance of Scale Economies and Organizational
Efficiency.
Barry Stein suggested that whatever the increased costs resulting from belowoptimum-size production facilities, they pale in comparison to the variations in cost
resulting from greater or lesser efficiencies within facilities of any given size.
The normal neoclassical approach, according to Stein, is to treat the firm's internal
functioning as a "black box":
One of the characteristics of classical economists' view of business organization is a
tendency to view firms as entities operating at near-optimal efficiency within
whatever constraints size, industry, and the environment impose. The treatment of
economies of scale and of other questions related to efficiency thus have generally
focused on the allocative aspects; that is, the extent to which resources or factors of
production have been optimally distributed to firms and establishments within the
economic system. Within that framework, firms are assumed to operate on the
frontier of their specific production functions.22
As an example, he quoted Robert Dorfman:
businessmen determine the cost of attaining any [desired] output by choosing the
combination of factors [labor, materials, or capital] with which to produce that
output.... The production function incorporates all the technical data about

20

Ibid., pp. 112-13.
Adams and Brock, pp. 36-37; see also Murray Bookchin, Post-Scarcity Anarchism (Berleley, Ca.: The
Ramparts Press, 1971), pp. 108-110.
22
Stein, p. 27.
21

production; it shows the greatest amount of output that can be obtained by the use of
every possible combination of input quantities.23
Stein continued:
If this describes the actual situation, then questions of allocation become critical.
However, there is very good reason to believe that industrial firms operate not on or
near their production frontier, but well inside it, and, correspondingly, measures
assuming the ideal case are likely to be misleading.
There are two points to be made. The lesser is related to utilization of capacity. It
is clear that what might be theoretically true with regard to the efficiency of a plant
that is operating at design capacity, with all fixed assets properly contributing their
share to output, will hardly be true when some fraction of the assets are, in effect,
idle....
But excess capacity is the minor point. More important is the fact that while
economists focus on problems of allocation, businessmen have always spent more
time on problems of internal efficiency, in the obvious belief that it can be
increased....24
By way of contrast to the neoclassical assumption that the production elves were
magically running things in an optimal manner inside the black box, Stein brought in
Harvey Leibenstein's key concept of "x-efficiency" (about which more in Chapter Seven).
This was anticipated in the 1950s by the so-called "Solow residual," which showed that
some 80% of economic growth could not be explained by the accumulation of labor or
capital stocks.25
Leibenstein suggested "an approach to the theory of the firm that does not depend on
the assumption of cost-minimization by all firms."
The level of unit cost depends in some measure on the degree of x-efficiency, which
in turn depends on the degree of competitive pressure, as well as on other
motivational factors. The responses to such pressures, whether in the nature of effort,
search, or the utilization of new information, is a significant part of the residual
[unexplained increase] in economic growth.26

23

Robert Dorfman, Prices and Markets (Englewood Cliffs, N.J.: Prentice-Hall, 1967), pp. 67-68, in Stein
pp. 27-28; Stein commented, in fn1 p. 98: "Of course, no one assumes that the production function is
either known with precision or ideally followed, but the assumption is that businesses, by and large, operate
sufficiently close to their production frontier so that attention can shift to the exogenous variables
influencing the firm."
24
Stein, p. 28.
25
Robert U. Eyres, "Lecture 5: Economic Growth (and Cheap Oil)," p. 4.
26
Harvey Leibenstein, "Allocative Efficiency vs. X-Efficiency," American Economic Review (June 1966),
pp. 412-13.

...[F]irms and economies do not operate on an outer-bound production possibility
surface consistent with their resources. Rather they actually work on a production
surface that is well within that outer bound. This means that for a variety of reasons
people and organizations normally work neither as hard nor as effectively as they
could.27
As Stein commented,
In other words, the usual assumptions about the efficient use of resources within a
firm are simply not true. What is more, the extent of those inefficiencies is not small.
There is significant opportunity for firms to increase their output for any given array
of resources or, alternatively, to reduce their use of resources for any given level of
output....
It is at least arguable, and perhaps should be apparent, that there can be no perfect
utilization of available resources. Theories concerning the firm that assume that any
single specific parameter is responsible for observed behavior are positing an overly
simplistic assumption. Corporations, despite the legal fiction of personhood, do not
act uniquely as entities, but as a composite of human subsystems, each of which is
attempting to satisfy conflicting and comlex needs, some personal (keeping one's job,
doing more satisfying work, earning more money) and some organizational
(exceeding profit goals, developing new products, maintaining the corporate share of
market).... What is clear... is that the larger the firm and the more complex the
subsystem of interactions, the more the possibility that alternative solutions exist and
the likelihood that efficiency, however measured, can be improved.
Support for these views of potential loss of efficiency can also be gained from
simple observation of the extent to which companies "discover" during lean times that
they are perfectly capable of operating at the same level with substantially fewer
employees or, in some cases, facilities....
The significance of all this is simply that computations and estimates of
economies of scale, from whatever source, can be misleading or downright inaccurate,
since they typically assume that firms and plants operate efficiently within their
constraints. This is generally not the case; what actually is being measured, if
anything, is the relative productivity of various entities, all of which are capable of
increasing their efficiency by amounts and in ways that are uniquely related to that
entity. In addition, such savings as might in fact be available because of the real
economies of scale (ranging up to perhaps 20 or 25 percent for a substantial change in
size) are capable of being overhwelmed by the continuing increases due to
improvement in "x-efficiency."

27

Ibid., p. 413.

It may be that these inefficiencies help explain the great lack of consistency in the
many studies of economies of scale....28
D. Increased Distribution Costs
It's also important to remember that whatever reduction in unit production cost results
from internal economies of large-scale production is to some extent offset by the
diseconomies of large-scale distribution.
...[U]nit costs of production, which up to some point decrease with scale, must be
compared to unit costs of distribution, which tend to increase (other things being
equal) with the size of the area served.29
As Ralph Borsodi observed years ago, the larger the plant needed to achieve
economies of scale in production, the larger the market area it serves; hence, the longer
the distances over which the product must be distributed. His observation, stated simply
as Borsodi's Law: as production costs fall, distribution costs rise.
In most cases, the increased cost of distribution exceeds the reduced cost of
production at a level of output far lower than would be ideal for maximizing purely
internal economies of scale. The increase in unit production cost, even for significant
reductions in size below the optimum for productive economy of scale, is quite modest:
The 1975 study referenced earlier by Bain, surveying twelve industries in seven
industrialized nations, found a remarkably shallow cost curve for plants below optimal
size: in half the industries surveyed, a plant with output at a third of the optimal level
suffered unit cost increases of less than 5%.30 Compare this to the reductions in
distribution cost for a market area reduced by two-thirds.
Distribution costs are increased still further by the fact that larger-scale production
and greater levels of capital intensiveness increase the unit costs resulting from idle
capacity, and thereby greatly increase dependence on high-pressure, "push" forms of
marketing. Borsodi wrote extensively on the phenomenon--the increased reliance on
brand differentiation, packaging, and advertising--in The Distribution Age. That entire
book was an elaboration of the fact that, as he stated in the Preface, production costs fell
by perhaps a fifth between 1870 and 1920, while at the same time the cost of marketing
and distribution had nearly tripled (we've already examined the marketing aspect of the
phenomenon in Chapter One).31 "[E]very part of our economic structure," he wrote, was
"being strained by the strenuous effort to market profitably what modern industry can

28

Stein, pp. 28-30.
Ibid., p. 65.
30
Bain, pp. 114-15.
31
Ralph Borsodi, The Distribution Age (New York and London: D. Appleton and Company, 1929), p. v.

29

produce."32
Kirkpatrick Sale described the even greater relative costs of marketing gimmickry, in
addition to physical distribution costs, in the 1970s:
...the complications [in cost-differential between large- and small-scale
production] rose not in the plants but far downstream from the lathes and belts
and assembly lines. First, distribution. The more goods that are produced, the
wider the market area must be, hence the more expensive the costs of
distribution... throughout that area; it is now an accepted standard in the U.S. that,
particularly for consumer goods, the unit costs of distribution will be higher than
those of production, and they will increase as the price of gasoline goes up.
Second, advertising. Mass production naturally necessitates sufficient advertising
to create a mass market, and the more extensive it is the more expensive--which is
why name-brand items are always more expensive than generics.
I can't help but interject here on another reason that name-brand items are more
expensive. The sale of generic products by the manufacturers of name-brand goods is a
form of dumping, directly analogous to the dumping of surplus product overseas by
domestic manufacturers protected behind tariff walls. The overbuilt manufacturing
corporation must minimize idle capacity to keep unit costs down, but cannot dispose of
its full product at cartel prices when running at full capacity. The solution is to sell as
much of the product as possible at oligopoly markup, and then dispose of the rest at
whatever price it will bring--whether by foreign dumping or by repackaging as generics.
It's otherwise known as price discrimination, the classic monopolist's strategy of setting
different prices for the same product based on ability to pay. Anyway, Sale continues:
(The high cost of advertising also tends to keep smaller and cheaper firms out of a
market--creating an "entry barrier," in economic terms--thus reducing the competition
that might lead to lower consumer prices.) Finally, promotion and packaging. In
markets that are saturated, and where Brand A is not especially different from Brand
B, it is necessary to find gimmicks that make a product stand out--bigger boxes,
added partitions, toys, contests--and lead to added costs.33
As with "x-efficiency" in our discussion above, the costs of the "push" distribution
made necessary by large scale probably outweigh any savings in unit cost resulting from
economy of scale iself. As we already saw in Chapter One, the shift from bulk
commodity sales to brand specification and pre-packaging resulted in a price increase of
some 300% for essentially the same goods. Barry Stein compared the price of Consumer
Value Stores' private brand to the price CVS charged for the nationally branded version of
the same goods. Typical was the CVS store brand of multi-vitamin, which sold for $1.39

32
33

Ibid., p. 4.
Kirkpatrick Sale, Human Scale (New York: Coward, McCann, & Geoghegan, 1980), pp. 315-16.

per 100 compared to $2.13 for 100 1-a-Day vitamins, and $0.63 per 100 buffered aspirin
compared to $1.00 for 100 Bufferin. And as Stein points out, CVS being a discount
store, the price it charged for national brands was itself considerably lower than the
manufacturers' suggested list prices. The latter was $2.98 and $1.67 for 1-a-Day vitamins
and Bufferin, respectively.
.... the CVS products are all attractively packaged and in no obvious way inferior in
appearance or presentation to the national brands (therefore, no great savings are
being made by cheaper packaging).... [And] it is likely, from CVS' own description
of its program, that these products, by and large, are being manufactured on order by
relatively small firms (such as manufacturing chemists). If this is not the case and
they are, in fact, being produced by the same type of large firm as the national
products, one can still clearly conclude that, at least for products of this class,
whatever economies of scale exist in production, they are being dwarfed by
diseconomies in advertising, promotion, and physical distribution.34
In other words, the alleged economies of large-scale production result in such
expensive, high-capacity facilities that large corporations are required to take heroic
measures--often more expensive than the supposed unit cost savings from large scale--to
move enough of their product to keep the plants running at full capacity.
Increased unit costs from idle capacity, given the high overhead of large-scale
production, are the chief motive behind the push distribution model. Even so, the
restrained competition of an oligopoly market limits the competitive disadvantage
resulting from idle capacity--so long as the leading firms in an industry are running at
roughly comparable percentages of capacity, and can pass their overhead costs onto the
customer. The oligopoly mark-up included in consumer price reflects the high costs of
excess capacity.
It is difficult to estimate how large a part of the nation's production facilities are
normally in use. One particularly able observer of economic tendencies, Colonel
Leonard P. Ayres, uses the number of blast furnaces in operation as a barometer of
business conditions. When blast furnaces are in 60 per cent. operation, conditions are
normal. When this figure is exceeded, productive industry is experiencing a period of
good times; and when it falls below that figure, it is in for a period of hard times.
It is obvious, if 60 per cent. represents normality, that consumers of such a basic
commodity as pig iron must pay dividends upon an investment capable of producing
two-thirds more pig iron than the country uses in normal times.
Borsodi also found that flour mills, steel plants, shoe factories, copper smelters, lumber
mills, automobiles, and rayon manufacturers were running at similar or lower percentages

34

Stein, pp. 67-68.

of total capacity.35 Either way, it is the consumer who pays for overaccumulation: both
for the brand name markup and marketing cost of distributing overproduced goods when
industry runs at full capacity, and for the high overhead when the firms in an oligopoly
market all run at low capacity and pass their unit costs on through administered pricing.
Furthermore, Borsodi's law does not apply merely to the relative efficiencies of large
versus small factories; it also applies to the relative efficiencies of factory versus home
production. Borsodi argued that for most light goods like food, textiles, and furniture, the
overall costs were actually lower to manufacture them in one's own home. The reasons
were the same ones put forward by Kropotkin and Mumford, with which we will deal
more closely under our discussion of neotechnic in Part Four: the electric motor put
small-scale production machinery in the home on the same footing as large machinery in
the factory. Although economies of scale in production are available, on an ever
diminishing level, up to a considerable scale of production, the majority of economies of
machine production are captured with the bare adoption of the machinery itself, even with
household electrical machinery. After that, the production cost curve is very shallow,
while the distribution cost curve is steep.
Borsodi's first study of the economics of home manufacture involved the home-grown
tomatoes that his wife canned. Expressing some doubts in response to Mrs. Borsodi's
confidence that it "paid" to do it, he systematically examined all the costs going into the
tomatoes, including the market value of the labor they put into growing them and canning
them, the cost of the household electricity used, and every other cost they could think of.
Even with all these things factored in, Bordodi still found the home product cost 20-30%
less than the canned tomatoes at the market. The reason? The home product was
produced at the point of consumption, and had zero distribution cost. The admittedly (if
modest) unit cost savings from large-scale machinery were not enough to offset the
enormous cost of distribution and marketing.36
Borsodi went on to experiment with home production of clothing with loom and
sewing machine, and with building furniture in the home workshop.
I discovered that more than two-thirds of the things which the average family now buys
could be produced more economically at home than they could be brought factory made;
--that the average man and woman could earn more by producing at home than by
working for money in an office or factory and that, therefore, the less time they spent
working away from home and the more time they spent working at home, the better off they
would be;
--finally, that the home itself was still capable of being made into a productive and

35

The Distribution Age, pp. 42-43.
Ralph Borsodi, Flight From the City: An Experiment in Creative Living on the Land (New York,
Evanston, San Francisco, London: Harper & Row, 1933, 1972), pp. 10-15.
36

creative institution and that an investment in a homestead equipped with efficient domestic
machinery would yield larger returns per dollar of investment than investments in insurance,
in mortgages, in stocks and bonds....
These discoveries led to our experimenting year after year with domestic appliances and
machines. We began to experiment with the problem of bringing back into the house, and
thus under our own direct control, the various machines which the textile-mill, the cannery
and packing house, the flour-mill, the clothing and garment factory, had taken over from the
home during the past two hundred years....
In the main the economies of factory production, which are so obvious and which have
led economists so far astray, consist of three things: (1) quantity buying of materials and
supplies; (2) the division of labor with each worker in industry confined to the performance
of a single operation; and (3) the use of power to eliminate labor and permit the operation of
automatic machinery. Of these, the use of power is unquestionably the most important.
today, however, power is something which the home can use to reduce costs of production
just as well as can the factory. The situation which prevailed in the days when water power
and steam-engines furnished the only forms of power is at an end. As long as the only
available form of power was centralized power, the transfer of machinery and production
from the home and the individual, to the factory and the group, was inevitable. But with the
development of the gas-engine and the electric motor, power became available in
decentralized forms. The home, so far as power was concerned, had been put in position to
compete with the factory.
With this advantage of the factory nullified, its other advantages are in themselves
insufficient to offset the burden of distribution costs on most products....
The average factory, no doubt, does produce food and clothing cheaper than we produce
them even with our power-driven machinery on the Borsodi homestead. But factory costs,
because of the problem of distribution, are only first costs. They cannot, therefore, be
compared with home costs, which are final costs.37

Even the internal economies of the factory, it should be added, were balanced by other
internal diseconomies: the overhead costs of superintendence and administration, and the
dividends and interest on capital.38 Since first reading Borsodi's account I have
encountered arguments that his experience was misleading or atypical, given that he was
a natural polymath and therefore perhaps a quicker study than most, and therefore failed
to include learning time in his estimate of costs. Still, Borsodi's case studies are a useful
counter to claims that economies of scale are inherent in the greater technical efficiency
of large-scale machinery. And the savings in unit cost Borsodi demonstrated, if true,
would be sufficient to compensate a fair amount of learning time.
The internal economies resulting from division of labor, specifically (which Borsodi
acknowledged in the quote above), are also greatly exaggerated. Stephen Marglin argued

37
38

Ibid., pp. 17-19.
Ralph Borsodi, This Ugly Civilization (Philadelphia: Porcupine Press, 1929, 1975), pp. 34, 37.

that the economies in question resulted, not from division of labor as such, but from the
separation and sequencing of tasks. Nearly the same economies could be achieved by a
single workman or group of workmen in a small shop, by such separation and sequencing.
To illustrate, he took Adam Smith's famous example of the pin factory and stood it on its
head. An individual cottage workman, instead of painstakingly making one pin at a time,
might draw out and straighten the wire for an entire run of production, then cut all the
wire, then sharpen it all, etc., dividing the total operation into the very same subtasks as in
Smith's pin factory.39
One alleged reason for economies of large-scale production is that large scale permits
ever more specialized production machinery. But as Adam Smith pointed out, the
profitability of division of labor is determined by market size; and as we shall see in
Chapter Thirteen, when transportation ceases to be subsidized, so that the savings from
maximal automation with highly specialized machines are offset by the true cost of longdistance distribution, the spurious economies of excessive division of labor disappear.
When all costs are taken into account, it is more efficient overall to produce most goods
in short production runs, for local markets, on general purpose machinery. Without
artificially large market areas resulting from artificially cheap distribution, the demand in
the smaller market areas would be insufficient in most cases to operate expensive
specialized machinery at full capacity. Unit costs would be lower with frequent changes
of product line on the same general-purpose machinery.
And even in the case of the largest existing corporations under state capitalism, with
artificially large market areas resulting from subsidized transportation, their attachment to
the largest-scale machinery is often misguided. While individual machines may be
"super-efficient" from the standpoint of minimizing unit costs of that particular stage of
production, they are often quite disruptive and inefficient from the standpoint of the
overall flow of production. Their adoption is typically associated with the "batch-andqueue" operation of American Sloanist industry (about which more in Chapter Eight),
which (as the authors of Natural Capitalism put it) optimizes the efficiency of individual
steps in the production process at the expense of pessimizing the overall flow of
production. Their excessive "efficiency," from the perspective of the overall production
process, means that they generate excess inventories and buffer stocks that raise costs and
disrupt flow. On the other hand, a smaller and less "efficient" machine that is compatible
with the other stages of production may result in improved flow and greatly reduced
overall cost, despite the higher unit costs of that particular stage. Consider the case of
Pratt & Whitney:
Traditional substitutions of complex machines for people can backfire, as Pratt &
Whitney discovered. The world's largest maker of jet engines for aircraft had paid $80
million for a "monument"--state-of-the-art German robotic grinders to make turbine blades.
The grinders were wonderfully fast, but their complex computer controls required about as
many technicians as the old manual production system had required machinists. Moreover,

39

Stephen Marglin, "What Do Bosses Do?"

the fast grinders required supporting processes that were costly and polluting. Since the fast
grinders were meant to produce big, uniform batches of product, but Pratt & Whitney needed
agile production of small, diverse batches, the twelve fancy grinders were replaced with eight
simple ones costing one-fourth as much. Grinding time increased from 3 to 75 minutes, but
the throughput time for the entire process decreased from 10 days to 75 minutes because the
nasty supporting processes were eliminated. Viewed from the whole-system perspective of
the complete production process, not just the grinding step, the big machines had been so fast
that they slowed down the process too much, and so automated that they required too many
workers. The revised production system, using a high-wage traditional workforce and simple
machines, produced $1 billion of annual value in a single room easily surveyable from a
doorway. It cost half as much, worked 100 times faster, cut changeover time from 8 hours to
100 seconds, and would have repaid its conversion costs in a year even if the sophisticated
grinders were simply scrapped.

When entire processes are taken into account, "excessive scale or speed at any stage of
production turns the smooth flow of materials into turbulent eddies and undertows that
suck down earnings and submerge entire industries."40
Another example comes from the cola industry, where the most "efficient" large-scale
machine creates enormous batches that are out of scale with the distribution system, and
result in higher unit costs overall than would modest-sized local machines that could
immediately scale production to demand-pull. The reason is the excess inventories that
glut the system, and the "pervasive costs and losses of handling, transport, and storage
between all the elephantine parts of the production process."41
Of course the authors of Natural Capitalism exaggerate the market penalties of
inefficiency in such cases. The pressure to remedy such over-specialization and overautomation is hardly overwhelming in most cases. Large industry often operates with
forms of production that are capital-intensive and specialized far beyond the point of
increasing costs, simply because all the firms in an industry share the same institutional
culture and consequently need not be overly concerned with any competitive pressure to
minimize costs. Without cartelized markets and subsidies, the issue of jet engine
manufacturing technology would probably be moot; in an unregulated market, with
unimpaired competition and fully internalized costs, there likely wouldn't be any jet
engine manufacturers in the first place.
E. The Link Between Size and Innovation.
The superior innovativeness of the large corporation, such a sacred cow for
Schumpeter and Galbraith, is also questionable at best.

40

Paul Hawken Amory Lovins, and L. Hunter Lovins. Natural Capitalism: Creating the Next Industrial
Revolution (Boston, New York, London: Little, Brown, and Company, 1999), pp. 128-29.
41
Ibid., p. 129.

T.K. Quinn, a former Vice President of GE (writing in the heyday of managerialist
liberalism), viewed the oligopoly firm's role in the innovation process as largely parasitic:
I know of no original product invention, not even electric shavers or heating pads, made
by any of the giant laboratories or corporations, with the exception of the household garbage
grinder.... The record of the giants is one of moving in, buying out, and absorbing the
smaller creators.42

Paul Baran and Paul Sweezy, in Monopoly Capital, commented on Quinn's rhetorical
bombshell:
...the corporation knows how to use for its own ends the very weaknesses of the small
enterprise which it has outgrown. When a new industry or field of operation is being opened
up, the big corporation tends to hold back deliberately and to allow individual entrepreneurs
or small businesses to do the vital pioneering work. Many fail and drop out of the picture,
but those which succeed trace out the most promising lines of development for the future.43

John Jewkes, surveying the period from 1900 to 1958, found that comparatively few
of the major inventions of the 20th century came from large organizations. Out of 61 of
the most important inventions, 33 were individual efforts, seven were of mixed or unclear
origins, and only 21 the product of corporate research labs. In even the latter group, five
of the inventions came from smaller corporations. And the inventions coming out of the
large corporations often involved research teams that were quite small,44 what today
might be called "skunk works." To take one example:
At a $5 billion survey comany, three of the last five new-product introductions have come
from a classic skunk works. It consists at any one time of eight to ten people, and is located
in a dingy second-floor loft six miles from the corporate headquarters. The technical genius
is a fellow whose highest degree is a high-school equivalency diploma... (although the
company has literally thousands of Ph.D. scientists and engineers on its payroll)....
The group's first product, now a $300 million per year sales item, was fully developed
(prototyped) in twenty-eight days. Last year a major corporate product bombed. A skunk
works member asked for and got permission to take two samples home and set them up in his
basement. He used one as a benchmark. He tinkered with the other for about three weeks
and corrected virtually all of its flaws (with nickel and dime items), actually improving
performance over original design specs by a factor of three. The president visited his
basement and approved design changes on the spot. The latest of the group's successes was
designed in (covert) competition with a corporate engineering "team" of almost 700 people.45

42

T.K. Quinn, Giant Business: Threat to Democracy: The Autobiography of an Insider (New York,
1953), p. 117, cited in Paul Baran and Paul Sweezy, Monopoly Capital, p. 49.
43
Ibid., p. 49.
44
John Jewkes, David Sawers, and Richard Stillerman, The Sources of Invention (London: MacMillan &
Co Ltd, 1958), pp. 72-88.
45
Tom Peters, In Search of Excellence: Lessons from America's Best-Run Companies (New York: Warner
Books, 1982), pp. 211-212.

Arnold Cooper found, likewise, that the small firm made better use of its R&D
dollars, and that its technical workers were on average more capable.46 And Jacob
Schmookler testified before Congress in 1965 that there is an inverse relationship
between firm size and productivity per research dollar:
Existing comprehensive indexes of outputs of new technical knowledge suggest that beyond
a certain not very large size, the bigger the firm, the less efficient its knowledge-producing
activities are likely to be. Evidently, as the size of the firm increases, there is a decrease per
dollar of R&D in (a) the number of patented inventions, (b) the percentage of patented
inventions used commercially, and (c) the number of significant inventions.47

A National Science Foundation study of technical innovation between 1953 and 1973
found that the smallest firms produced "about 4 times" as many major innovations per
R&D dollar as did the mid-sized firms, and 24 times as many as the largest firms.48
Adams and Brock contrast the innovativeness of the pre-WWII auto industry, with its
many modest-sized firms, with the stagnation under the Big Three during the first decades
of the postwar era.
...[W]ith the demise of the independents and the concentration of industry control in the
hands of three giant firms, the pace of product innovation slackened significantly.
Innovations like front-wheel drive, disc brakes, fuel injection, fuel-efficient subcompacts,
and utilitarian minivans languished in the hands of the Big Three.... "The major features of
today's automobiles--V-8 engines, automatic transmissions, power steering, and power
brakes--are all prewar innovations. These have been considerably improved and refined over
the past twenty-five years," [economist Lawrence J. White] concluded in 1971, "but still the
industry has been uninterested in pursuing alternatives. The suspension, ignition,
carburetion, and exhaust systems are fundamentally the same."49

Paul Goodman also viewed the automobile industry as a typical example of this
aspect of oligopoly behavior: "Three or four manufacturers control the automobile
market, competing with fixed prices and slowly spooned-out improvements."50 As just
one example, consider the way the Big Four automakers colluded to suppress
antipollution devices. They agreed that no company would announce or install any
innovation in antipollution exhaust devices without an agreement of the other three. They
exchanged patents and agreed on a formula for sharing the costs of patents acquired from
third parties.51

46

"R&D is More Efficient in Small Companies," Harvard Business Review (May-June 1964), in Barry
Stein, p. 35.
47
Quoted in Stein, p. 34.
48
Adams and Brock, The Bigness Complex. 1st edition, p. 52.
49
The Bigness Complex, 2nd ed., pp. 48-49.
50
Paul Goodman, People or Personnel, p. 58, in People or Personnel and Like a Conquered Province
(New York: Vintage Books, 1963, 1965), p. 58.
51
Mark J. Green, et al., The Closed Enterprise System, pp. 254-256.

In the computer field, Intel saw the main market for its micro-processors as giant
institutional clients, and IBM dismissed the idea of small computers for the home. The
desktop computer was created by members of the Homebrew Computer Club, who,
"playing with electronic junk..., combined Intel's microprocessor with spare parts," and
built the first cheap computers able to "run on the kitchen table."52 Apple produced its
first desktop computers for the commercial market in Steve Jobs' garage.53
Harvey Leibenstein noted that the adoption of even known technologies and best
practices--even when they are known to result in astronomical increases in productivity-occurs at a glacial pace in concentrated industries with little competitive pressure.
...there is a great deal of evidence that the delay time between invention and
innovation is often exceedingly long (sometimes more than 50 years), and the lag time
between the use of new methods in the "best practice" firms in an industry and other
firms is often a matter of years. Salter in his study on Productivity and Technical
Change... points to the following striking example: "In the United States copper
mines, electric locomotives allow a cost saving of 67 per cent yet although first used
in the mid-twenties, by 1940 less than a third of locomotives in use were electric."54
The drug industry's massive R&D spending is almost entirely directed toward gaming
the patent system, rather than genuine innovation. A majority of R&D spending goes
toward tweaking existing drugs on the verge of going generic just enough to justify a new
patent for the "me, too" version of the old cash cow, rather than to fundamentally new
drugs ("new molecular entities").55 Even when fundamentally new drugs are developed, a
majority of the total cost of is not for developing the drug itself, but for testing all the
possible variants of the drug in order to secure patent lockdown against competition.
"Quasibill," a frequent commenter at my blog with a background in engineering, is
eloquent on the subject:
What generally gets included in the accounting for research costs are some
amazing things, that I can't do justice to on a blog - I get surprised everytime I talk to
my friends in the industry about how much waste is involved - but it's all invisible to
them. It's just "how it needs to be for the FDA to keep track of everything." If you
want, I can give you some examples, but I'd rather focus on another point for now.
Namely that what big pharma is researching is cancer meds. It's not. In the rare
instances that big pharma produces and markets such medicines, it has purchased
them from small start-ups that themselves are the result normally of a university

52

Johan Soderberg, Hacking Capitalism: The Free and Open Source Software Movement (New York and
London: Routledge, 2008), p. 17.
53
Adams and Brock, 2nd edition, pp. 52-56.
54
Leibenstein, "Allocative Efficiency vs. 'X-Efficeincy,'" p. 403.
55
Ibid., pp. 57-58.

laboratory's work. When big pharma cites to billions of research costs, what it is
talking about is the process whereby they literally test millions of very closely related
compounds to find out if they have a solid therapeutic window. This type of research
is directly related to the patent system, as changing one functional group can get you
around most patents, eventually. So you like to bulk up your catalogue and patent all
closely related compounds, while choosing only the best among them, or, if you're
second to market, one that hasn't yet been patented.
This work is incredibly data intensive, and requires many Ph.D's, assistants, and
high powered computers and testing equipment to achieve. But it is hardly necessary
in the absence of a patent regime. In the absence of patents, (and of course the FDA),
you could just focus on finding a sufficient therapeutic window, and cut out the
remaining tests. It would be an issue of marginal costs to determine whether someone
would go to the effort to find a "better" therapeutic window, or related parameter.56
Quasibill also noted that Big Pharma displayed the general cultural atmosphere of
waste that we normally identify with the Land of Cost-Plus Pricing, usually found in
military contractors and the like.
Have you ever been to a Merck campus (yes, they are campuses, not buildings or
sites)? If you look at the structure of the business, the first thing that strikes you is that
it looks like Detroit, circa 1980. And there's only one reason for that - government
protection of their profit margin. A good friend of mine works there - makes over
100G a year in a union job, where he gets written up if he does too MUCH work. And
yet while Detroit has suffered and is still paying for employing such a business model,
Pharma's been posting huge profits. Why's that?57
And a great deal of Big Pharma's drug R&D is conducted at taxpayer expense, either
through subsidies to the drug giants, or through research actually carried out in university
and government agency labs.58
The one thing the massive organizational size and expenditure aren't very good at,
according to Michael Perelman, is innovation. They attempt to compensate for their
mediocre performance in developing new drugs "by more intensive marketing, taking
over smaller, more innovative companies, and laying off workers."59 He quotes a Wall
Street Journal article:

56

Comment on Kevin Carson, "Intellectual Property Stifles Innovation," Mutualist Blog, May 21, 2006.
http://mutualist.blogspot.com/2006/05/intellectual-property-stifles.html
57
Comment on Ronald Bailey, "This Is One Reason People Hate Drug Companies," Reason Magazine
Hit&Run blog, February 24, 2006, http://www.reason.com/blog/show/112756.html.
58
Adams and Brock, 2nd edition, p. 58.
59
Michael Perelman, "Pharmaceutical Crackup?" EconoSpeak, December 8, 2007.
<http://econospeak.blogspot.com/2007/12/pharmaceutical-crackup.html>

The rise of generics wouldn't matter so much if research labs were creating a stream of
new hits. But that isn't happening. During the five years from 2002 through 2006, the
industry brought to market 43% fewer new chemical-based drugs than in the last five years
of the 1990s, despite more than doubling research-and-development spending...
The dearth of new products has led the industry to invest heavily in marketing and legal
tactics that squeeze as much revenue as possible out of existing products. Companies have
raised prices; the average price per pill has risen 63% since 2002, according to Michael
Krensavage, Raymond James analyst. Companies raised advertising spending to $5.3 billion
in 2006 from $2.5 billion in 2001 and since 1995 have nearly tripled the number of industry
sales representatives to 100,000....
The industry spent $155 million on lobbying from January 2005 to June 2006, according
to the Center for Public Integrity, on "a variety of issues ranging from protecting lucrative
drug patents to keeping lower-priced Canadian drugs from being imported." The industry
also successfully lobbied against allowing the federal government to negotiate Medicare drug
prices, the center said. The lobbying has drawn fire from politicians, doctors and payers, and
damaged the industry's public image.60

After a decade or so of relative fluidity caused by the disruptive onset of
globalization, global capital has settled back (with joint ventures and strategic alliances)
into the same oligopoly pattern as that of the old American economy. That's especially
true of the auto industry. After a brief period of admittedly traumatic shock, when they
first encountered vigorous Japanese and European competition,
the Big Three began to spin a far-reaching web of joint ventures and alliances with
their major foreign competitors. Thus, General Motors (still the world's biggest auto
manufacturer) has joned with Toyota (then the largest importer of automobiles into
the U.S. market) to jointly produce compact cars in California. G also has acquired
sizable ownership in Japanese carmakers Isuzu and Suzuki, built a jointly owned
production plant with Suzuki in Canada, and acquired half-ownership of Swedish
manufacturer SAAB. Ford, for its part, acquired a 25 percent ownership stake in
Mazda (later expanded); joined with Mazda to acquire an ownership stake in the
Korean car firm Kia; joined with Mazda to build a production facility in Flat Rock,
Michigan; combined its Latin American operations with Volkswagen (subsequently
dissolved); and engaged in partnerships with Nissan to jointly produce vehicles (in
addition to more recently acquiring outright control of Jaguar, Volvo, and rolls
Royce). Chrysler joined with Mitsubishi to build the Diamond Star Motors assembly
facility in Bloomington, Illinois, while spawning a variety of partnership pacts with
other global car firms.
At the same time, the major American and European auto manufacturers participate in

60

Barbara Martinez and Jacob Goldstein, "Big Pharma Faces Grim Prognosis: Industry Fails to Find New
Drugs to Replace Wonders Like Lipitor," Wall Street Journal, December 6, 2007, in Ibid.

the respective USCAR and EUCAR R&D consortia.61 So thanks to joint ventures,
foreign automakers have reason to view themselves more as partners than as competitors
to the American firms in this country. Lawrence Wilkinson brilliantly described the way
in which corporations regulate innovation, as oligopoly reasserts itself:
We're headed to a world that's more oligopolylike, a transition from a period of
robust change to a period of lock in.... All over, there's a settling down, a slowing
of the pace of change. Companies aren't really killing innovation -- they're
rationalizing it to manage its pace. The definition of oligopolistic economics is
three or so players behaving in lockstep with the marketplace. They don't
necessarily collude, but they develop ways of signaling pricing and containing
innovation.62
F. Economy of Scale in Agriculture.
If there is one industry in which the triumphalist rhetoric of "superior efficiency" of
large size is unjustified by reality, it is large-scale agribusiness. The reader has surely
heard the rhetoric: claims that without "Green Revolution" techniques "the world would
starve," ADM's boasts that "we feed the world," etc.
But the claimed "superior efficiency" of the large-scale agribusiness operation over
the family farm is illusory. Likewise unfounded is the claimed superiority of mechanized,
chemical agriculture, whether family or corporate, over more labor-and soil-intensive
forms of production. The large agribusiness operation, with mechanized row-cropping
and monocultures, is the most efficient "solution" to an artificial problem. The
techniques of the so-called Green Revolution are only more efficient if one assumes from
the outset the goals of the latifundistas and other state-privileged landed oligarchs in the
Third World, and of the giant agribusiness interests in the West.
According to a 1973 USDA pamphlet (of all things), even mechanized farming
reaches peak efficiency at a fairly small scale. Like all other internal economies of scale,
economy of scale in mechanized farming relies mainly on making full use of equipment:
The fully mechanized one-man farm, producing the maximum acreage of crops of
which the man and his machines are capable, is generally a technically efficient farm.
From the standpoint of costs per unit of production, this size farm captures most of
the economies associated with size.... Beyond that range there may be diseconomies
due to the increasing burden of supervision and communication between supervisor
and workers.... The incentive for increasing farm size beyond the technically
optimum one-man form is not to reduce costs per unit of production, but to increase

61
62

Adams and Brock, 2nd edition, pp. 160-61.
Quoted in Harriet Rubin, "Power," Fast Company No. 65 (November 2002), p. 76.

the volume of business, output, and total income.63
More specifically, USDA studies have found that the optimal size farm for raising
vegetables (using conventional mechanized techniques) is around 200 acres, while the
optimal cereal farm in the Midwest tops out at 800 acres.64
The secret to the success of large-scale agribusiness is not greater internal efficiency,
but its greater efficiency at manipulating the state for benefits. The real difference in
profitability comes from the channeling of state-subsidized inputs to large-scale
agribusiness. As California family farmer Berge Bulbulian testified to Congress,
...Probably the biggest obstacle we face in our struggle to save the family farm is
the attitude of many Americans, including some farm people, that the family farm is
obsolete, it is inefficient, and therefore unable to compete with the efficient and wellfinanced conglomerates. Well-financed they are, but efficient they are not. I
challenge any giant agribusiness corporation to match my efficiency. There is no way
a large concern with various levels of bureaucracy and managed by absentee owners
can compete in terms of true efficiency with a small, owner-operated concern....
....No, I can't sell for a loss and make it up in taxes, nor can I lose on the farming
end of the business and make it up at another level as a vertically integrated operation
can....
I have no political clout and lobbying to me means writing a letter to my
Congressman or Senator. But that is not what efficiency is all about.
Efficiency has to do with the relation between input and output. No, the big
agribusiness firms are not efficient except in farming the government.65
The family farm is more efficient than the large agribusiness operation (what Mason
Gaffney calls "latifundia") in terms of output per acre. Gaffney found that while big
corporate farms have somewhat higher output per man-hour, their output per acre is
actually less than that of small farms.
One may at least firmly conclude that large farm units are less improved and less
peopled than small and medium-sized farms. There are two possible interpretations.
One is that big farms are more efficient, getting more from less, but that is refuted by

63

W.R. Bailey, The One-Man Farm (Washington, D.C.: USDA Agriculture Economic Research Service,
1973), pp. v, 3. Quoted in L.S. Stavrianos, The Promise of the Coming Dark Age (San Francisco: W.H.
Freeman and Company, 1976), p. 38.
64
Kirkpatrick Sale, Human Scale, p. 233.
65
Farmworkers in Rural America 1971-1972. Hearings before the Subcommittee on Children and Youth
of the Committee on Labor and Public Welfare, United States Senate, 92nd Congress, 11 January 1972,
Part 3A, p. 1156. In L.S. Stavrianos, The Promise of the Coming Dark Age, pp. 38-39.

their getting less output per $L. The other is that Veblen was right, many of them are
oversized stores of value, held first to park slack money and only secondly to produce
food and fiber, and complement the owner's workmanship. The Florida 9 may
represent a home grown rural "third world" of large, underutilized landholdings that
preempt the best land and force median farmers onto small farms on low-grade land.66
According to Frances Moore Lappé, large landowners--both in the U.S. and in the
Third World--are not only least productive in terms of output per acre, but they hold huge
tracts of arable land out of cultivation. In Colombia, for example, a 1960 study found that
the largest landowners, who controlled 70% of the land, planted only 6% of it.67 The best
land, belonging to the large landholders, was often used for grazing cattle instead of
growing staple crops.68 In Guatemala, Del Monte planted only 9,000 of its 57,000
acres.69 Small cultivators are consistently found to produce greater outputs per acre. In
India, the smallest farms produce per-acre outputs a third higher than the larger ones. In
Thailand, farms of 2-4 acres produce 60% more rice per acre than farms of over 140
acres. A World Bank study in Latin America found a three- to fourteen-fold difference in
yield per acre between small and large farms.70
And bear in mind that these comparative figures on optimal economy of scale apply
only when the large- and small-scale operations are both engaged in conventional
mechanized row-cropping. The use of intensive raised-bed techniques for vegetables (the
biointensive method of John Jeavons, for example) is far more productive than
conventional commercial agriculture in terms of output per acre.
[T]he small farmer working with his own labour on a family holding, has been shown
in a wide variety of developing countries... to produce more per acre than big estates.
Some of the highest yields are to be found in countries where acre limitations are
strictly enforced. This productivity is secured not by heavy machines which drink
gasoline and can easily damage fragile soils, but by hard work with light equipment
which is by definition less prone to generate ecological risks. Fertilizers and
pesticides are less lavishly used, human and animal wastes are more carefully
husbanded. Greater personal care keeps terraces in trim, shade trees planted, gullies
forested. And earnings are not spent, as is often the case in semi-feudal economies,
on acquiring more land for extensive use, thus pushing up land prices and driving
working farmers away from the soil. Nor are they withdrawn altogether from the
rural economy, by the development of 'Western' standards of consumption or an over-

66

Mason Gaffney, from Chapter 10 of Ownership, Tenure, and Taxation of Agricultural Land, edited by
Gene Wunderlich (Westview Press), excerpted in Dan Sullivan's seminar on "The Myth of Corporate
Efficiency" at SavingCommunities.Org http://savingcommunities.org/seminars/corpefficiency.html.
67
Frances Moore Lappé, Food First: Beyond the Myth of Scarcity (New York: Ballantine Books, 1977),
p. 14.
68
Ibid., p. 42.
69
Ibid., p. 107.
70
Ibid., pp. 183-84.

affection for numbered accounts in Swiss banks.71
John Jeavons, in developing successive versions of his biointensive farming
techniques,72 has managed to reduce to four or five thousand square feet the space needed
to meet the bare subsistence requirements of the average person. Of course, it is a
relatively spare and monotonous diet, with the vast majority of the space devoted to high
carbohydrate cereal grains, legumes or tubers that concentrate a great deal of caloric value
in a small area. Only a small fraction of the space, perhaps 20%, can be spared for fruits
and vegetables to supplement the diet with vitamins. But 4000 square feet is about half
the space available even on a standard suburban residential lot. Even for the cul-de-sac
denizen, that leaves considerable space for additional vegetable beds, a few dwarf fruit or
nut trees and berry bushes, and a patch of alfalfa or some extra corn to feed chickens and
rabbits. The careful prevention of rainwater runoff, the saving of surplus rain in cisterns
for dry season irrigation, the composting of kitchen scraps and human waste--all these
things would make possible a nearly closed loop of food production.
In fact, some 15% of the world's total food production currently takes place in cities.
In China, back garden, rooftop and small lot production together supply 85% of urban
vegetable consumption, along with significant amounts tree crops and meat.73
All this is not to say that complete household sufficiency in food, or the elimination of
division of labor between town and country, is either necessary or desirable. It only
means that it is possible. A return to agriculture based on intensive work with the spade,
u-bar and fork would not mean starvation. It would mean greater output per acre than is
presently the case. And based on Borsodi's experience, even if the production process
itself is more labor-intensive in such small-scale production than mechanized
conventional farming, the overall labor required might still be less from the point of view
of the subsistence farmer substituting labor in direct production for wage labor to earn the
money to buy food; the wage laborer buying store food must, after all, work enough to
pay the transportation and marketing costs, which comprise more of the typical food
dollar than the actual production.
It's especially important to remember that there's no such thing as generic or
immaculate "technology," independent of the purposes of those who design it. The
decision to develop one techology, rather than another, is made from the perspective of
someone's interest. The choice of a particular technology is an answer to a question--so
we should always be aware of who's asking the question. The avenues of technological
development taken by the Green Revolution reflect a conscious political decision to
develop technologies of use primarily to large-scale agribusiness with access to

71

Barbara Ward and Rene Dubos, Only One Earth, in Godfrey Boyle and Peter Harper, eds. Radical
Technology (New York: Pantheon Books, 1976), p. 249.
72
John Jeavons, How to Grow More Vegetables (Berkeley and Toronto: Ten Speed Press, 1974).
73
Hawken et al, Natural Capitalism, p. 200.

government-subsidized irrigation water and other inputs, rather than technologies that
would increase the productivity of the peasant smallholder without subsidized water.
Large-scale plantation agribusiness, typically, flourishes only when supported by
government-subsidized irrigation projects. For example, a large share of American
produce comes from rain-poor areas of the West: vegetables are actually imported by
rain-rich regions like New England, because subsidized irrigation water makes the
Western operations artificially competitive. It is far more cost-effective in semi-arid
regions, when irrigation is not subsidized, to use cisterns to save water from the limited
rainy seasons for use through the dry period. For a subsistence farmer making intensive
use of small spaces, runoff from the rainy season may well be sufficient to provide
irrigation water during the dry spell. The main technical problem is providing enough
storage tanks. The ITDG was quite successful in designing cheap water tanks made from
local materials.74 And biointensive horticulture, which minimizes plant spacings and
maximizes soil cover, requires up to 88% less water than conventional large-scale
farming.75
The so-called "Green Revolution" in the Third World, particularly, occurred in the
context of a colonial history where peasant cultivators were pushed off of the best land
and onto marginal land, and the most fertile, level land was used for plantation farming of
cash crops. It is a myth that Third World hunger results mainly from primitive farming
techniques, or that the solution is a technocratic fix. Hunger results from the fact that land
once used to grow staple foods for the people working it is now used to grow cash crops
for urban elites or for the export markets, while the former peasant proprietors are
without a livelihood.
The techniques of subsistence production were often well-suited to the existing
situation.
Colonialism destroyed the cultural patterns of production and exchange by which
traditional societies in "underdeveloped" countries had previously met the needs of
the people. Many precolonial social structures, while dominated by exploitative elites,
had evolved a system of mutual obligations among the classes that helped to ensure at
least a minimal diet for all.... The misery of starvation in the streets of Calcutta can
only be understood as the end-point of a long historical process--one that has
destroyed a traditional social system.76
(It's also worth mentioning that colonial administrations, by ruling through the abovementioned "exploitative elites," often removed all the traditional checks on their power.

74

George McRobie. Small is Possible: A factual account of who is doing what, where, to put into practice
the ideas expressed in E. F. Schumacher's SMALL IS BEAUTIFUL (New York: Harper & Row, 1981)., p.
45.
75
Hawken et al, Natural Capitalism, p. 210.
76
Lappé, Food First, p. 100.

The British, e.g., turned the village headman in India into a tax farmer, and thus abrogated
the cutomary peasant control of land in the village communes. The general phenomenon,
turning local elites into landlords with absolute title in the modern European sense, was
widespread throughout the colonial world.)
Native farming techniques, often derided by colonizers as primitive or backward,
were in fact well-suited to local tradition as the result of generations of experience.
Lappé cites A. J. Voelker, a British agricultural scientist in India during the 1890s:
Nowhere would one find better instances of keeping land scrupulously clean from
weeds, of ingenuity in device of water-raising appliances, of knowledge of soils and
their capabilities, as well as of the exact time to sow and reap, as one would find in
Indian agriculture. It is wonderful, too, how much is known of rotation, the system of
"mixed crops" and of fallowing.... I, at least, have never seen a more perfect picture
of cultivation.77
Colonial agricultural policy focused all subsidies to research and innovation on export
crops, leaving subsistence techniques to stagnate. Slaves and hired farm laborers had no
incentive for preserving traditional knowledge, let alone refining technique. To the
contrary, farm laborers had every incentive to do the bare minimum, reduce output, and
even sabotage production. (I believe Adam Smith had similar observations about the
incentive effects of absentee land ownership in England.) The African peasant "went into
colonialism with a hoe and came out with a hoe." The most important effect of plantation
culture, perhaps, was a "narrowing of the experience of agriculture to plantation work...
[which] over generations robbed entire populations of basic peasant farming skills."78
Lappé cited the observations of Pascal de Pury, a WCC agronomist, that
often [appropriate] technology turns out to be rediscoveries of a people's traditional
practices that Western arrogance caused them to be ashamed of. Over and over again
he finds peasant cultures that had refined and adopted techniques over centuries to be
losing them in our time. What stands to be irretrievably lost is... successful,
productive techniques uniquely suited to local conditions....79
It is impossible to understand the so-called Green Revolution as it occurred in the
Third World, unless one first understands the political context in which it took place.
The central facet of that context was the process by which the land of subsistence farmers
was expropriated and turned over to cash crop cultivation, native populations were
reduced to dependency, and formerly independent peasants were often forced to engage in
cash crop production. The best land was often taken over by the colonial powers and
handed over to settlers, and the former subsistence cultivators transformed into farm
laborers.

77

Ibid., pp. 101-02.
Ibid., p. 113.
79
Ibid., p. 173.

78

...Throughout the colonies, it became standard practice to declare all
"uncultivated" land to be the property of the colonial administration. At a stroke,
local communities were denied legal title to lands they had traditionally set aside as
fallow and to the forests, grazing lands and streams they relied upon for hunting,
gathering, fishing and herding.
Where, as was frequently the case, the colonial authorities found that the lands
they sought to exploit were already "cultivated", the problem was remedied by
restricting the indigenous population to tracts of low quality land deemed unsuitable
for European settlement. In Kenya, such "reserves" were "structured to allow the
Europeans, who accounted for less than one per cent of the population, to have full
access to the agriculturally rich uplands that constituted 20 per cent of the country. In
Southern Rhodesia, white colonists, who constituted just five per cent of the
population, became the new owners of two-thirds of the land.... Once secured, the
commons appropriated by the colonial administration were typically leased out to
commercial concerns for plantations, mining and logging, or sold to white settlers.80
Sometimes the labor of the dispossessed was secured by slavery and other forms of
forced labor, although the colonial powers usually preferred to use direct taxation on
people, land and houses to compel the native population to enter the wage labor market.
Lappé presents some instances of her own. For example, in 1815, following the
British conquest of the Kandyan Kingdom (present day Sri Lanka), all central parts of the
island were designated as crown land and sold for nominal prices to coffee planters, with
government funding of surveying and road-building costs. In Java, the Dutch
administration "authorized" village headmen (usually under the influence of bribes) to
lease communal land to Dutch plantation companies. Often entire villages were thus
"sold" to foreign planters, without the consent of the rightful owners of the land.81
Colonial authorities worldwide similarly abrogated the traditional status of land, when it
was the inalienable property of a village commune or clan, by making it--in violation of
native law--usable as a pledge for debt. Likewise, such communally-owned land was
often made seizable for non-payment of taxes by the individual cultivator.82
In addition, colonial authorities simultaneously granted protectionist privileges to
settler plantations and imposed legal disabilities on independent native producers,
through the mercantilist policies of shipping companies and produce marketing boards.83

80

"Development as Enclosure: The Establishment of the Global Economy," The Ecologist (July/August
1992) 133.
81
Lappé, Food First, pp. 103-06.
82
Ibid., pp. 114-15.
83
Walter Rodney, "Chapter Five. Africa's Contribution to the Capitalist Development of Europe: The
Colonial Period," in How Europe Underdeveloped Africa (Dar-Es-Salaam: Bogle-L'Overture Publications,
London and Tanzanian Publishing House, 1973) Transcribed by Joaquin Arriola

Given this maldistribution of land through state-abetted land theft (either by colonial
regimes or by landed oligarchies in collusion with Western agribusiness interests), the
logical next step is for the state to divert inputs like subsidized irrigation systems, roads,
and so forth, disproportionately to the large plantations while denying them to subsistence
farmers. The state's direct subsidies and loan programs are set up so that only large
holdings, with access to preferential benefits like state-subsidized irrigation, can qualify.
Heavily state-subsidized agricultural R&D, likewise, is channelled in directions geared to
increasing the profits of cash crop agriculture on the big plantations, rather than to
increasing the productivity of small peasant holdings.
The "high-yielding variety" (HYV) seeds associated with the so-called Green
Revolution are normally productive only under the most favorable conditions, like those
prevailing on the big agribusiness plantations. The Green Revolution was a statesubsidized research project to develop plant varieties tailored to the prevailing conditions
in the state-subsidized agribusiness sector. They are deliberately designed to be
productive, in other words, under precisely the conditions provided by corporate
agribusiness.
...[T]he term "high-yielding varities is a misnomer because it implies tha the new
seeds are high-yielding in and of themselves. The distinguishing feature of the seeds,
however, is that they are highly responsive to certain key inputs such as irrigation and
fertilizer.... [W]e have chosen to use the term "high-response varieties" (HRV's) as
much more revealing of the true character of the seeds.... Unless the poor farmers can
afford to ensure the ideal conditions that will make these new seeds respond..., their
new seeds are just not going to grow as well as the ones planted by better-off
farmers....
Just as significant for the majority of the world's farmers is that the new seeds
show a greater yield variability than the seeds they replace. The HRV's are more
sensitive to drought and flood than their traditional predecessors....
HRV's are often less resistant to disease and pests. [They supplant] varieties that
had evolved over centuries in response to natural threats in that environment.84
They are, in other words, "highly responsive" to plentiful water from subsidized
irrigation projects, large-scale inputs of chemical fertilizer and pesticides, and
monocultural growing conditions. And they are also most responsive on the kind of
especially fertile, well-watered land that just happened to be stolen by landed elites under

<http://www.marxists.org/subject/africa/rodney-walter/how-europe/index.htm>.
84
Lappé, pp. 130-31.

the colonial regimes or post-colonial landed oligarchies.
Under the conditions of peasant subsistence farming, the traditional drought- and
pest-resistant varieties are far more productive. Locally adapted varieties tend to be
drought-resistant and hardy, and to produce steady yields under harsh conditions.85
Locally adapted varieties are also highly responsive to the kinds of inputs that are
more likely to be within the means of the small subsistence farmer: for example, better
plowing and harrowing techniques and weed elimination, crop rotation, green manuring,
better soil conservation, and better moisture retention in the soil.86
"Green Revolution" seeds are like a genetically engineered superman who will die
outside of his plastic bubble.
In Mexico, 97.7% of land devoted to corn and most land devoted to wheat lacked
irrigation. The Institute for Agricultural Investigation, a Mexican research organization,
set out to develop varieties of corn and wheat that would produce greater yields on small
non-irrigated farms. But the Rockefeller Foundation concentrated on developing varieties
that produced high yields in response to high levels of irrigation and synthetic fertilizer.
...The resulting new "miracle" strains enabled Mexico to become self-sufficient in
wheat, but the beneficiaries were the wealthy landowners, who could afford the
fertilizers and irrigation. The mass of the Mexican peasants have experienced
increased unemployment or underemployment with the growing mechanization of the
large estates.
The same pattern prevailed in India, Pakistan and the Philippines, where research
went to developing seed varieties primarily of benefit to large landowners with access to
subsidized irrigation water and fertilizer, rather than to the 70-90% farming non-irrigated
land. At the same time, the resulting land hunger on the part of the great subsidized
farmers has led to pressure to expropriate smallholders by abrogating traditional rights of
land tenure, and to evict tenant farmers paying rent on land that is rightfully theirs. The
landless and the underemployed rural proletariat, in turn, swell the urban slums with
people who once fed themselves.87 In addition, as Lappé observed (or perhaps, rather,
recycled an observation at least as old as Henry George) that the increased productivity
from Green Revolution seeds drives up rents, with crop share rents increasing from the
traditional 50% to 70%.88 Naturally, this further increases the tendency toward eviction
of small holders and the consolidation of the large estates.
It is a widespread observation that the large plantations benefiting from Green

85

Ibid. p. 130.
Ibid. pp. 150-51.
87
Stavrianos, The Promise of the Coming Dark Age, pp. 42-44.
88
Lappé, Food First, pp. 136.
86

Revolution techniques are likely to receive highly preferential access to subsidized inputs
like irrigation water. According to Michael Perelman,
...It is true that the Green Revolution has increased the amount of wheat and rice
produced in Asia. But it is also true that the adoption of this technology requires
heavy government subsidies in the form of cheap credit, favorable foreign exchange
rates, and high government support prices.... Much of the increase comes from the
use of irrigation for prime agricultural lands. Extending irrigation is expensive and
some observers even question whether it is possible to continue irrigating without
depleting the ground water.89
As a good example of the big landed interests' privileged access to subsidized
irrigation water, consider the case of Pakistan. The big landowners seek new dams to
provide more subsidized water for their agribusiness plantations--and since they don't pay
for it themselves, they're not very careful about how they use it:
We, as a nation, tend to build, neglect and throw away, only to build again. There
is no concept of maintenance. Pakistan has the largest contiguous irrigation system in
the world. It is supposed to be a miracle of engineering that has helped increase our
food production. But we don't maintain it. Operation, maintenance, and replacement
costs a lot of money. Where is that money coming from?
Some of the data in the recent World Bank report, "Pakistan's water economy
running dry," is quite frightening. When comparing Pakistan with Australia, the report
shows that in Australia, the entire cost of efficient operation, maintenance and
replacement is paid by the actual users, whereas taxpayers pay the interest on any
loans that may have been accrued in putting that water system into place.
In Pakistan, taxpayers - not users - are paying most of the operation and
maintenance costs, no one is paying for replacement.... When we can't even look after
our existing infrastructure, is there even a case for building new infrastructure?....
We have little additional water to mobilise. We've already used up everything that
exists in our water cycle so when we say we're putting up another dam or reservoir, it
doesn't necessarily mean there will be additional water coming in, we are just reappropriating what's already in the system. Who's going to pay for the additional
investment? We've taken so many loans to be returned over a long term period, how
much more can we sustain? Our water resource base is severely degraded because of
pollution and atrophying and overuse, groundwater is being over-exploited. Flooding
and drainage problems are also going to get worse, partly because of climate change
but also because of the way we manage our water system. The water infrastructure is

89

Michael Perelman, "Farming for Profit in a Hungry World: The Myth of Agricultural Efficiency," in
Louis Junker, ed., The Political Economy of Food and Energy (Ann Arbor: University of Michigan, 1977),
p. 34.

in terrible disrepair - everything is broken, there are leakages, powerful people create
their own direct links. We have poor governance, low levels of trust, water
productivity is extremely low, what we produce per acre, regardless of the crop, is
still less than what others are producing....
Water rights in Pakistan is tied to ownership of land, so in spite of so many
reforms, we still have very big farms owned by very powerful people, (rather than
smaller farm owners) and landless peoples who actually work the land. The biggest
farms are in southern Punjab and upper Sindh, while northern Punjab has smaller,
more owner-worked farms. Where we have bigger landlords with their rent-seeking
behaviour on the land, their payment for water is not a major consideration. Where
sharecropping arrangements have been perpetuated, there isn't much impetus to
change because the system suits the landowners.
So all we hear about is a demand for more water. The entire world is going on to
use less water and grow more crops but here we are shouting for more water to
maintain some of the lowest productivity not only in the world, but also in the
subcontinent. There are so many cheap technologies available - drip and sprinkler
irrigation and there are already people here producing this equipment. In our rural
economy, the whole use of labour on farms suits those in power, while others have no
voice.90
The same resources currently put into subsidizing the needs of agribusiness, if put
into research efforts in the interest of small-scale farmers, would have meant a
fundamentally different direction of technical development. L.S. Stavrianos wrote:
Large corporations are... virtually the sole beneficiaries of agriculture research
financed by the federal, state, and county governments. Research oriented toward
benefiting family farms would devise cooperative-ownership systems and credit
schemes; develop low-cost simple machinery; provide information on the purchase,
operation, and maintenance of machinery; and promote biological control of insect
pests. Instead, scientists with research grants develop complicated and tremendously
expensive machines. They breed new food varieties better adapted to mechanical
cultivation.... Paramount has been the vision of rural America as a factory producing
food, fiber, and profits for vertical monopolies extending from the fields to the
supermarket checkout counter.91
The administration of Lazaro Cardenas in Mexico, during the 1930s, is a good
example of the result when state policy is less one-sided. His agrarian reform, starting in a
country where two percent of the population owned 97% of the land, resulted in 42% of
the agricultural population owning 47% of the land and producing 52% of agricultural

90

"Interview--Simi Kamal" Newsline (Pakistan) February 2006.
http://www.newsline.com.pk/NewsFeb2006/interviewfeb2006.htm.
91
The Promise of the Coming Dark Age, p. 35.

output. Under Cardenas, state loans and technical support were aimed primarily at the
needs of small-scale agriculture. The result was an explosive increase in the rural
standard of living. As for state-funded agricultural R&D,
...The purpose... was not to "modernize" agriculture in imitation of United States
agriculture but to improve on traditional farming methods. Researchers began to
develop improved varieties of wheat and especially corn, the main staple of the rural
population, always concentrating on what could be utilized by small farmers who had
little money and less than ideal farm conditions.
Social and economic progress was being achieved not through dependence on
foreign expertise or costly imported agricultural inputs but rather with the abundant,
underutilized resources of local peasants.... Freed from the fear of landlords, bosses,
and moneylenders, peasants were motivated to produce, knowing that at last they
would benefit from their own labor.92
The groups alienated by Cardenas--the great rural landowners, the urban commercial
elites, and (as you might expect) the U.S. government--reasserted their political control
under Cardenas' post-1940 successor, Avila Camacho. Rather than small farms and
cooperatives, development spending was directed, on the American model, toward
electric power, highways, dams, airports, telecommunications, and urban services
that would serve privately owned, commercial agriculture and urban
industrialization....93
The Camacho administration, naturally, was heavily involved in the postwar
Green Revolution. The direction of the new big research program was diametrically
opposite to that under Cardenas.
...Policy choices systematically discarded research alternatives oriented toward
the nonirrigated, subsistence sector of Mexican agriculture. Instead, all effort went
to the development of a capital-intensive technology applicable only to the
relatively best-endowed areas or those that could be created by massive irrigation
projects.94
Under Camacho, huge irrigation projects were developed for favorably situated
land owned by big landed elites, and massive state subsidies were provided for the
importation of mechanized equipment.
As Lappé writes, the Camacho approach could not coexist with that of Cardenas.
The Cardenas agenda of increasing the productivity of peasant proprietors would have

92

Lappé, Food First, pp. 123-24.
Ibid., p. 124.
94
Ibid., pp. 125-26.

93

increased their standard of living; in so doing, it would have reduced the surplus
going to urban and export markets rather than domestic consumption, and also
reduced the flow of landless refugees to the cities. In other words, the Cardenas
policies threatened the supply of cheap wage labor for industrialization, and the
supply of cheap food to feed it.
The point to all this is not that Cardenas' version of state intervention was
desirable, but 1) that the present system touted by neoliberals as the "free market"
involves at least as much state intervention; and 2) that there is no such thing as
neutral, politically immaculate technology that can be divorced from questions of
power relationships. Criteria of technical "efficiency" depend on the nature of the
organizational structures which will be adopting a technology. And the forms of state
R&D subsidy and other development aid entailed in the Green Revolution artificially
promoted capital-intensive plantation agriculture, despite
overwhelming evidence from around the world that small, carefully farmed plots
are more productive per acre than large estates and use fewer costly inputs...95
What's more, the high-response varieties developed by the Green Revolution
crowded out equally viable alternatives that were more appropriate to traditional
smallholder agriculture. Any just assessment of the Green Revolution must take into
consideration the path not taken (or Bastiat's "unseen"). The Green Revolution,
coming as it did on the heels of land expropriation, channelled innovation in the
directions most favoring the land-grabbers. It was a subsidy to the richest growers,
artificially increasing their competitiveness against the subsistence sector.
...Historically, the Green Revolution represented a choice to breed seed
varieties that produce high yields under optimum conditions. It was a choice not
to start by developing seeds better able to withstand drought or pests. It was a
choice not to concentrate first on improving traditional methods of increasing
yields, such as mixed cropping. It was a choice not to develop technology that was
productive, labor-intensive, and independent of foreign input supply. It was a
choice not to concentrate on reinforcing the balanced, traditional diets of grains
plus legumes.96
HRVs are actually less hardy and durable under the conditions prevailing on
subsistence farms--less drought-resistant, for example. Locally improved varieties
are specifically adapted to be productive under conditions of low rainfall, and more
resistant to insects and fungi without costly chemical inputs. Local seed varieties,
combined with intensive techniques and the creative use of biological processes,
result in levels of output comparable in many cases to that of Green Revolution seed

95
96

Ibid., p. 127.
Ibid., p. 153.

varieties combined with heavy chemical inputs and subsidized irrigation. Even
setting aside the long-term costs of soil depletion, good husbandry with local varieties
of seed produce almost as much corn and sorghum output per acre. An experiment in
Bangladesh--ceasing pesticide use in order to raise fish in rice paddies--resulted in a
25% increase in rice production, along with the high quality protein from the fish.
The fish controlled insects more efficiently than chemical pesticides, and fertilized the
rice.97
A rural development agenda geared toward the interests of peasant proprietors
would have emphasized, not increasing the yield of seeds in response to expensive
irrigation and chemical inputs, but improving the soil.
This brings us back to our earlier consideration of the concept of "efficiency."
The discussion above gives the lie to vulgar Coasean arguments that justice in
holdings doesn't matter, as long as they wind up in the "most efficient" hands. For one
thing, it matters a great deal to the person who was robbed; it matters a great deal
whether you're producing enough staple crops on your own land to feed your family,
or instead holding a begging bowl in the streets of Calcutta or living in some tinroofed shantytown on the outskirts of Mexico, while your stolen land is being used to
grow export crops for those with the purchasing power to buy them. But more
importantly, the Green Revolution and the alternatives it crowded out demonstrate-again--that there's no such thing as generic "efficiency" in the use of resources. The
"most efficient" use of a piece of land depends mightily on who owns it, and what
their needs are. An "efficient" technique for the land thief is entirely different from
what would have been efficient for the land's rightful owner.
One can afford to be a lot less efficient in the use of inputs that he gets for free.
Capital-intensive techniques that increase output per man-hour, but reduce output per
acre, are suited to the interests of American-style agribusiness. They're perfect for
large landowners who, as a historical legacy, have preferential access to large tracts of
land (to the extent that they can even afford to hold significant parts of it out of use),
but want to reduce their dependence on hired labor. In areas with underutilized land
and unemployed population, on the other hand, it makes a lot more sense to increase
output per acre by adding labor inputs. And this is exactly the pattern that prevails in
small-scale agriculture. Lappé found, in a survey of studies from around the world,
that small farms were universally more productive--far more productive--per acre than
large plantations. Depending on the region and the crop, small farms were from onethird to fourteen times more productive. The efficiency of small proprietors working
their own land, compared to plantation agribusiness using wage or tenant labor, is
analogous to that of the small family plots in the old USSR compared to the state
farms. Plantation agriculture is able to outcompete the peasant proprietor only through

97

Ibid., p. 127.

"preferential access to credit and government-subsidized technology...."98
Mechanized, large-scale production is more efficient, not in terms of food output
per acre, but in terms of dollar output per laborer. That makes perfect sense if you're
a capitalist farmer with more land than you can use (thanks to the state), and you want
to minimize labor costs and agency problems through a strategy of capital
substitution. But it doesn't make much sense where there's millions of unemployed
people who would rather be working the land than squatting in the streets of Calcutta
or the shantytowns of Mexico City.
Green Revolution techniques are very "efficient" indeed--but only given the
artificial objectives of those who stole the land.
The same general observations apply to agribusiness in the developed world. As
Michael Perelman observes, the intensive raised bed techniques of early modern
Europe compare quite favorably to the outputs per acre of today's mechanized
agribusiness. For example, he mentions a seventeenth century Paris gentleman who
produced 44 tons of vegetables per acre; modern methods in the U.S. produce only
15 tons of onions or 8.6 tons of tomatoes--the highest-yielding crops--per acre.99 In
the modern Green Revolution,
the really revolutionary changes in American agriculture have not been directed
toward increasing yields.... Actually, the unique achievement of U.S. agriculture
is not the production of maximum crop yields [per acre] but the harnessing of
fossil fuel energy to replace human energy in agriculture.100
Conclusion.
Overall, the importance of economy of scale was summed up very well by Barry
Stein, in his concluding remarks on a survey of the empirical literature:
Such uncertainty and variability suggest that technical economies of scale are not
the primary determinant of either competitive ability or true efficiency. Available
data indicate first, that in most industries the penalties for operating plants well below
the apparent optimal scale are not great; second, the presence of substantial relatively
constant costs (added to those directly associated with production) dilutes even those
clear advantages of greater productive scale; and third, there is no strong case to be
made for significant economies of firm (as against plant) size.101

98

Lappé, Food First, p. 189.
Michael Perelman, Classical Political Economy: Primitive Accumulation and the Social Division of
Labor (Totowa, N.J.: Rowman & Allanheld; London: F. Pinter, 1984, c 1983) pp. 41-42.
100
Michael Perelman, "Farming for Profit in a Hungry World," pp. 40-41.
101
Size, Efficiency, and Community Enterprise, pp. 24-25.
99

So why are giant corporations able to survive, despite such manifest violation of all
the laws of efficiency? There are really two questions involved here that we need to
attend to separately.
First, the evidence above demonstrates that most large plants, let alone multiplant
firms, operate far beyond optimal size for economy of scale. Yet they are still profitable
despite being less efficient in terms of unit costs even under the conditions of the existing
state capitalist economy. Why is this?
The reason is twofold. First, they are protected, by state intervention, from the
competitive disadvantages resulting from inefficiency. A state-cartelized oligopoly firm
can operate at higher costs and pass its costs on to the consumer, because it is protected
from the full vigor of competition from smaller and more efficient producers.
Second, as we already mentioned at the outset of this chapter, the figures above for
optimal economy of scale assume the existing input costs, without considering the extent
to which the state subsidizes inputs and externalizes a wide range of operating costs on
the taxpayer.
In the next chapter, we will consider the whole range of measures by which the state
restricts competition and subsidizes inefficiency costs.