S~~~\~~     1253) 
POLICY RESEARCH WORKING PAPER   13 3 1
The Myth of Monopoly                                              V largefinnsaremore
prevalent in the United States
A  New  View  of Industrial Structure                            than in Ria.And there'is
in  Russia                                                       little evidence in Russia of
in RUSSia
industrial concentration in
Nnational narkets. lnstead -
Annette N. Browrz
',  barriers to competion in
Ba-rry W. Ickes
Razndi Ryterman                                                  Russ'ia arise as a resulc of
highy segmented product
marke. Corequently.
traditional policy remedies.
a'appropriate for problems of
concentration (such as
antitrust policy and import
competition) may be Hi-
advised or inadequate for
addressing problems of
imperft compeition in the
Russian econory. The.
presciption for healthier
competion: improve Russia's
distributon systen  and
fa-cilitate the entry of new: -
firrns-
The World Bank
Policy Research Department
Transition Eoonomics Division
August 1994  



POLICY RESEARCH WORKING PAPER 1331
Summary findings
Discussions of economic reform in the Russian              Instead, barriers to competition in Russia arise as a
Federation arc colored by the conventional view of       result of highbly segmented product markets. In large part,
Russia's industrial strucrurc. Both in Russia and in the  this segmentation can bc viewed as a legacy of central
Wesr, Russian industry is characterized as very large    planning. Under the prior regime, enterprises were
enterprises operating in highly concentrated industries.  highly isolated, divided along both ministerial and
Brown, Icke  and Ryterman challenge the                geographic lines. Presently, these barriers are reinforced
conventional view. They assess Russian industrial        by some fe&tures of the transitional environment that
concentration by comparing the Russian industrial        continue to undermine the efficient distribution of
structure (as revealed in the 1989 Soviet Census of      goods.
Industry) with that in the United States and other         Brown, Ickes, and Ryterman conclude that the
countries.                                               traditional policy remedies appropriate for problems of
They find that very large firms are more prevalent in  concentration (snxch as antitrust policy and import
the United States than in Russia. This empirical fact    compedtion) may be ill-advised or inadequate for
suggests that planners economized on the costs of central    addressing problems of imperfect competition in the
economic coordination not by building unusually large    Russian economy.
enterprises, but by not building very small enteprises.    They argue instead that improving the distrbution
Their most important finding: That there is little     system and other market infrastructure that supports
aggregate or industry concentration at the national level    trade and facliting the entry of new finns should be the
in Russia. Monopolies and oligopolics actually account   most critical elements of competition policy in Russia.
for only a small share of national employment and
production.
This paper-a producr of the Transition Economics Division, Policy Research Department-is parr of a larger effort in
the deparmnent to understand the adjustment of enterprises in economies in transition. Copies of the paper are available
free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Maxine Berg, room N11-054,
extension 36969 (68 pages). August 1994.
Tec Policy Rearch Working Paper Sris dsemnates the fmdiw of wait as progres to encoure the exchacng of ide about
dnudopmensue Anobceah,e of thesfies is toga thefns outquickly, even tife presemman rekss thanfudfl polishe& The
papers carry the names of the auors anddsbd be usd and cid accordingly. The flding, iterpretatknt and condusions are the
adtors' own and sbold not beatątibsaed to the World Bank its Execsate Board of Dirtrs, or any of is menmbr countries.
Produced by the Policy Research Dissemination Center



The Myth of Monopoly:
A New View of Industrial Structure in Russia*
Annette N. Brown
Department of Economics
University of Michigan
Ann Arbor, MI 48109
(313) 994-0440
anbrown@umich.edu
Barry W. Ickes
Deparment of Economics
The Pennsylvania State University
University Park, PA 16802
(814) 863-2652
i04@psuvm.psu.edu
Randi Ryterman
The World Bank
1818 H Street, NW
Washington, D.C. 20433
(202) 473-7037
nryterman@worldbank.org
Correspondence to: Barry W. Ickes, Department of Economics, The Peansylvania State University,
University Park, PA  16802
Slhe authors thank Abram Bergson, Joseph Berliner, Abe Becker, Rick Ericson, Gregory Grossman,
Gur Ofer, Rughvir Khemani, Bryan Roberts, Alan Gelb, and workshop participants at Indiana,
Michigan, Penn State, RAND, and the World Bank for helpful comments. We also thank Yvonne
Ying and Miodrag Deric for research assistance. Barry Ickes and Randi Ryterman are grateful to the
National Council for Soviet and East European Research for its support This work was completed
while Annett Brown was a visitor at the World Bank and she is grateful for its support.



The Myth of Monopoly:
A New View of Industrial Structure in Russia
1.    Introduction
Discussions of economic reform in the Russian Federation are colored by the conventional
view of Russian industrial structure. This view, held both in Russia and in the west, is that Russian
industry is characterized by very large enterprises operating in highly concentrated industries. This
perception of industrial structure has important implications for policy, and for the interpretation of
developments in Russia.' For example, based on the view that Russian industry is dominated by
monopoly enterprises, Russian policymakers have reintroduced price controls on a wide variety of
commodities.
The conventional wisdom about Russian industrial structure is based on a generation of
research on the Soviet system of central planning, research that appeared to be supported by available
evidence. According to this view, planners econoinized on the costs of central planning by creating a
highly concentrated industrial sector with a small number of very large enterprises in each industry.
Before transition, the conventional costs associated with imperfect competition - higher prices and
restricted output - could be overcome by the control of prices and the setting of output targets by
central planners. Now, however, as the Russian economy adopts the market system, problems of
industrial structure become important and affect the design of econoric reftbmL Consequently, it is
crucial that the conventional wisdom be re-examined.
In this paper, we challenge the conventional wisdom. We estimate Russian industrial
concentradon by examining the Soviet Industrial Census of 1989, and by comparing the Russian
industrial structure to that in the United States and other countries. We find that the conventional
wisdom about Russian industrial structure is seriously misleading. We find, for example, that very
large firms are more prevalent in the United States than in Russia, as are very small firms. Our
analysis suggests that planners economized on the costs of central economic coordinauion, not by
buiding wwsually large enterprises, but by not buiding very small enterprises. Because innovation
was centrally directed, small firms did not play the role they play in a market economy, and thus
industry evolved under a completely different process of economic selection.
Our most important finding is that there is little aggregate or industry concentration at the
national level in Russia. Monopolies and oligopolies actually account for a very small share of
national employment and production. Our analysis suggests instead that the barriers to comperition in
'A frequently offered explanation of the output drop that fbllowed price liberalizatipn is that
monopolies reduced output to raise prices.



ThduaWd Concamtraooa
Russia arise as the result ofproduct markets ta are highly segmente'J? In large part, this
segmentation can be viewed. as a legacy of the system of central planning. Nevertheless, some
features of the transition environment continue to undermine the efficient distribution of goods,
reinforcing these barriers. Based on our finding, we conclude that the traditional policy remedies
appropriate for problems of concentration, such as anti-trust policy and import competition, may be
ill-advised or inadequate for addressing problems of imperfect competition in the economy. We
argue instead that imwroving the disbution system and fiacilitating the entry of new firms are the
most critical elements of competition policy in Russia.
The remainder of the paper is as follows. In the rest of section 1, we focus on the process of
industrial evolution under central planning. We also discuss the conventional wisdom, and explain
why we think a further examination is warranted. Section 2 discusses data and methodology. In
section 3, we present our findings in the form of a series of "myths" and wrealities." We then turn,
in section 4, to analyze the barriers to competition that do exist in Russia today. Section 5 discusses
the implications of our findings for economic reform in Russia. Section 6 concludes the paper.
1.1   Industril Evolution Under Central Planning
The industrial structure that exists in the Russian Federation today was crealid mostly as a
consequence of decisions made during the prior economic regime. Therefore, we begin our
discussion of industrial concentration in Russia by considering the process of industrial evolution
under central planning. Some of the onfusion that arises over the present industrial structure is a
direct consequence of differences in views over the important features of this process. We present
these two views below.
Much of the convenlonal view about industrial evolution under central planning arises from a
particular folk model (the 'cookie cutter" model) of planner or ministerial decision making. In this
model, a Stalinist minister is assigned the responsibility of building a new industry. The minister.
studies the set of technologies available to produce the product to identify the efficient scale of
production within a centrally planned sewing. After this efficient scale is identified, the mirister uses
the cookie cutter to carve a set of identical factories at this scale to satisfy industrial development
needs.
'Mhe predominant source of monopoly power under central planning was the sellers' market
created by price controls. Price liberalization then is perhaps the most important element of
competition policy in Russia and other countries of the former socialist world.
2



Indusud Co_wmn__n_
Clearly, the conventional model supports the view that Russian industry is very homogeneous.
Each industry is populated by identical firms: . Also, the model is often interpreted to suggest that
industry is highly concentrated and dominated by very large firms. It is generally believed that, by
building fewer and, thus, larger enterprises; the administaive costs of central planning could be
economized.' Moreover, Stalinist ministers are generally believed to have identified scales of
production that were quite large, a phenomenon known as- gigantomania.' Eva Ehrlich [1985: 293]
relates this bias quite simply, 'In the socialist countries, large size and economic efficiency were
thought to be synonymous." Stalin, especially, preferred large scales of production because such
enterprises stood out as impressive examples of Soviet industrialization.5
An alternative view of industrial evolution under central planning is implicit in the more
recent literature on ministerial decision-making under central planning. The older literature
emphasizes the similarity of technologies within an industry among firms when they are first built; the
newer literature emphasizes the differences between firms that become more pronounced as they age.
This newer literature presents models that stress the role of hidden information and hidden
action in the behavior of decision makers under planning. In these models, industrial ministers must
allocate production targets to enterprises with imperfect knowledge of their true productivity. To
elicit information from enterprises about their potential productivity, the minister must provide
enterprise directors with sufficient incentives to meet their targets.
These models provide several insights into the type of industrial structure that evolves as a
consequence of this system of incentives. First, implicit in the model is the belief that enterprises in
each industry are heterogeneous. They differ in productive potential, possibly because of differences
in managerial skills or behavior, location, access to suppliers, or even technology. Second, over
time, these differences may not be fully revealed. When the incentives directors face do not
adequately reward them for fully revealing inforrnation, the directors of higher productivity
'Fewer enterprises makes it easier to construct the plan and also reduces the costs of monitoring
its implementation.
'See, for example, Gregory and Stuart [1986: 143]. We note that, in early research on Soviet
enterprises, gigantomania referred to the tendency of the Soviet ministers to built gigantic plants.
More recently, however, this concept has been used to describe enterprise size.
'Discussing the fixation with large enterprises, Peter Wiles 11962: 3041 writes: "There is
something 'socialist' and 'progressive' about mere size, even if unaccompanied by lower costs.
Gigantomania as such, then, reinforces the view that large capital expenditures are a good thing, even
where smaller ones will do."
3



ladueria Concgaden
enterprises will choose to conceal the true potential of their enteririses, precisely because they lbow
that ministers will use this information against them in designing fiture production targets. This
dynamic incentives problem, the "ratchet effect,' impairs economic performance because the treat of
higher targets causes enterprise directors to demand greater rewards in return for full revelation. This
makes it more costly for planners to obtain important information.'
Ickes and Ryterman [19931 use these observations to develop a model of industrial evolution
under central planning. In most of the literature on managerial decision making, the number of
enterprises is given. The problem considered in the Ickes-Ryterman model is the determination, by
the relevant mi ister, of how many enterprises should be built in a given industry. In this model,
planners provide industrial ministers with a stream of aggregate output targets to be met over time.
The industrial minister, in turn, must disaggregate these targets and award them to specific enterprises
in the ministry, building new enterprises and expanding existing enterprises to meet this goal. The
minister cannot, however, shut down enterprises that are observed to be high cost because the absence
of exit is one of the distinguishing features of centrally planned economies. Of course, the productive
potential of new enterprises is learned only over time as the ministers observe the ability of enterprise
directors to consistenty meet production targets at low cost.
This model also provides important insights into the evolution of industry under central
planning. Again, the model stresses the presence of heterogeneity within industry. Over time, as
ministers observe the performance of enterprises, they naturally award larger targets to enterprises
with a demonstrated ability to meet their production goals. Thus, the model predicts that a mature
industry will be populated by a mixture of enterprises - larger more productive enterprises and
smaller less productive ones. Thus, this and other more recent models raise the possibility that
industry, in fact, is not dominated by very large enterprises and, consequently, may not be highly
concentrated.
Ickes and Ryterman suggest that the maximum and minimum scales of enterprises in a
centrally planned industry were determined by an interplay of technology and the costs of centrally
planned production. All else equal, these administrative costs increase with the number of levels of
hierarchy as well as well as the span of control at each level. Therefore, these costs are viewed as
favoring the creation of fewer larger enterprises than would be created in a market setting. We
believe that diseconomies of scale that are created in very large enterprises quickly outweighed the
administrative advantages of very large firms. Larger firms require more layers of hierarchy to
'Mhe cassic reference is Berliner [1957]. Keren [19931 provides a survey of this literature.
4



monitor production, quickldy adding to costs. Moreover, small firms were not required to play the
important role - in fostering innovation - tat they play.in market economies. In market economies,
small firms enter the market, experiment with a new product or process, then grow or fail based on
their success. Under central planning, a different process of economic selection was implemented,
one in which product innovation was typically produced in larger science-production associations.
Therefore, we believe that ministers economized on coordination costs predominantly by choosing not
to build very small enterprises.
Industrial concentration, in turn, is determined by the interplay of technology and the costs of
centrally planned production, on the one hand, and demand, measured by the size of output targets,
on the other. For a given technology and costs of coordination, industrial concentration will be
higher in industries that were presented with lower aggregate output targets. This feature suggests
that, to some extent, important industries for which product demand was high are less likely to be
concentrated than less important industries.7 More importantly, given the large size of the former
Soviet Union, this feature suggests that industrial concentration in national markets is unlikely to
exist.
1.2   Problems with the Conventional Wisdom
In par-, the conventional view has been so compelling historically because some empirical
evidence does appear to support it. In this section, we explain why we believe this evidence is
misleading.
Heidi Kroll, in recent work [19911, presents some evidence for bodt large firm sizes and
industrial concentration in the Soviet Union. Concerning enterprise size, she states that, since the
:960's, the size of Soviet enterprises has been increasing,
The average number of employees per enterprise rose to 813 in 1987-88, and 73.4 percent of
the labor force now work in enterprises employing more than 1000 workers; indeed,
enterprises with 10,000 workers or more employ 21.6 percent of the labor force, while those
with 500 workers or fewer employ only 14.9 percent of the labor force...3
7However, we do recognize that some small industries may have strategic importance to an
economy.
Kroll [1991: 147].



Induarld Cencrntraion
She gives many of the conventional reasons to expLain why Soviet enterprises were so large. While
these numbers seem compelling, they are haed if not impossible to interpret without comparison with
other countries.
Eva Ehrlich [1985: 267-295] does compare enterprise and establishment sizes between
capitalist and socialist economies. While her survey does not include data from the Soviet Union, her
methodology and conclusions are relevant to our study. She employs two measures to describe
industrial structure: average employment and size distribution of firms. Her tables show quite clearly
that average enterprise sizes in the socialist economies, especially Poland and Hungary, are greater
than in even the 'large-type' capitalist economies. The tables also show that size distributions of
firms vary distinctly between the two types of economies: the socialist economies have a greater
percentage of employment in the large firms and a lower percentage in the small firms than the
capitalist countries do. Ehrlich relates the following metaphor for socialist industrial structure,
A Hungarian economist compares the size structure of the Hungarian economic system to a
pyramid turned upside down, characterized not by large enterprises relying on a broad base of
small- and medium-size firms but, on the contrary, by a preponderance of big enterprises and
a sirnificant lack of small and medium ones.2 
Given their understanding of the distinct size distributions, what Ehrlich and Kroll fail to do is
to weight or correct the averages they use according to the different size distributions and ihe
different overall numbers of firms in these economies. We can illustrate this problem with a simple
example. Consider two economies, each with 10 firms of different sizes, but distributed equally
-across the two economies so that the average firm size is the same. Now, give one economy, A, 10
more firms all of which are smaller than the smallest original firm, and leave economy B the same.
The average firm size in A is now much smaller, but that does not mean that economy B has more
large firms than A. In. other words, this change does not winvert" B's pyramid, but rather just builds
to the bottom of A's. In this paper, we try to draw a clear picture of size structure both by using
country comparisons and by correcting or explaining our measures in terms of size distribution.
Kroll also provides evidence of industrial concentration.
According to Goskomstat SSSR, more than one-third of the most important types of machine-
building products are produced by a single enterprise, and approximately the same share is
produced by only two enterprises.. .According to Gossnab, 80 percent of the volume of output
in machine building is manufactured by monopolists, and 77 percent of machine-building
'Ehrlich [1985: 294].
6



Industdn  Ciucantnadon
enterprises are monopolists.. .Another statistic from Gossnab is that about 2,000 eniterprises in
the country are the sole producers of.a specific type of product...
The World Bank [1992: 82] provides even more startling statistics. Under the former regime, the
State Conunittee for Material Technical Supply, Gossnab, organized the delivery of 7,664 distinct
product groups. According to the World Bank, 77 percent of these products were produced by single
enterprises."
These stsics measure a specific type of concentration-product concentration-that is, they
measure concentration in terms of the ability of the consumer to find alternative suppliers of the exact
same product. Due to central planning, however, product categories in the former Soviet Union were
defined very narrowly. In order for the plinners to ensure that the input needs of each producer
would be met. they (or the appropriate industrial minister) assigned very specific targets for each
intermediate good. For example, in principle, the output target for one centimeter nails would be
distinct from the target for two centimeter nails. As a consequence, the number of different product
categories used by planners in the former Soviet Union was enormous. Moreover, enterprises
specialized in production more highly than in the West.12 Under central planning, there was no
incentive, let alone authority, for the enterprise to diversify its production. Thus, while concentration
measured in terms of products will naturally appear greater than concentration measured at a more
aggregated level, this disaggregation effect is exaggerated in the Soviet case." In this context, the
statistics quoted above are not all that surprising.
"Kroll [1991: 144-145].
"Kahn and Peck [1991; 62-67] also discuss the prevalence of monopoly in Soviet industry.
'2Granick [19671 is the classic reference on the forces that went into the design of enterprises and
factories in the Soviet metal fabricating industry. One of his important insights was the link between
product specialization and scale (1967: 361: "If each plant were to limit its output to a single product -
- or to a small range of products if its output of one item would exceed the total planned consumption
of the entire USSR - and concentrate all its facilities on such production, then each plant could gain
economies of scale."
"This point is illustrated, albeit unwittingly, by the IMF-World Bank-OECD-EBRD joint study on
the Soviet economy [1991: 16]: "Industrial production in most sectors tends to be highly concentrated
in one or a few enterprises. For example, in almost two-thirds of the 38 product groups included
under sledge-press machines, the largest enterprise accounted for 75 percent ore more of total
production in 1988..." The key point, however, is that with the demise of central planning
enterprises that produce different types of sledge-press machines can compete against each other.
7



lgddrd Cohwiado
The other type of concentration one can measure is tndusty concentration, that is,
concentration measured in terms of the ability of the producer to supply substitute or directly
competing products in the short term. This measure is a broader measure than the one abve,
requiring enterprises to be classified and compared based on their industry rather than the specific,
products they produce. It is based on the assumption that enterprises compete by producing the same
products, producing substitute products, or by being able to easily alter production in order to make
the same or substitute products.
In a market seting, where an enterprise's choice of product mix is not centrally determined,
this approach more correcty measures potential competiftion between firms that produce similar types
of products. Potential competition is an important feature of an industrial structure. In a
decentralized enviromnent, the threat of entry, most easily from other firms producing similar
products, often serves to discipline existing firms in a particular product market. If existing firms set
prices too high (or quantities too low), the incentive for near competitors to enter the product market
and share in oligopoly rents is raised. Thus, if a market is contestable, competitive conditions may
exist even if only one or a few firms produce a product.
The question, then, is which measure is most appropriate for this analysis. During the Soviet
period, there was no competition and little opportunity for enterprise directors to choose their product
mix. Now, with enterprise reform, decisions about product mix are decentralized, and there is the
potential for competition. In fact, evidence from an assortment of surveys on enterprise behavior
suggests that some enterprises, both state-owned and privatized, are adapting their product mix in
order to survive. Thus, we believe that measures of industry concentration more accurately reflect
the incentives related to competition in the economy. Furthermore, from a practical perspective,
studies of market economies typically use industry measures rather than product measures to discuss
concentration and competition. Not only does this suggest that industry measures are more
appropriate; it means that by using industry measures, we are able to compare Russia to other
countries.
Kahn and Peck [19911 also find that measures of product concentration exaggerate the role of
monopoly and oligopoly in the Russian economy. They provide data on the number of industries in
Russia that appear to be concentrated, and compare these statistics to similar ones for the United
States. They find, as we do, that a larger number of industries in Russia appear to be concentrated.
The problem with their analysis, however, is that it fails to take into account the importance of these
B



laraatlh CoImCmraUo
industries."' As we) demonstrate, below, this correction is critical to understanding the degree of
industrial concentr;ition that actually exists i the Russian economy.
2.    Data and Methodology
We compile statistics on industrial concentration in the Russian Federation using data
collected for the 1989 Soviet Census of Industry. This data set includes all civilian enterprises
engaged in production activities defined by the Soviets as industrial. For a summary of the
characteristics of enterprises in the data set, see Table 1.
The enterprises in this data set are classified based on the primary conunodity they produce
and, correspondingly, are assigned a U.S. Standard Industrial Classification (SIC) code." As our
primary concern is industrial concentration, we focus our attention at the four-digit SIC level."'
Further disaggregation would take us into the realm of product concentration which, as we have
argued above, is of less interest.
Typically, in studies such as ours, sales are used as the principal measure of a firm's size.
Unfortunately, in countries such as the former Soviet Union, data measured in value terms do not
provide useful measures of a firm's activity. Aggregate measuves of economic behavior that are
expressed in value terms have ambiguous meaning because prices were determiined by administrative
flat, and not by market interactions. Thus, although we provide some statistics based on sales, we
primarily use employment to measure the market position of firms within an industry.
Our choice of 1989 was dictated by circumstance; that is the year for which we have the data.
There are, however, some distinct advantages to this year. The survey methodology used by
"4Kahn and Peck [1991: 65] acknowledge this problem, but did not have the data needed to
properly weight industries according to their importance in the economy.
"Unfbrtunately, we do not know the industrial code assigned to the enterprises by Goskomstat
(the State Statistical Office). Consequently, we cannot aggregate our data into the standard branch
divisions used in Soviet publications, which are quoted often in western analyses. Thus, for example,
we are not able to look at machine building as a separate branch.
"Although many enterprises are assigned codes at the five-digit level, we use four-digit codes in
our analysis. We base our decision on two factors. First, comparison at the four-digit level reveals
evidence of competition and potential competition within an industry, while comparison at the five-
digit lcvel would reveal information about product competition only. Moreover, this type of analysis
is nearly always conducted at the four-digit level for the U.S. and other western countries. Thus,
analysis at the four-digit level allows us to evaluate industrial concentration in the Russian Federation
in a broader context.
9



lnad uifa Coucmatmtfou
Goskomstar (the State Statistical Office) was based on central planning institutions. As those
institutions began to deteriorate, the quality ;f the survey frame and data also began to deteriorate.
Analysts are thus faced with a tradeoff. Earlier years most likely orovide more accurate data, but are
of less interest for their implications about the transition. In our assessment, 1989 was the last year
in which Goskomstat was able to conduct a survey at an sufficient level of quality for this analysis.
Fortnately, 1989 constiutes a good base year to assess the initial conditions of reform.
The staistics for the Russian Federion are presentea in context of statistics for the U.S. and,
to a lesser extent, for the O.E.C.D. We chose the U.S. as tie dominant country for comparison
because of its size and level of industrial development. The statistics on industrial structure in the
U.S. are based on U.S. Census Bureau data1, usually from 1987, although we do make some
comparisons which use other years; we specify when 1987 is not the comparator.
2.1    The Unit of Analysis
Conducting comparisons of industrial structure between the Russia and the United States
raises issues of the proper unit of comparison. The U.S. Census Bureau collects data at the company
and establishment levels, where an establishment is defined as all plants owned by a company that are
engaged in similar activities at one location. Russian data, however, are collected at the enterprise
level. Technically, an enterprise is a company. However, it differs from a western company in an
important way.
Unlike companies, enterprises are seldom multi-divisional firms in the western sense.
Although they may produce several products for sale, these products are typically in closely related
product groups. However, in addition, many enterprises produce products that do not reach the
market. Often, enterprises are vertically integrated, producing output that they consume as inputs.
Also, many enterprises engage in side activities such as farming to provide food for workers. Thus,
in practice, Russian enterprises might be quite diversified."' Many enterprises are made up of several
"We use data from both the Census of Manufacturers, Concentration Ratios in Manufacturing,
1987 and the 1987 Enterprise Statistics.
"'A good example is the production of machinery. According to Hewett [1988: 172]: 'Some
departments in nonmachinebuilding enterprises also produce machinery. Forty-five percent of all
metalworking equipment in the Soviet Union can be found there, a stock that by itself exceeds in
value the entire capital stock of the U.S. machinebuilding sector."
10



radutd Coceutwie
plants (zawd), often in different cities (and, during the Soviet period, even in different republics).
Hence, in certain respects, an enterprise is ldss than a company, but more than an establishment.'
For evaluating firm size and aggregate concentration, we compare Russian enterprises to
U.S. companies, where U.S. companies are measured by their domesic employment. Very large
enterprises in Russia are almost always multi-plant enterprises and, consequiently, are more like
companies than estdblishments. Moreover, for the discussion of policy implications related to firm
size and aggregate concentration, company is a more insightful measure. In contrast, for evaluating
industry concentration, we compare Russian enterprises to U.S. establishment-groups. Establishment-
groups are all domestic establishments in a single company that are classified with the same 4-digit
SIC code.?° The U.S. Bureau of the Census uses this unit to partition the company and allocate its
activities to different industries. This partition enables Census to compute concentration statistics that
reflect the ability of domestic producers to supply substitute or direcdy competing products in the
short term. To the extent that Russian enterprises are like establishments (in that their priniary
products for sale are closely related), this comparison is direct. To the extent that Russian enterprises
are, in fact, diversified companies, concentration measures will overstate the level of concentration in
the Russian economy. Consequently, using the enterprise data actually biases the Russian statistics
against our case.
The use of enterprise data does raise an important issue, however. Beginning with a decree
of Brezhnev in 1973, Soviet enterprises were organized into associations (obyedineniya).2 The
enterprises that comprise an obyedineniye were operated under a single management22 Hence, when
1"he extreme vertical integration of enterprises presents an important additional source of
potential competition in the Russian economy. In our analysis, we measure concentration based on
the industry to which the enterprise belongs. To the extent that the enterpr.se produces products that
belong to a different SIC code, we are understaing the potential for competition.
2Ihese statistics are based on establishment level data which are then aggregated into the
establishment-group unit of measure. The concenation ratios that we present at the two-dgit level
are based on company data with company as the unit of measure. The bias is not great here since
few companies will have establishments in two different two-digit industries.
2'The primary purpose of the reform was to reduce the administrative burdens on the planners, as
discussed above.
'According to Conyngham [1982: 228] the average number Of enterprises in an obyedinemiye
varied across sectors: "in the machine-building industries, the ptoduction unions average five or six
enterprises. In light industry, the average is nine... in the chemical, coal, and other extractive
industries, however, (associations) usually incorporate twenty-five or more enterprises."
11



lisdEsgrfo Concemnidon
we examine enterprise data, we may be treating units that belong to a single, larger structure as if
they were independent units. It is important to keep in mind, however, that our purpose in this paper
is to examine industrial strmcture&in Russia as it pertains to the development of a market economy. In
this context, enterprises seem to be the appropriate unit of analysis, because these are the units that
are typically being privatized.
2.2    The Military Industrial Complex (MIC)
The primary difference between the Russian and the U.S. data sets is the fact that the Russian
data set includes data on the civilian sector only. The military-industrial complex (MIC) in Russia is
very important Consequently, we cannot simply ignore its presence in our analysis. To deal with
this problem, we have assumed that the industrial structure of the MIC closely resembles that of
heavy industry in the civilian sector. We then use the observed distributions of heavy industry in the
civilian sector to produce esimates of the industrial structure of the MIC. We base our choice on the
observation that the process of industrial evolution in the MIC closely resembled the process of
evolution of heavy industry in the civilian sector. In fact, our calculations suggests that the mean size
of firms in the military-industrial sector is smaller than the mean size of firms in civilian heavy
industry?. In our analysis, we estimate the total number of firms in the military-industrial complex?'
and assume the distribution of these firms by size resembles the distribution in civilian heavy
industry. Although this process produces an upward bias in the distribution of firms in the military-
industrial complex (toward larger firns than is statistically evident), we feel this more conservative
approach is necessary to persuade readers to our point of view.
Specifically, to estimate the size distribution of enterprises in the military-industrial complex,
we computed the size distribution of enterprises in the following civilian branches: chemicals,.
industrial machinery and equipment, electronics, transport equipment, and instruments. We selected
these branches based on our belief that enterprises in the military-industrial complex produce similar
2'The average firm size for our proxy group of branches (chemicals, industrial machinery and
equipment, electronics, transport equipment, and instruments) is 1,750. We know that adding the
MIC to our data would add 5,309 enterprises and 7,979,161 workers. This implies an average firm
size of 1,503 for the MIC. The largest enterprise in Russia, for example, is Autovaz, a civilian, not
a military enterprise.
'Including fuels and energy production.
12



Inasdd  ncenwtean
types of products.2? Then, we applied this distribution to the number of enterprises that are kmown to
exist in the miliary-ipdistrial sector.' See Table 2 for more detailed information.
Although this method is useful in correcting biases inrroduced in estimates of enterprise size,
we do not use it in calculations of industry concentration. While omission of enterprises that are part
of the military-industrial complex may bias statistics related to firm size downward (for example, the
average size of the 100 largest firms), it is unlikely to do so in the case of measures of industrial
concentration. Enterprises in the military-industrial complex are of two types. Some produce
products that compete with those produced by enterprises in the civilian sector. In these cases, our
(unadjusted) statistics overstate the level of concentration in Russia, a bias that only strengthens the
force of our conclusions Alternatively, some enterprises in the military-industrial complex produce
products that are not produced in the civilian sector. In these cases, we suggest the reader look to
levels of concentration that are measured in heavy civilian industry for a prediction of the levels of
concentration in these industries.
3.    Myths and Reality
In this section, we discuss some of the most commonly held myths concerning firm size and
concentration in the Russian economy. We begin with a discussion of the size distribution of
enterprises. We then turn to industrial concentration at the national level, and finally to concentration
at the industry level.
3.1   Size Distribution of Enterprises
The conventional view of industrial evolution in Russia emphasizes the dominant role played
by very large firms in the economy. As a result, it views Russian industry as highly concentrated. In
2'We did not adjust the implicit weights assigned to any of the civilian branches in the
computation of the size distribution. We base this decision on the observation that the military-
industrial complex may have beer. fairly autarkic; as a consequence, its structure should reflect an
industrial balance not unlike that observed in the remainder of the economy.
'We note that the mean size of enterprises in heavy civilian industry is larger than the mean size
of enterprises in the military-industrial complex. As a consequence, when we use the size distribution
of firms in heavy civilian industry to approximate the size distribution of the military-industrial
complex, we unavoidably inflate the number of workers employed in this sector. Specifically; official
data suggests that 7,979,161 workers are employed in the military-industrial complex; however, when
we use the distribution of employment in heavy civilian industry to estimate the size of firms in the
military-industrial complex, we estimate that 9,289,726 workers are employed in the sector.
13



lnduadagr  C.ace&ubiea.
contrast, the more recent work on industrial evolution suggests that Russian industry is fairly
heterogeneous in terms of size, opening the possibility that industry is not dominated by very large
firms and, in fact, may not be highly concentrated. Clearly, then, the size distribution of enterprises
is at the very heart of our discussion of industrial concentration in Russia. Thus, we begin by
presenting evidence related to the size distribution of firms.
Myth 1: Many Russian enterprises are very large. The Russian economy suffers
from gigantomania.
Reality 1: In fact, Russia's largest enterprises are actually smaller than the largest
finns in many O.E.C.D. countries.
In Table 3, we compare the size of the largest finrs in O.E.C.D countries with that of
Russia. The ten largest civflian firms in Russia employed an average of 62,649 workers. We
estimate that the ten largest firms, including those in the military-industrial complex, employed an
average of approximatly 92,698 workers (table 8).27 We can compare these average-employment
statistics to those for the top ten firms in a sample of other countries using 1985 statistics found in
Scherer and Ross [1990: 63j. Table 3 shows that these statistics for Russia are notably smaller than.
those for the United States, Japan, West Germany, the United Kingdom, and France, and, on
average, about the same as that for Holland. The result is similar when one compares the average
firm sizes of the top 20 firms in these countries; that is, the average-employment of the top 20
Russian enterprises is notably smaller. In fact, there are only 113 civilian and approximately 217
total industrial enterprises in Russia that have 10,000 workers or more.
These findings are actually not that surprising. Most of the very large companies in the
O.E.C.D. countries are multinational enterprises that have both multinational labor markets - they
2tTo estimate the number of enterprises in the military-industrial complex that employ 10,000
workers or more, we calculated the percent of civilian enterprises in heavy industry that employed
10,000 to 19,999 workers, 20,000 to 29,000 workers, ..., 90,000 to 100,000 workers, and more than
100,000 workers and applied these percents to the number of enterprises in the military-industrial
complex. To estimate the size of the enterprises in the military-industrial complex, we calculated the
mean size of civilian enterprises in each size class and assigned the mean to all the non-civilian
enterprises widiin the class. We note hiat, due to the small number of firms involved in the
calculation of the ten largest finns in the economy (cited above), the particular estimate is subject to
large potential error.
14



J'udIL&WI Cmeawuauude
employ people at home and abroad - as well as multinational product markets for their goods. They
tend to operate in environments in which communication and other infrastructure facilitating large-
scale organization is good. Thus, their large size can represent an efficient scale of production.
Large Russian enterprises, on the other hand, cannot be considered conventional multi-national firms.
While they were often built to serve the C.M.E.A. market, they rarely located production outside
national boundaries. But, there is no reason to conclude a priori that their sizes are too large for
their national and regional markets.
The conventional belief, or myth, that Russian firms are unnaturally large is often used to
suggest that they must, in fact, be inefficiently large. This conclusion then supports the position of
those who favor breaking up Russian enterprises during transition. Our evidence shows, however,
that it is incorrect to assume that Russian enterprises are inefficient based on size alone. If Russian
enterprises are inefficient, their inefficiency likely comes from internal organization rather than from
scale of production. The inernal organization of many Russian firms reflects historic circumstances,
rather than market conditions. The assertion that Russian enterprises are "too large' is really an
assertion about the shape of the cost functions of these units. Without such data one cannot really
address the issue. But the comparison with other O.E.C.D. countres does provide some perspective
on the size of Russian enterprises. Organizational inefficiencies can be found in firms of all sizes and
require a much different set of remedies tan simply breaking up large firms into their constituent
plants.
Myth 2: The size distribution of Russian enterprises can be represented by an
inverted pyramid. There are many large enterprises and very few medium or small
enterprises.
Reality 2: The size distnrbution is better represented by an uprightpyramid. The
significant difference between the industrial structure of market economies and that of
Russia is that Russia lacks the myriad of very small firms found in market economies.
15



Indartlol Coacentradon
Tables 4, 5, and 6 show the distributions of Russian and U.S. manufacturing enterprise?s by
enterprise employment.' In
table 4, we see that, while Russia does have more large finns and the U.S. has dramatically more
small firms, for both countries the number of firms as a percentage of total firms in a size class
decreases as the size of firms in the class increases. Thus, the image created by the industrial
structure in both Russia and the U.S. is an upright pyramid, in which many smaller finns support
fewer larger firms. In table 5, we see that an inverted pyramid is present in the Russian case, but it
describes the distribution of small firms only. That is, the number of firms as a percentage of total
small firms in a size class increases with the size of firms in the class in the case of Russia, but
decreases in the case of the U.S. We consider this observation very important. It supports the
conclusion from the historical evidence that there was, in effect, a lower bound on enterprise size in
the former Soviet Union.
The relative absence of small enterprises in Russian industry suggests that comparing Russia
and the United States by looking at average firms size for industry as a whole may be misleading If
-we take the arithmetic mean, for example, we find that manufacturing firms in Russia employ an
average of 670 workers in the.civilian sector and an estimated average of 925 workers when firms in
the military-industrial complex also are considered. In contrast, manufacturng firms in the U.S.
enploy an average of 70 workers. We can correct for the small-frm bias by computing average
emnployment per firm for all manufacturing firms employing above a lower bound of 249 employees.
In Russia, these firms employ an average of 1,297 workers in the civilian sector and an estimated
average of 1,621 workers in both sectors, while, in the U.S., these manufacturing firms employ
2Manufacturing enterprises are a subset of industrial enterprises. See table-l for a breakdown of
all industrial enterprises in the Census into one-digit SIC codes.
2-We base the size categories on the current Russian definitions, with one exception. In tables 4
through 6, we define small enterprises based on the U.S. definition of 1 to 249 workers..
(Consequently, medium firms begin with 250 workers.) The 'very small' categories are based on the
categories in U-S. tables. In further tables which just present Russian data, small is 1 to 199
employees.
To- estimate the number of non-civilian firms in each size class, we calculated the share of
enterprises in heavy civilian industry in each size class and applied these shares to the total number of
firms in the military-industrial complex. Similarly, we used mean employment by enterprises in
heavy civilian industry in each size class as the approximate employment of non-civilian firms
assigned to the class.
16



lndWd Cancuantles
2,103 workers.30 Average employment by manufacturing firms with 50 workers or more is 761
workers in the civilian sector in Russia, an estimated 1,025 workers in both sectors, and 498 workers
in the U.S?1
The distribution of employment (rather than firms) across size categories presents a more
striking difference between the two countries (the lower panel of table 4). In Russia 91.5 percent of
civilian employment and an estimated 94.5 percent of total employment in manufacturing is provided
by enterprises with employment of 250 or greater, while only 73.1 percent of U.S. manufacturing
employment is provided by similar firms. But, these statistics obscure the dominant role of extra-
large firms in the provision of employment in the U.S. A closer examination of the size distribution
of firms in the two countries reveals this role. Three-fourths of Russian manufacturing employment
falls in the middle of the distribution, that is, in medium and large enterprises. In contrast, two-thirds
of U.S. manufacturng employment falls in the tails, in sinall and extra-large firms. Strikingly, 40.2
percent of the U.S. employment is provided by extra-large firms, while only 15.3 percent of Russian
civilian employment and an estimated 20.5 percent of total manufacturing employment are provided
by similar firms. Thus, in comparison with the U.S., the Russian economy is not dominated by
gigantic firms and Russian employment is not dominated by employment in very large firms.
The importance of medium and large enterprises in Russian industry is also evident in table
7, where we break down the size distribution of firms for separate branches of civilian industry in
Russia.32 This table also snows that production by very large firms is mostly concentrated in a few
industrial branches. Just four branches contain 84 of the 113 extra-large enterprises in the civilian
sector, and one of these, mining, is a non-manufacturing branch. The other three are industrial
machinery and equipment, primary metals, and transportation equipment. These branches not only
have a large share of their employment in extra-large enterprises - an average of 45 percent - they
also represent a large share of industrial employment - an average of 8.7 percent each. The results
from table 7 suggest that, for some branches, restructuring, especially reorganization of the large
enterprises, will have a significant effect on their relevant labor markets. But" this is clearly not true
'Mhe total numbers of firms used for these calculations are 8,131 for civilian manufacturing in
Russia, 12,131 for total manufacturing, and 7,454 for the U.S.
'"The total numbers of firms used for these calculations are 15,066 for civilian and 20,233 for
total manufacturing in Russia, and 37,604 for manufacturing in the U.S.
2The branch distinctions for the manufacturing enterprises are 2-digit SIC categories. The other
branches are 1-digit groups, except for construction and mining which have been divided into two.
17



IndaadL ConecatnmIon
for all branches of industry, nor even for branches with large shares of industrial employment. The
other branches with high shares of national ihdustrial employment - Food and Kindred Products,
Lumber and Wood Products, Stone, Clay, and Glass, and Textile Mill Products - have zero to little
employment in extra-large firms. Thus, while restructuring in the industries in these branches will
significantly impact the economy, the issues involved will be somewhat different.
3.2   Industrial Concentration
There are several measures of industrial concentration that are commonly utilized. Ownership
concentration measures how diffuse is the ownership of industry. Russian industry in 1989 was
entirely state-owned, leaving little to study?' Product concentration measures the degree to which
products are produced by few or many enterprises. This measure, however, says very little about
potential competition, because it disaggregates markets too finely, especially given the very distinct
product categories used in Soviet planning. Moreover, given the fact that central planners
deternined the assortment plans of enterprises, product concentration says very little about what
enterprises can produce. Rather, it is more a measure of the assortment planners and ministers chose.
Consequently, we focus our attention on measures of concentration both in the aggregate and
by industrial branch. The former essentially measures how large enterprises are relative to the size of
the economy. The latter is a measure of market power. We consider these in turn.
3.2.1 Aggregate Concentration
Perhaps the most salient component of the conventional wisdom is the belief that the Russian
economy is dominated by large enterprises.
Myth 3: The largest Russian enterprises account for an unusually large share of
national production and employnrent.
Reality 3: The largest enterprises account for only a moderate share of national
production and employment.
3The Russian privatization plan, since it gives advantages to insiders, wil most likely lead to a
rather diffuse ownership at least in the first stages. It will be interesting to study what happens to
ownership concentration in the faure.
18



Indua1 Concahude
Table 8 presents statistics describing the role, of the largest enterprises in industrial
employment and production. The ten largest enterprises by employment account for 4.6 percent of
national civilian employment; the top 100, 16.3 percent.  Our estimates suggest, that if enterprises
in the military-industrial complex were included in our sample, the top 100 enterprises in the entire
economy would account for a smaller share, approximately 14.3 percent of total employment.
Scherer and Ross present aggregate concentration shares for the United States for 1982r; table 9 is a
supplemented version of their table. They show that, in the U.S., the largest 100 manufacturing
corporations accounted for 23.8 percent of total U.S. employment. In Russia, sales of the largest 100
civilian firms accounted for 21.6 percent of total sales, while sales of the largest 100.firms in the
U.S. accounted for 31.8 percent of total sales of civilian goods. The comparisons for the largest 200
firms offer the same conclusions.
Table 10 compares aggregate concentration in the U.S. and Russia at higher level&. In spite
of the fact that Russia has 1118 the number of manufacturing 'irms as the U.S., the manufacturing
four-firn concentration ratio for Russia is three percentage points lower than for the U.S., and this
margin persists in the other groupings as well. Thus, aggregate concentration in Russia is less than in
the market economy of the U.S. These results relate to those above which show that enterprises are
not as gigantic as we once presumed.
Table 8 also shows how aggregate civilian coricentration is distributed across industries and
regions. The top 10 civilian firms only represent four branches; in fact, these firms represent only
four industries at the 4-digit S.I.C. level. This concentration of Russia's major enterprises in a few
number of branches and industries is also true when examining the top 25, 50, or 100 civilian
enterprises. These major enterprises seem to be rather broadly distributed across oblasts (provinces),
however. Seven different oblasts have enterprises in the top 10, and these fall in six different
economic regions. Thirty-nine out of the 78 oblasts in Russia contain enterprises with 10,000
workers or more, and these fall in all 12 of the economic regions. Thus, the major enterprises are
not geographically concentrated. Only one oblast, Kemerovskaya, stands out as having a large share
'Scherer and Ross [1990: 591. Scherer and Ross explain that aggregate concentration in the
U.S., in terms of domestic manufacturing activities alone, has risen insignificantly since the 1960's.
Thus, the use of 1982 data for the U.S. in comparison with 1989 data for Russia does not present a
significant problem.
19



IndusUla Cacentraai
of the major civilian enterprises. It has six of the top 25 civilian firms and seven of the top 50
civilian firms.
3.3   Industry Concentration
Prior to presenting our evidence, we should say a few words about measuring, comparing,and
thinking about industry concentration Any comparison of concentration between the U.S. and Russia
is tricky because the U.S. has both many more industries and many more firms than Russia. In
manufacturing alone, the U.S. has 448 industries at tde 4-digit S.I.C. level, while Russia has only
350 4-digit industries.6 Even when using percentages of concentrated industries versus actual
numbers, the comparison could be biased by the type of industries that Russia does not have. The
U.S. also has almost 18 times as many firms. In terms of firms per industry, the U.S. has an average
of 685.5 firms per industry in manufacturing, while Russia's average is only 49.1. In these simple
terms, Russia's production is clearly more concentrated. This does not mean, however, that Russia's
industrial markets are necessarily non-competitive.
We employ three measures of industry concentration. The most well known is the four-firm
concentration ratio (CR4). This is a ratio of the sum of the measure for the largest four firms
(according to the same measure) to the sum of the measure over all firms in the industry. More
specifically, we usually calculate the sum of employment in the four firms with the most employment
in the industry as a percent of the sum of employment in all the firms in the industry. We calculate
CRB's in the same manner for the largest eight firms in an industry?'
We also classify enterprises according to how many enterprises there are in that industry, that
is, we identify monopolists, duopolists, oligopolists, and others which fall into larger categories.
Finally, we use the Russian measure of a 'domidant' firm; a dominant firm is one which accounts for
a 35 percent or greater share of its industry's market in terms of employment or sales. In the tables
35Kemerovskaya oblst is located on the east border of Western Siberia in the south.
MThis number for the U.S. does include military production and thus likely includes industries
which are not counted in the Russian number.
"Without information about costs we cannot calculate Lerner Indices. We calculate concentration
ratios rather than Herfindahl-Hirschman Indices (HHI's). As Scherer and Ross explain, CR4's and
HHI's are highly correlated (Scherer and Ross 11990: 74]), and the analysis we do is not precise
enough to warrant the more complicated statistic.
20



Indaad  CoMceanden
looking at Russia alone; and unless otherwise specified, all measures are calculated at the four-digit
industry level. That means we report the highest possible measure of industry concentration.
We therefore implicitly define the relev;ant market for competition as the four-digit industry
level. This assumption is no more than a guess or a proxy. For some industries, the relevant
markets are more distinct, for some more aggregated.? As we explain earlier in the paper, we
intentionally measure industry concentration instead of product concentration. The four-digit level is
then the most disaggregated market we can choose. In the case where the relevant markets are larger,
we actually overstate concentration. We try to con;rol for this problem in the U.S.-Russia
comparisons by calculating the statistics at the same level in each country.
Myth 4: Russian industry is highly monopolized. These monopolies are very large
enterprises.
Reaity 4  When measured at the national level, there are very few monopolistic
enterprises in Russia, and most of these firms are relatively small.
Table 11 shows that only 43 of the 21,391 civilian Russian enterprises are monopolies in their
four-digit industries at the national level. While this represents 10.6 percent of all industries, it only
represents 0.2 percent of civilian firms, 0.2 percent of civilian employment, and 0.2 percent of
civilian output This measure of industy concentration and, thus, potential competition is much
lower than the statistics often quoted, for example the ones from Koll [19911 above, which is that
around 80 percent of Russian products are produced by monopolies. Table 11 also shows that, while
the mean size of a monopoly is 726 workers, the median is only 285 workers. Even 726 workers is
less than the means in all other competition classes, except the class with the most number of finms.
Looking at table 12, we see that there are no extra-large firms that are monopolies, or duopolies for
that matter. In fact, 32 of the 43 monopolies have less than 1,000 employees.
If we define oligopoly quite liberally to mean four or fewer firms in an industry, we see in
table 11 that, while 26.4 percent of industries in this sample are oligopolies, they only account for 1.1
percent of all these firms and 1.9 percent of all this employment. At the same time, 70.3 percent of
firms and 41.8 percent of employment falls in enterprises in industries which have more than 100
'Scherer and Ross [1990: 73] offer a good discussion of defining markets according to SIC
codes.
21



liadarld CDnecaindo
firms. The highest employment mean is in the industry category with five to ten firms, and table 12
shows that the plurality (and one-third of allj extra-large firms are in industries with 21 to 50 firms.
Mammoth monopolies simply do not dominate the civilian industrial sector.
Myth 5: Russian industry is heavily concentrated.
Reality 5.1: Measured at the two-digit level, U.S. manufacturing is at least as
concentrated as Russian manufacturing.
Table 13 lists CR4's and CR8's by employment for companies in two-digit manufacturing
branches in the U.S. and Russia." If one took the argument seriously that the industrial structure of
Russia was arbitrary, then one would expect there to be no correlation between the concentr.tion
ratios in Russia and the United States. In fact, however, the Pearson correlations between the U.S.
and Russia for the two sets are .72 and .82 respectively. These results suggest that, in terms of the
larger firms in these branches, the U.S. and Russian industrial structures are actually quite similar - a
surprising result if one believes that the Russian industrial structure should appear artificial or
unnatural as a consequence of central planning. Further, the mean CR4 and CR8 for the U.S. are
both greater than those for Russia; the hypothesis that they are equal is rejected with 97 percent
confidence or better. Table 14 presents similar information, but with the CR4's and CR8's calculated
by sales rather than employment. Here again, the Pearson statistic shows positive correlation with
values of .67 and .77. While the mean ratios for the U.S. are greater than those for Russia, in this
case one cannot reject the null that they are equal. This is still a surprising result when one expects
Russian manufacturing to be much more concentrated.
These tables present another interesting comparison - that between the CR's calculated in
terms of employment and those calculated in terms of sales within the two countries. In all four
comparisons, U.S. and Russia for four-firm and eight-firm, the concentration ratios calculated in
terms of sales are higher than those for employment. But, this difference is much greater for Russia
than for the U.S. The null hypothesis that the U.S. four-firm concentration ratios for employment
'Comparisons of the U.S. and Russia are potentially biased in favor of the result that the U.S. is
less concentrated. The reason is that the U.S. Census Bureau does not publish concentration ratios
when those numbers might reveal information about specific enterprises in an industry. These
missing data then are concentration ratios that are quite high; six such observations are deleted from
this analysis. This bias causes the analysis to understate concentration in the U.S.
22



lndaltu CenwaulXn
and sales have equal means cannot be rejected, but the null for the other three sets can be rejected,
and for the Russian numbers, the confidence level is greater than 99 percent. The mean concentration
ratios calculated by sales in the Russian series are at least six percentage points higher than those
calculated in terms of employment, while for the U.S. the difference is only 1.8 points. This finding
implies that, relative to the U.S., the large firms in Russian manufacturing represent a larger
percentage of output than they do of labor. This conclusion continues to hold if we look at industry
as a whole. Extra-large enterprises in civilian manufacturing account for 25.6 percent of
manufacturing output and only 15.3 percent of manufacturing labor; extra-large enterprises in all of
civilian industry account for 22.9 percent of industrial output and just 17.3 percent of industrial labor.
The small and .nedium-sized firms, on the other hand, account for a greater percentage of labor than
of output.
Reality 5.2: Measured at the four-digit level, the industrial structures of the U.S. and
Russian manufacturing sectors are quite different, and the Russian sector appears more
concentrated.
Comparing CR4's across 331 observations of matching U.S. and Russian four-digit industries,
we find that the Pearson Correlation is only .06. The mean CR4 for Russian manufacturing is 69
while the mean for the U.S. is 36, and, not surprisingly, the null that the means are equal is rejected
with 100 percent probability. It is interesting that the structures of two economies seem to diverge so
much when viewed at a lower level of aggregation. Upon examination of the data, we discover that,
in striking contrast to the U.S. industrial structure, many industries in Russia, are in fact, very small.
Thus, a more meaningful way to compute the Pearson correlations is to weight the comparisons of
industrial concentration by industry size.
One simple way of weighting is to compare only the predominant industries in Russia with the
same industries in the United States. Comparing the ratios for the top 25 percent of industries in
Russia in terms of employment, we find that the means are not significantly different; both are about
40. As we take larger groupings of the major industries, the mean CR4 for Russia increases while
that for the U.S. remains about the same. Even for the top 75 percent of industries though, the
Russian mean is still under 60, that is, on average, concentration is not high enough to be considered
a barrier.
23



iadaur  Conmman
'Reality 5.3: At the four-digit industry level, many more Russian industries are
concentrated than U.S. industries; however these industries account for a minority
share of Russian industrial activity.
When one thinks about the above analysis, it may not seem significant that the concentration
levels are roughly similar between the U.S. and Russia for only the top 25 percent of industries in
Russia. Further analysis reveals, however, that the top 25 percent of industries accounts for almost  .,
80 percent of employment in the 331-industry sample. The top 75 percent, whose mean CR4 is less
than 60, accounts for over 99 percent of the employment.
Table 15 presents more statistics on industrial concentration in Russia using the full sample of
Russian industries and grouping industries by concentration ratios instead of ordering by employment
From this table, we see that 55.2 percent of four-digit industries in all Russian industry have fc z'r-
firm concentration ratios of employment of 61 percent or more; 62.1 percent have four-firm ratios of
sales of 61 percent or more. This 55.2 percent of industries in the top half of the table represents
only 17.5 percent of industrial employment and 17.4 percent of industrial sales. To provide a context
for these statistics, we compare them to those for the U.S. Table 16 offers a comparison of U.S. and
Russian four-digit industry concentration in manufacturing using categories of four-firm concentration
ratios.' Here we see that, in the U.S., only 17.6 percent of industries have CR4's of 60 percent or
greater, while in Russia, 64.8 percent do. There is little doubt that more Russian manufacturing
industries are concentrated. However, the 17.6 percent of industries in the U.S. account for 19.8
percent otvalue added, while the 64.8 percent in.Russia account for only 30.9 percent of sales and
25.0 percent of employment. In other words, the percentage of industries in the U.S. that are
concentrated is much more indicative of shares of economic activity than in Russia. Thus, the
statistics describing the number of concentrated industriec are misleading whtn assessing the
importance of concentration to the economy.
Rea5Lty 5A: Concentrated industries in Russia have a different industrial structure
than industry in general.
'We use calculations from Scherer and Ross [1990: 83J which are based on 1982 establishment
data and use establishment groups as the unit of measure.
24



IaduaI Cwacen-wes
Concentrated industries in Russia have proportionately more medium, large, and extra-large
enterprises and fewer small ones than in industry in general. This demonstrates that the problem of
* very few smalf firms is even more important in concentrated industries. Concentrated industries
(defined, for example, as industries with CR4's of over 40 percent) are thus different from other
industries, not only in terms of the industries' top enterprises, but also in the overall distribution of
enterprises within the industries. This is evident from the data in the bottom half of table 12. In
addition, note from table 16 that concentrated industries account for a larger share of output than
employment, while in other industries just the opposite is tue. This seems to suggest that enterprises
in highly concentrated industries have higher labor productivity than in industry in general. One must
be careful about such inferences, however, since they are based on data generated under the system of
regulated prices and planned outputs.
Reality 5.5: In the set of industries that are highly concentrated, concentration is
mostly due to having few firms in the industry rather than to having principal firms in
the industry.
When industries are arranged by CR4's, into deciles for example, one might expect that the
share of total industries in a decile would be roughly similar to the share of total value added or
nationa employment in that decile. This is in fact the case for the US. For Russia, on the other
hand, while the top decile represents the lowest share of output and employment, it accounts for more
four-digit industries than the other three combined.on the top four deciles of industries by
concentration ratios (table 15). Table 16 shows a similar distribution - almost half of all industries
fall in t.e top category of the CR4 ranges. If we arranged the data in table 16 from quintiles into
deciles of concentration ratios, we would find that most of the 45 percent would indeed fall in the 90
to 100 category, but account for a small share of employment and sales. The very high CR4's
suggest one of two things: either production in these industries is highly concentrated in the largest
firms, or these industries have very few firms. Table 17 shows that the latter is true. One hundred
seven of the 125 industries in the 91 to 100 percent decile have four or fewer firms, and none of
these industries has more than 10 firms. There is, in fact, a triangle of zeros in the lower left portion
of the table - concentrated industries have few firms. No industry with an employment CR4 greater
than 60 percent has more than 50 firms.
It may seem obvious that this latter explanation is true. But, limited evidence on the U.S.
suggests that the opposite is true, Xt is, that high CR4's in the U.S. indicate a high concentration of
25



IuduArts CeizeWung
production in the larger firms, but do not necessarily indicate a very small number of firms. Table
18 presents comparisons of sales CR4's and'CRB's for selected four-digit industries in the U.S. and
Russia.4' In this selected sample alone, there are two four-digit industries in the top decile wit far
more than ten firms - one has 352 firms. Of the 16 industries with CR4's of 60 percent or greater,
there are nine with more than 50 firms. Thus, based on this sample evidence alone, we know that a
table 17 for the U.S. would not have zeros in the lower left triangle. In all of the 350 industries, of
the 53 U.S. industries with CR4's of 60 or greater, there are 19 industries with more than 50 firms,
and of the 10 in the top decile (including those with undisclosed ratios), six of them have more than
10 firms-the average is 121 firms.
The Russian evidence points to three explanations. The first is that the industries in our data
set with few firms have parallel four-digit industries in the defense sector. We know that many
civilian goods were produced by enterprises in the military-industrial complex and, thus, are not
represented in our data.'2 Where this is the case then, the industries actually have many more firms
and are likely less concentrated than our data show. If we had these finms in our data, then, not only
would we find fewer concentrated industries, but we would also see that the match up between
industries and shares of both employment and output is more even. In other words, it is not that
these industries have such a small share of employment and output relative to th- others in our data,
but rather we just do not see how much they actually account for. Looking at uble 18, we guess that
the exclusion of finns in the military-industrial complex probably explains the high concentration and
low firm numbers for household television receivers, semiconductors, and even possibly screw
machine products. However, it probably does not explain the concentration in the women's and
misses' dresses industry.
'As it is not feasible to include a table with all 350 manufacturing industries, we present a
selected sample in 18. To create a "random" sample for comparing the U.S. and Russia, we adopt
the selected industries which Scherer and Ross [1990: 77J, with no intention of comparing to Russia,
present in their text. The U.S. numbers in our table are updated to 1987, however, and a we change
a few of the industries up or down a category in order to exactly match as many industries as
possible. We also reordered the listing so that the 1987 CR4's for the U.S. industries are in
descending order.
'2Another reason why some firms may be missing is that some of these products are consumed by
the enterprises that produce them. Therefore, these products never reach the market. This
explanation is especially important for critical inputs whose delivery was very uncertain under central
planning. To eliminate the uncertainty, enterprises often. developed the internal capacity to produce
their inputs ("universalism").
26



The second explanation is that some industries there were some industries which, Soviet
planners intentionally kept smnal. These industries were probably low priority sectors - light industry
and consumer goods. This explains both the low mimber of firms and the low share of output and
employment. For these industries, then, the conventional wisdom about Russian industrial concentra-
tion does hold, although the high concentration is probably just a consequence of low priority, and
thus low output and few firms, rather than the explicit intent of the planners.
The final explanation is that some industries are smnall because they are relatively new
industries. As of 1989, the planners had not had the chance to build many enterprises in these
industries, plus, with new technologies, they may have been reluctant to invest large amounts
initially. For example, plastic pipe and plastic foam products are two industries that have very few
firms and thus are very concentrated in Russia, but have many firns and low concentration ratios in
the U.S. The closely related four-digit industry, plastic bottles, did not even exist in Russia in 1989.
Here, high concentration is a consequence of youth and, thus, low output and few firms, rather than
the explicit intent of the planners.
Understanding these three explanations is very important. When we can identify the reasons
why an industry has a small number of firms, we can then predict which industries might indeed
suffer from oligopolistic behavior after price liberalization. The new industries likely have good
incentives for entry and sLould not present a problem, but the others, especially those that produce
intermediate goods, likely represent bo'. "anecks in the new economy. For these, the impact of their
concentrto will be greater than their share of economic activity in general.
In sum, there is, in fact, a group of industries which resemble the conventional wisdom about
concentration in the Russian economy. They represent, however, a small share of the economy.
There is another group which appears highly concentrated in our data, but in fact these industries art
augmented by production of civilian goods in the military-industrial complex. Apart from these
industries, and even including those augmented by production in the military-industrial complex,
Russian industry is not highly concentrated when measured at the national level. The vast majority of
industries have enough firms that, with a national market, competition should exist. Although this
seems to contradict the historical evidence, one feature of the command economy does indeed suggest
this result. As we explain later, ministers preferred to keep entire chains of production within their
ministry to minimize reliance on firms outside of their direct control. Thus, for many intermediate
goods, each ministry wanted its own enterprise. For these industries, we should find at least as many
enterprises as there were ministries that used the products.
27



lndasiL Concentrution
Reality 5.6: The-difference between Russian and U.S. concentration arises not in the largest
firms in each industry, but, rather, in the secondary firms in bach industry.
Here we use table 18 to look at specific comparisons between four-digit industries. For many
industries, the U.S. and Russia have very similar concentration ratios, but Russia has only a fraction
of the firms that the U.S. does. For example, the CR4 and CR8 for the Russian storage-batteries
industry are only slightly different while Russia has one tenth as many firms - 13 where the U.S. has
125. For farm machinery and equipment, Russia also has similar concentration ratios, but only has
147 enterprises where the U.S. has 1,576. The U.S. and Russian ratios are about similar for the
metal-cutting machine tools industry, but there are 381 such firms in the U.S. and only 51 in Russia.
Thus, while Russia does have fewer firms in each industry in general, that does not mean that
industries are controlled by oligopolies any more so than in a market economy like the U.S. Rather,
for industries with similar concentration of large firms, the big differences appear in the secondary
finns. In the last example, each secondary (ninth largest or smaller) firm in the U.S. accounts for an
average of .16 percent of sales, whie each in Russia accounts fbr 1.26 percent
These findings are the logical conclusion of the results of the analysis on enterprise size.
Russia is characterized by medium and large firms, while the U.S. is characterized by very small and
extra-large firms. So, we find that, while the large firms in Russia may represent the same share of
their industry that the extra-large firms in the U.S. rpresent of theirs, the remainder of the industry's
production in Russia is filled by a small number of medium-sized firms while that in the U.S. is filled
by a myriad of very small firms. In terms of price competition, these medium-sized firns are
probably more likely to compete with the primary firms given a national market in Russia than the
small ones do with the manmmoth ones in the U.S. This is good news for short-term monopoly
concerns in Russia. The small firms in the U.S., however, offer dynamic advantages to the whole
industrial structure as we discuss above.
Myth 6: The Russian economy is controlled by a large number of dominant
enterprises.
Reality 6: Dominant enterprises do not, in fact, dominate the national economy.
28



Anadzarfa Concenradon
In table 19, we present the results of analysis using the Russian definition of dominant
enterprise. This definition was created in thf context of anti-monopoly policy. An enterprise is
dominant if it commands a 35 percent or greater share of its market. For the purposes of policy,
markets are defined differently for different industries. Here, we start by looking at the national
market. We find that less than 1 percent of all enterprises and less than 4 percent of all employment
are accounted for by finns with 35 percent or greater market shares at the national level. The results
are similar when measured using sales except that dominant fiim sales as a percentage of total is 7.6
percent. Some of these industries have more than one dominant firm. While such industries would
be considered concentrated in the analysis above, these individual firms have less market power than
if only one were dominant. The table shows that, if we ignore industries with two dominant firms,
the share of industries with dominant firms is somewhat diminished. Table 20 shows dominant firms
and dominated industries across branches of production. Four manufacturing branches seem to cover
much of the dominance: electronics, fabricated metal, instruments, and paper. Mining also exhibits a
high proportion of dominant firms and dominated industries.
4.    Barriers to Competition
In the preceding section we have presented evidence that supports our view that, at.the
national level, industrial concentration - the presence of too few firms or of powerful firms - is not
responsible for problems of imperfect competition in Russia. This still leaves open the question of
whether there are structural impediments to competition in Russia. In this section, we argue that
important barriers to competition do exit in Russia. These barriers, however, are not the result of
industrial concentaon, but rather are primarily the result of markets that are highly segmented.43
Under the prior regime, enterprises were highly isolated, divided along both ministerial and,
often, geographic lines." In part, this segmentation can be viewed as a legacy of central planning.
Unfortunately, certain features of the transition environment strengthen these divisions, undermining
3As Ofer [1992: 91] points out "...inertia in distribution links and in supply and marketing
routes, and the remaining main core of production according to ministerial flat may preserve
monopolistic power and produce monopolistic prices." Kahn and Peck [1991: 66-67] also emphasizes
that problems in distribution may create local monopolies.
MTis point was noted in the IMF-World Bank-OECD-EBRD joint study: "Even where more than
one enterprise exists, the national aggregates hide a high degree of regional monopoly power that is
protected by generally poor communications and transportation and by administered marketing
channels which, in turn, are insulated from one another by ministerial lines of responsibility
[1991: 16].
29



lwdased Cenceizron
the efficient distribution of goods. In this section, we discuss the nature of these segmentations, both
ministerial and geographic, in more detail.
4.1    Ministerial Segmentation
The success of central planning relied on the ability of planners and industrial ministers to
retain control over the important decisions of enterprises. The roles of Gosplan (the State Planning
Committee) and the industrial ministries in guiding production and investnent decision making are
well known." Under central planning, the distribution of goods was also coordinated centrally, by
Gossnab (the State Committee for Material Technical Supply). Gossnab was responsible for creating
and managing the wholesale trade system, including identifying appropriate trading partners, setting
the contractual terms of delivery, and arranging for the transportation of goods.
This system was designed to allow planners at Gossnab to control the system of distribution.
Preventing enterprises from developing their own trading links was an important element in limiting
enterprise autonomy and forcing adherence to the plan. Given the sheer size of the task of supply
control, Gossnab planners relied heavily on historic linkages between enterprises when designing their
distribution plans. In many cases, this inertia meant that enterprises were forced to remain in
relationships that, over time, become obsolete due to the creation of alternative and potentially more
appropriate partners.
Moreover, many enterprises were assigned trading partners that were unable tp fulfill their
contractual obligations on a timely basis. This uncertainty undermined the ability of enterprise
directors to meet their production targets and, consequendy, to receive adequate financial rewards.
Unfortunately, the economic and legal structure provided little recourse for the director, pressing him
or her to find alternative sources for important inputs." In some cases, the director independently
developed the internal capacity to produce the needed inputs. In other cases, the industrial minister
took the initiative and established the capacity to produce important inputs, particularly when so doing
'5For more infornation, see Gregory and Stuart [1986].
"One important direction this effort took is the development of informal distribution lines between
enterprises. Most enteprises employed a tolach (expediter) whose job it was to procure inputs
through informal channels. Although these efforts were widespread, they were formally illegal, and
inhibited the development of economy-wide supply information. For more on informal aspects of
plan fulfillment, see Powell [1979].
30



laduri ConcenrtIon
reduced his or her reliance on enterprises outside the ministry.' This latter feature of central
planning alone suggests that, for important commodities, there must be at least as many firms as there
are industrial ministries.
The dominant feature of this system of distribution was the absence of institutions to provide
enterprises with the information they would require to establish links with other firms on a
decentralized basis. In effect, Gossnab and the industrial ministries created a barrier to insulate
enterprises from their trading partners. As a consequence, enterprises tended to become highly
isolated, without knowledge of national and, in some cases, local market structure.
Vertical dependence among enterprises is, to a great extent, the consequence of the arbitrary
demarcation between processes in Soviet industry. Enterprises within an industrial ministry can
usefiully be thought of as processes aiong an assembly line. While there are logical ways of dividing
of an assembly line into its constituent parts, ministers made divisions for reasons of control, rather
than economic rationality. This is of little consequence when the enterprises are subordinate to a
ministry that fits them together. With the collapse of the industrial ministries enterprises are now free
to seek out new customers and new suppliers, but this ability is checked by the arbitrary ways in
which the assembly line is divided. Enterprise directors suppose that they are tied to a vertical chain
that it is very difficult to escape from.
Thus, an important legacy of central planning is an industrial structure that is highly
segmented based on historic trading relationships. We call this type of segmentation ministerial
because it arose out of the ministerial system that included both Gossnab and the industrial ministries.
Unfortunately, the information problem that arose out of the ministerial system continues to persist
today. Currently, much of distribution is organized by wholesale organizations, many of which are
vestiges of the system of central planning. They continue to distribute for the supply organizations
and industrial ministries to which they previously corresponded even when the latter have been
privatized or decentralized. Thus, they act to maintain and reinforce the ministerial distinctions that
arose prior to the introduction of markets. Until new wholesale firms are created to compete with
these firms, old patterns of production and distribution are likely to persist.
"7An extreme example of this "ministerial autarky" occurs with respect to timber. As described in
Hewett [1988: 173]: "Minergo (Energy and Electrification), for example, ships sawn timber produced
by construction firms at the Bratsk and Krasnoiarsk hydroelectric stations in Siberia 3,000-5,000
kilometers away to its enterprises in the European USSR. Simultaneously Minlesbumprom (timber,
pulp, paper, and wood) ships sawn timber to Siberia from its enterprises in the European USSR.'
31



Industri Cencentraton
Ministerial segmentation has important implications for industrial concentration in the Russian
economy. It has produced well-defined and persistent vertical linkages between enterprises, linkages
which, in some sense, can be considered a form of vertical integration. As enterprises re-create the
vertical chains of the assembly line, they represent both fewer and larger producers in the economy.
However, the vertical integration that we observe in the Russian Federation is quite distinct from
vertical integration in western countries. In the Russian Federation, the integration does not take a
legal form nor is it motivated by conventional economic interests." Rather, it is created by a lack of
knowledge of alternative trading partners. In the extreme, one could view the Russian economy as
segmented along historically determined chains of production, in which each firm in the chain may be
acting as both a monopsonist and a monopolist.4'' Once fims become better informed about their
trading alternatives, we can expect some of the more inefficient chains of production to break down.5
4.2    Geographic Segmentation
Many markets in the Russian Federation, which one would naturally expect to be national,
are, in fact, regional or local. In part, this geographic segmentation is a vestige of the system of
central planning. In the prior regime, the production of many commodities of lesser importance, such
as clothing or footwear, was planned by regional, not national, authorities, and thus these enterprises
are only experienced in selling within local markets. Moreover, currentlv the distribution of these
'In the Williamsonian tradition the primary explanation of vertical integration is the reduction in
transactions costs that occur when asset specificity is present. In the case of Russia, however, it is
not asset specificity, but the lack of knowledge of alternative suppliers and customers that creates the
potential for integration.
"Rughvir Khemani suggests that this type of market segmentation offers another reason for
heterogeneity of firm size within industries, that is, a firm in a given industry was established or
maintained at the size necessary for its vertical, or ministerial, market regardle$s of the sizes of other
firms in the same industry.
'CEd Hewett [1988: 170-174] discusses the reasons for strong vertical linkages as well as
physical vertical integration. He concludes, "As a consequence, the successful enterprise is the
vertically integrated enterprise, and the successful ministry, the vertically integrated ministry."
5"To date most of the evidence supporting the presence of ministerial segmentation is anecdotal,
but non-contradicting. in fall 1992, we interviewed 75 firms across western Russia The survey was
conducted by the authors in collaboration with Alan Gelb and I.J. Singh from the World Bank and
Valeriy Makarov and other economists from the Central Economics and Mathematics Institute in
Moscow. These interviews revealed that enterprise directors were often not aware of alternative
trading partners, even when they were known (by the interviewers) to exist.
32



Indual Concentradox
and other goods is arranged by the wholesale trade organizations that we discuss above. In certain
areas, many of these wholesale trade organizations have only single outlets that act as regional
monopolies. As a consequence, the markets for certain goods are highly localized.'2
Naturally, one expects that after liberalization, the size of the markets for these types of goods
will expand quickly. However, certain features of the transition environment suggest that some
barriers to this expansion do exist. Specifically, some regional or oblasr governments have
implemented restrictions on the free flow of irater-oblast trade. Such restrictions have historic
precedents in both legal and illegal activity. Under the prior regime, the transpormtion of goods
between cities required special permits. Any official could stop a truck and inspect its load to
determine whether the delivery was authorized or not. And, if the delivery was, in fact,
unauthorized, the truck driver may have offered the official a bribe to ignore the transgression.
The use of licenses and other regulations to restrict the free flow of trade between oblasts
appears to persist during the transition, although in ways that are presently umneasurable- Much of
the evidence is anecdotal. Many obMast governments have introduced explicit controls restricting the
export of important goods from their region. Typically, these governments still control the local
prices of important consumer commodities and therefore require export restrictions in order to prevent
the flow of these conunodities into neighboring, high-price regions. Private entrepreneurs often
complain that, in addition to these export restrictions, they encounter the extra cost of formnal and
informal tariffs when transporting good across oblast borders. However, the extent to which these
added costs are the consequence of explicit policy is unknown. The failure of the federal
government to invalidate old laws and the activation of new, often conflicting, laws provide local
officials with wide discretion in the enforcement of policy. To some extent, local officials appear to
be using this lack of clarity to collect bribes from firms engaged in the transportation of goods,
although, again, the pervasiveness o. this phenomenon is not known.
Barriers to inter-oblast commerce also are created by problems in the system of
transportation. Unfortunately, the present trasportation system in the Russian Federation was
designed to support the unique institutions of central planning. Certain types of transportation
infrastructure, such as roadways, are presently underdeveloped because they threatened the ability of
govermment authorities to maintain central control over economic behavior.3 For example, nearly all
'2Kahn and Peck [1991: 66] emphasize the importance of regional, as opposed to national,
markets as a barrier to competition.
"Gregory and Stuart [1986].
33



IaduaSd Cenesan
transport in terms of tonkilometers is served by trains and not by trucks. Trains are easier to
administer and control centrally, while trucks require roads. Roads complicate enforcement of
restrictions on internal travel. Moreover, the present system of incentives prevents the efficient use
of available transportation. For example, central control over rail transportation (and the absence of
freight-forwarding institutions) has led to conditions where freight must be scheduled at least six
months in advance.
How important are these barriers in creating local monopolies and oligopolies?
Unfortunately, we do not have direct evidence that enables us to identify which markets have become
localized. Nor do we precisely know the level of localization - the economic region, the oblast, or
the city or town. However, we have prepared some tables to indicate the potential impact of local
markets on the presence of imperfect competition in Russia based on the assumption that geographic
segmentation in Russia is present.
To begin, refer to table 21. Table 21 introduces the reader to the twelve economic regions
and describes each of them in terms of their dependence on particular branches of the civilian
economy. Table 19 shows that, if markets are segmented based on the twelve economic regions, a
larger share of industries have dominant firms. They account for about 10 percent of firms and
about one-third of employment and sales. Ib table 22, we calculate the number of monopolies and
oligopolies that would be present, again based on the assumption that markets are largely contained
within economic regions. )We find that in three major economic regions - Chernozem, Northwest,
and Volgo-Vyatka - regional monopolies and oligopolies may employ nearly half of all civilian
workers or more. It is, of course, not surprising that as we disaggregate on a geographical level that
concentration should increase because, in the Soviet period, industrial location decisions generally
were made on ministerial, not regional lines.?
Much of the anecdotal evidence suggests that markets may even be segmented to a finer
geographic level, to the oblasT level or, in some cases, to the level of the city or town. Clearly, the
smaller the geographic market, the greater the potential for imperfect competition. Unfortunately,
high levels of industrial concentration in localized markets not only lead to higher prices and lower
output, they also exacerbate the existing probiems of highly localized labor markets. During the
Soviet period, the government severely restricted internal migration. Citizens were issued domestic
passports and not permitted free travel outside of their city of residence. Thus labor markets were
"The exception, of course, was the sovnarkhoz - regional planning ministries - that were
implemented by Khrushchev. This experiment ended in 1964.
34



Indi&al Concugrenon
limited to cities or towns except in the cases where the government wanted workers to move, for
example when it enticed workers to the Far East with higher wages. Currently, although the
migration laws have changed, the shortage of housing at current prices and other factors continue to
restrict labor mobility. As a consequence, firms that may be monopolists or oligopolists in their local
goods markets may also be monopsonists or oligopsonists in their local labor markets.
Table 23 presents statistics that describe the average number of civilian firms and industries in
both oblasts and in cities or towns. This table suggests that, while oblasts contain many firms, they
also contain many industries. Thus, although oblasts may be industrially diverse, the presence of
barriers to free trade suggests that many of these industries may be local monopolies or oligopolies.
This feature has mixed implications for reform. With industrial diversity, oblasts are relatively
insulated from economic shocks that affect particular .ndustries, such as shocks to certain types of
heavy industry. On the other hand, the presence of local concentration raises political pressure for
price controls, regulations, and other forms anti-monopoly policy that may impede the progress of
economic reform.
A potentially more important problem is the number of towns in Russia with very few
industries or enterprises. More than 90 percent of all cities and towns in Russia have nine or fewer
civilian firms or industies. Further, as shown in table 24, almost one-half of all cities have only one
firm and more than three-quarters have four firms or fewer. In an ervironment with labor mobility
and little market segmentation, the particular spatial distribution of firms would not pose a serious
threat to reform. Under such conditions, even if the only employer in the town shuts down, workers
can -find employment elsewhere. However, when labor is highly immobile,55 entire cities and towns
are open to potentially large unemployment shocks, if the dominant local industry experiences a
downturn. Under these conditions, workers in these firms are likely to pressure their local
govermnents to intervene and try to find them subsidies, undermining the process of micro-
adjustment.   Although there are many cities with four firms or fewer, these cities accodnt for only
12.2 percent of all civilian employment in industry. Typically, these cities are small towns that
predominantly host small firms. In table 25, we observe that the largest firms are not in cities with
four finms or less. They tend to be in cities with a moderate number of other firms, suggesting that
55The issue, of course, is whether observed low mobility of labor in Russia is due to a lack of
inter-regional employment that resulted from the absence of enterprise failure under the old regime,
or whether this is due to impediments to labor mobility, such as lack of housing. The latter effect
seems to be very important in Russia.
35



IndutarL Conuudou
either larger firms require complementary goods to be produced or the local work force requires a
sufficiently diversified local economy.m
We emphasize that local monopolies and oligopolies in these cities and towns exist only as a
consequence of segmented markets, not as a consequence of having too few firms in their particular
industries. In table 26, we present statistics on the potential competitiveness of industries in cities of
different sizes. We find that very few firms in cities with four firms or fewer are in industries that
are highly concentrated when measured at the national level. In fact, most of the firms in cities and
towns with few firms are in industries with many firms. This fact suggests either that these firms are
in industries that naturally serve only local markets (such as bakeries), or that these firms are in
industries that, with improvements in the system of distribution, are potentially very competitive.
S.    Implications for Economic Reform
The hct that barriers to competition in Russia arise, not from industrial concentration at the
national level, but from ministerial and geographic segmentation of markets, has important'
implications for economic reform. We now turn to the implications of our findings for competition
policy and economic reform.
With respect to competition policy, a whole set of issues arises surrounding the relative
importance and appropriate timing of anti-trust policies. Contestable markets theory argues that the
facilitation of free entry and exit alone should induce competitive behavior in firms 'through the threat
of entry. Even when monopoly power is exercised, many would contend that, in the Russian case,
the conventionally measured welfare losses are less than the benefits from faster privatization due to
the attractiveness of owning monopolies and oligopolies. Moreover, if the threat of entry does not
inhibit monopoly profits, these profits will attract new entry into the market. Thus, the govermnent
should keep its hands off enterprises.
Although entry and the threat of entry are important mechanisms for eliminating the
concentration problems imposed by highly segmented markets, there are nonetheless good reasons for
considering the role of an active competition policy in Russia. Three features of the current
environment suggest that enterprise directors will not perceive a threat for entry or will not care.
Thus, the persistence of non-competitive behavior may warrant competition policy. First, the
S6 It should be noted that'in some of these small cities there may be military enterprises that are
not included in our data set. Such additional firms as do exist would mitigate the vulnerability of
those towns to potentially volatile local markets.
36



indafll Cenmavwitou
enviroment is very uncertain. 'This uncertainty.tends to shrink the time horizon for decisionmakers.
In this case, enterprise directors will substantially discount the future fall in profits which could be the
result of attracting entry or regulation in the current period. Second, directors realize that the lejal
and administrative complexity of starting a new firm acts as an effective barrier against many
potential entrants into their markets. Third, the high cost of capital (to agents not in the state sector)
and the difficulties in acquiring facilities make it very difficult to start up new businesses!,
In certain cases then, an active competition policy may be warranted. It then becomes
important to design the policy appropriately. This clearly depends on the causes of imperfect
competition. We argue that traditional anti-trust policies are inappropriate when the sources of
imperfect competition are ministerial and geographic segmentation.
Anti-trust policy in Russia includes two types of distinct actions. First, the Russian
government has established anti-monopoly committees at regional and local levels. These committees
use product categories and current relevant market sizes to identify "monopolies" - those firms with a
35 percent or more market share - to be regulated. Anti-monopoly price regulations are based on
the belief that this market power comes from industrial stmcture. However, as we have emphasized,
the dominant cause of market power is market segmentation. Not only do price regulations fail to
address the real problern then, they probably exacerbate it by eliminating gains from inter-regional
trade and therefore reinforcing the segmentation. This type of regulation is also vulnerable to a
degree of mismanagement and corruption which could pose a real threat to the process of enterprise
reform in Russia. Any flexibility that local officials have in defining markets provides them with
wide discretion in identifying firms and thus, provides them with a tool to potentially punish any firm
that pursues its own, rather than the government's, objectives.
Second, anti-trust activity often involves breaking up lar!,er firms into smaller ones. In the
case of Russia, this type of action is often discussed in the context of privatization. The idea that
monopolistic or oligopolistic firms should bebroken apart into smaller enterprises is partly based on
the conventional belief that these firns are inefficiently large. Fawever, as we show above, the
evidence does not support this belief. Thus, we question the ability of the government to determine,
ex ante, the appropriate size for firms in an industry. We also question whether the government has
sufficient information to be able to determine, ex ante, whether a particular organizational structure is
suited for market competition.
"7Iis factor is clearly more important in sectors where the minimum efficient scale of the firm is
large. Hence new entry in the retail sector has been quite dramatic.
37



Induafi Cevacanudon
Moreover, when the real probWen is market segmentation, it is not at all clear that breaking
up enterprises will add to the effective number of potential competitors in the market. The breakup
of a large enterprise will not produce several identical small enterprises. In most cases, it will
involve the breakup of an integrated enterprise into its parts. If the market is segmented, this policy
will merely reproduce vertical dependence, as the former constituents of the enterprise will have still
rely heavily on each other. The optimal policy under such conditions must involve measures that
reduce the segmentation of markets (we discuss these types of measures below).
Moreover, both types of anti-trust policies often target specific sectors, in which the effects of
high prices create important economic or political consequences. For example, these policies may
target finns in one part of a chain of production only, such as in light industry. If this is the case,
then anti-trust policy runs the risk of creating worse market structures than those that are currendy in
place. For example, in most cases, firms are engaged in trade relationships which can be
characterized as bilateral monopolies. They purchase inputs from a limited number of firms that, in
urn, sell to only a limited number of customers. Thus, there is a mutual dependency between the
supplier and the customer. If, say, the limited number of customers are divided into many direct
competitors, then a power asymmetry is created, in which many downstream firms compete for a
limited number of upstream supplies. In this case, the relative bargaining power of the upstream
producer is increased and the costs of imperfect competition may be increased. Ignoring the role of
market segmentation in creating imperfect competition can thus lead to anti-trust pblicies that
exacerbate the situation.
The effectiveness of import competition as a remedy for market power depends on two
features of the economy. First, the economy must be open. Second, the economy must have a good
distribution system. Imported goods cannot easily flow into countries lacking seaports, airports, train
stations, and other centers of trade. The absence of such a system in Russia is at the center of our
discussion of imperfect competition. In our view, ministerial and geographic segmentation forces
trading relationships to be backward looking, promoting the maintenance of relationships developed
under central planning, rather than new ones. To promote enterprise-adjustment, a distribution
system must exist that is forward looking. This requires no unnecessary restrictions on the flow of
domestic trade; a good network of wholesale and retail enterprises to link producers with customers;
a good information system to allow firms to identify potential suppliers and customers; and a good
transportation system to move goods from the place of production to the place of consumption; a
good storage system to hold goods, to separate the time of production from the time of consumption;
a good communication system to allow firms to negotiate and modify contracts as needs change; a
38



I dutrc Concentradon
rapid payments and settlements system to facilitate financial compensation for products or services
provided; and, a system of enforceable contract law to enable firms with no history of relations to
contract with one another. Unfortunately, the Russian economy is faced with oblast-level restrictions
on commerce; wholesale- and retail-trade monopolies (and barriers to entry in wholesale and retail
trade); poor infrastructure and poor incentives in infrastructure (information, transportation,
communication, storage); long delays in the payment and settements system; and the absence of an
enforceable system of contract law.
Improvements in the system of distribution would require a combination of investment in
public infrastructure, improvements in the legal system, some privatization, and, most importantly,
elimination of all barriers to free internal trade and free entry.' Moreover, by improving
distribution, the creation of small trading firms will facilitate the entry of new industrial firms." We
would expect that these new firms will become an important source of innovation in the economy.Y
As more small new firms enter, we also anticipate that the size distribution of industrial finms will
change to reflect a more-market-oriented industrial structure.
6.     Conclusion
In this paper, we have presented evidence that calls into question the conventional wisdom
that Russia suffers from excessive industrial concentration. Concentration, measured at the national
level, is not significantly (in an economic sense) greater in Russia than in the United States. While
this indicates that Russia does not suffer from the problem of gigantomania - that is, production
concentrated in a few very large enterprises - national comparisons may obscure important issues
when distribution is costly and information is poor. Indeed, we have argued that the major barriers to
competition that do exist in Russia, arise from market and geographic segmentation.
5' Free entry, to the extent that it leads to the expansion of previously repressed sectors of the
economy - such as services and communications - may also play an important role in mitigating the
consequences of reductions in employment in industry that are associated with restructuring.
54One potential source of new entrants that could play an important role is the MIC. The decline
in orders for MIC output provides incentives for entry into civilian industry. Moreover, the natural
industries to enter are those where profits are high. This suggests that the MIC will be a dynamic
source of competition in Russia.
1DDearden, Ickes, and Samuelson (1990] show that the cost of inducing innovation is increasing in
the amount of hierarchy, and use this to explain the slow rate of innovation adoption in Soviet
industry.
39



laduarful Concamentoo
Analysis of market structure in Russia is important for understanding economic refrmn.
Market segmeniation has created a situation where enterprise directors believe that they are dependent
on a small number of customers and suppliers. This makes the transition environment much more
uncertain, and thus inhibits adjustment to the market.
It is often argued, that privatization of state-owned enterprises will lead to improved economic
performance solely because ownership will provide the proper incentives. It seems unlikely, however,
that without competitive pressures, ownership alone will be sufficient to change behavior.
Enterprises bent on survival will minimize changes that entail significant risks. Consequently, it will
be critical to remove the barriers to competition that exist in Russia. This means that it is crucial to
know the sources of these barriers. Our analysis strongly suggests that marker infrastrucrure -
communication, distribution, and information - is the most important element of a competition policy
in Russia.
40



indusrial Cncaooen
References
Berliner, Joseph, Factory and Manager in the USSR, Cambridge, MA: Harvard University
Press, 1957.
Conynghan, William J., The Modernization of Soviet Industrial Management, Cambridge:
Cambridge University Press, 1982.
Cooper, Julian, "Defence Industry Conversion in the East The Relevance of Western
Experience," working paper, May 1991.
Dearden, James, Barry W. Ickes, and Larry Samuelson, "To Innovate or Not to Innovate:
Incentives for Innovation in ZI'r,rarchies," American Economic Review, 80, 5, December 1990.
Erhlich, Eva, "The Size Structure of Manufacturing Establishments and Enterprises: An
International Comparison," Journal of Compararive Economics, vol. 9, 1985.
Granick, David, Soviet Metal-Fabricating and Economic Development, Madison, University
of Wisconsin Press, 1967.
Gregory, Paul and Robert C. Stuart, Soviet Economic Structure and Performance, 3rd edL,
New York: Harper & Row, Publishers, 1986.
Hewett, Ed A., Reforming the Soviet Economyy: Equality versus Efficiency, Washington, D.C.:
rhe Brooldngs Institution, 1988.
Hurwitz, Elliott., 'Concentration of Russian Defense Industrial Employment by Economic
Region and Oblast, memorandum, May 20, 1992.
Ickes, Barry W., and Randi Ryterman, "Entry Without Exit: Economic Selection Under
Socialism," working paper, 1993.
International Monetary Fund, The World Bank, Organization for Economic Cooperation and
Development, and European Bank for Reconstruction and Development, A Study of the Sovet
Economy, volume 2. Paris, 1991.
Kahn, Alfted and Merton J. Peck, "Price Deregulation, Corporatization, and Competition," in
Whw is To Be Done? Proposals fr the Soviet Transition to the Marker, M. J. Peck and T.
Richardson, eds. (New Haven, CT and London, England: Yale University Press), 1991.
Kered, Michael, "Optimal Tautness and the Economics of Incentives in Bureaucracies,"
Comparative Economic Studies, 35, 1, Spring 1993: 85-117.
Kroll, Heidi, "Monopoly and Transition to the Market," Soviet Economy, 1991, 7, 2, pp.
14-174.
41



InduiWda Coxatrutfan.
Ofer, Gur, 'Stabilizing and Restructuring the Former Soviet Economy: Big-Bang or Gradual
Sequencing?" in Vials of Transition, M. KeFen and G. Ofer, eds. (Boulder, CO; San Francisco, CA;
Oxford, England: Westview Press), 1992.
Scherer, F.M., and David Ross, Industrial Market Stuwre and Economic Performance,
Boston: Houghton Mifflin Company, 1990.
U.S. Department of Commerce, 1987 Census of Manufactures: Concentration Ratios in
Mwuifaatring, MC87-S-6, February 1992.
U.S. Department of Commerce, 1987 Enterprise Statistics: Company Swimay, ES87-3, June
1991.
The World Bank, Rssian Economic Reforn: Crossing the Threshold of Structural Change,
Washington, 1992.
Wiles, PJ.D., The Political Economy of Communism, Oxford, Basil Blackwell, 1962.
42



Table'
1989 densus of Manufacturers
Nuinher.ct..                  hCliuracteftics of.irm
.Industrial Hrms               Epvmn    -                                S
|921.391                       Mean              643           2 Digit       39
TotaLl Eniploynmnt    .        Median            211          3 Digit       180
In Industrial Fimr  .          Minimum             1          4 Digit       406
13,751,839          Maximum       100,605          5 Digit       489
Range         100,604
4374          Variance     4,578,015
t_4m    0hlas                 I Coef of Var   332.82
78
! * -   ..       |I   .: Fum Sazfe  ., *          .*        .-:.  
9IDigit.SIC.  : . Smail  :MbdiuIm                 .:z argw.. lotai.
Agriculture              1411         122         15        3    1551
Mining & Canstruction     307         259        140       25      731
Manufacturing:
food, textiles, wood,
print, chemicals        6518        4049 .    1082        14   11663
ManufacturinT.
rubber, leather
glass, metals, M&E      1410        2750       1304       69    5533
Transpart and Util.'s      238        572         84        1      895
Trade                        1          0          0        0        1
Fire, Insurance,
Real Estate                0           0         0         0       0
Services:
business, repair         430         515         65        1    1011
Services,
health, education           2           1         0        0        3
Public Administration        2           1         0        0        3
Total                   10319        8269      2690       113   21391
Table prepared by A. Brown, B. Ickes, and R. Ryterman.
43



Table 2. Estimates of Mean Employment by Finns in the
Military-lndustrial Complex in Russia for 1989
Number of    Mean Employ-
Sector                        Employment   Enterprses  mnent Per Finn
Civilian Industry                13,751,839      21,391           643
of which Heavy Civilian Industry  4,250,750     2,429         1,750
MIC                               7,979,161       5,309          1,503
Estimated MIC                  9,289,726       5,309          1,750
All Industry                     21,731,000      26,700           814
Sources: Goskomstat Ecnoamic Yearbook for 1990 and PlanEcon data.
44



Table 3. Aggegate Industrial Concentration Pattems in 1985
Average Size of Leading Fi;ms        Leading Company Employment
INumber of Employeesl                as a Percent of Total Industrial
E_nployment fin %)
Top 10       Top 20                Top 10       Top 20
Nation
United States         310,554      219,748                   13.1         18.6
Japan                 107,106       72,240                    7.3          9.9
West Germany          177,173      114,542                   20.1         26.0
United Kingdom        141,156     '108.010                   23.1         35.3
France                116,049       81,381                   23.2        32.5
South Korea            54,416          n.a.                  14.9         n.a.
Canada                 36,990       26,414                   15.3        21.9
Switzerland            60,039       36,602                   49.4         60.2
Holland                84,884       47,783                   84.5         95.1
Sweden                 48,538       32,893                   49.4         66.9
Russia (1989)          62,649       48,133                    4.6          7.0
U.S. statistics based on company data.
Sources: Scherer and Ross (1990, p. 631 and PlanEcon data.
Table prepared by A. Brown, B. Ickes, and R. Ryterman
45



Table 4. Comparison of the Size Distibutions of Russian and U.S. Manufacturing Finns
Size dam by employment
Small        Medium         Large        Ex-large       Total
Caurnry  Statistic                  1-249        250-959         lOUD-        1000
_ 9999         or mar
Russia    Number of firms           9,065         5,662          2,386          83          17,196
Witho &mat&edHMC`         10.374         7,651         4,292          188         22,505
As a percent of total
number af firms           52.7          32.9           13.9          0.5          100.O
in manaucturin*
*Wth estdatedŁfIC          46.1          34.0           19.1          0.8          100.0
US.      Number of firms           299.666        5.530          1,657         267         307.120
As a percent of totl
number of firms           97.6           1.0           0.5           0.1          100.0
in mandacturing
Russia    Numberof workers         974.721       2.874,640     5911.370      1.758.320    11.519,051
with estimatedMtC        1,151.649     3.959.560     11,440,500    4.257.068    20,808,777
As . percent of tatl
number of workers          8.5          25.0           51.3          15.3         100.0
in mnmufacturing
with esanstagd MIC          5.5          19.0          55.0           205         100.0
U.S.     Number of workers        5,777.592    2519.572       4,518.667     8.632.159    21.447.990
As a percent of total
number df workers         26.9          11.7          21.1           40.2         100.0
in mnufacturig
Rusia    Avg.#ofworkes               100  -        508           2.478        21.185         670
U.S.     Avg. Xof workers             19           456           2.727        32.330         70
'U.S. data are fronm U.S. Census 1987 Enterprise Statistcs and are company data.
"See Table 2 and the text for explanation
Table prepared by A.Brown, B. Ickes, and R. Byterman
46



7 ble 5. Com.parison of the Size Distribution of Russian and U.S. Small Manufacturing Firms
Sbe cion by employment
ountry   Statistic                     14          5 9        10-19      20149       50 99      100-249      Total
ussia    Number af firms                   27         205         512       1.386       2A476       4A459       9.065
with xtifiat;dMIC`                27         207         521       1,517      2762        5,340      10.374
As a porcent of tobl
number of small firnm            0.3         2.3         5.6        15.3        27.3        49.2       100.0
in manufacturing
bwth esdmated MIC                0.3         2.0        50         14.6        26.6        51.5       10;0.
.S.      Numberof firns               112,926      58S598      32,158      65.834      18.661      11.489     299.666
As a percent of total
number of small firms           37.7         19.6       10.7        22.0         62          3.8       100.0
in manufacturing_
usia    Number of worker                   76       1,513       7,435      48,645     180,81E     736.237     974.721
i,tb estimafifdAC                 76       1,533       7,576      53,336     202.165     88963   1,151.649
As a prment of total
number of workers               0.0          0.2        0.8         5.0        18.6        75.5       100.0
in manufacturing
small firms
*        with estizatidiC                 0.0         0.1         0.7         4.6        17.6        77.0       100.0
.S.      Number of worker             215.443     394,067     378,180   1,750.874   1.289.853   1,749,175   5.777,592
As a percent of total
number of workrs                3.7          68         6.5        30.3        22.3      - 30.3       100.0
in manufaturing
small finm
-Issia    Avg.#eof workers                  3           7          15         35           73         165        108
.S.      Avg. # of wrken                    2           7          12         27         . .69        152          19
J.S. data are from the U.S. Cansus 1987 Entepripse Statistics and aro company data.
'See Table 2 and the text far explanation.
3b1s prepared by A. Brown, B. lcides, and R. Rytarman
47



Table 6. Comparison of the Size Distribution of Russian and U.S. Manufacturing Firnns
Size dm by employmeit
Country           Statiodo              148          501699     1009249    2S0.99       1000.        ioaao        Total
9999       or more
fluei.    Numbar of firms                  2,130       2,476       4,459       5,662       2,386          83        17,196
witFt extjaadNX"e                2,Z72       Z782        5,340      7,651   4    4,292         188       272505
As a percnt of total
numbr of firm                12.4        14.4       25.9        32.9         13.9         0.5        100.0
in manufacturing
M*tk4w&tad  Mx                    10.1        12.3        23.7       34.0         19.1         0.8        1O0.0
US.       Number of finmn                269.516      18,661      11,489       5,530        1,657        267      307.120
As * percent of htol
numler of firmn              87.5        8.1         3.7         1.8          0.5         0.1        100.G
in manufacturinng
Renem     Humbwrofwmrkwe                  57,669     180.815     736,237   2.074,640    5,911,370   1,758.320   11.519,051
w M       A: estiwatdMiC        62szZi    202,165      886,963   3.959,560   11.440,50a   4,257,068   20.808,777
As a pernt of tol
number af merka                   0.5         1.9         6.4       25.0         51.3         15.3       100.0
in manufactnring
firms
with aujavatad ifX                  0.3        1.0         4.3        19.0        55.0        205          100.0
U.S.      Number of workes              2,73B.564   1,28.853   1,749,175   2.S19,572    4,518,667   8,63Z,159   21,447,990
An a percet of total
number of wrker                   12.8        LO          8.2        11-7        21.1        40.2        1W.0
in m  ufacturing
firms
nue.a    Avg.Dof workser                      27          73         165        508         2475       21185          670
U.S.      Avg.fofw wrke                       10 6a                  152        456         2727       32330           70
US. data an fom U.S. C    1987 Eueiesa Stabis an - cunpan daa
*'Se Tdd 2 and t test fr mnIaiaueL
Tabl ppad by A.Drown, B. ldwks, m R. Ryutmn n
48



Table 7. Size Characteristics of Firms by Industrial Branch.s In Russia In 1989
Number of Firms hr Employment I  e                  Shari o1     Share of Drench Employmauth bSIzr'      Share of
Branch                Small  Medium       Large  Ex-Large     NFirms Total Fitm.   Small    Madlum   Large       Ex-Lurge  Total Em
Agricullura'           1411       122        t5         3        1551         7.3      39.7      19.9     25.4      15.0        1.6
Apparel                 225       445       144         0         814         3.8      5.0      .38.2     56.9       0.0        3.8
Chemicals               103       188       150         5         454         2.1      1.9       14.2     75.0       8.9        4.9
Construction'            40        25         3         0          68 .       0.3      21.6     55.3      23.2       0.0        0.1
Electronics              37       100       118         3         256         1.2      1.1       11.1     78.8       9.0      . 3.2
Fabricated Metal       13        305        72         3         553         2.6      5.9       36.7     44.0      13.4        2.8
.2'                                  9                                      #                 a i,w,,  3 09.c t
furnitulre              110       218        62         0         390         1.8      6.7       42.5     51.8       0.0        1.7
K  ...... :idT 5.4.Y .  w a§fr' ;..TWTh ,..,    1;- .4.S. .. Yi~A
Instfuments             117        95        69         1         252         1.3      3.4    . 17.9      74.6       4.0        2.0
Leather                  64       123        73         0         260         1.2      3.4      24.6      72.0       0.0        1.7
Al.t.                                                                    . -ffV  ,.&IVt
Minlng"                2j7        234       137        25         683         3.1      2.3       9.4      37.8      50.5        8.0
Miscellmneous            96       217        43         0         356         1.7      6.8      54.2      39.0       0.0        1.3
Paper                    38        62        57         0         157        0.7       2.4      14.7   . 82.9       0.0         1.4
Petroleum                20        35        25         2          82        0.4       2.0    - 13.6      62.1      22.3        1.0
Printing               1256       146        22         0        1424         6.7     29.2      42.8      28.0       0.0        1.0
Rubber                   24        89        50         3         176         0.8      1.2      16.9      68.7   '  13.2       2.0
Servicesc    430                  GIs        65        .         1011        4.7       11.3     56.2      29.8      3.0        2.8
..     ;2P: :    . r        ;       .                                   421.7;, .       > 
Tgri'6   r ;       4 r 94;        24>    :2i2   r. ;[c i. ? J 601                                15ii  :"  .62  78 .r;:{ .3    -* 8.
Tobacco                   3        22         3         0          28         0.1      1.2      72.8      26.0      0.0         0.1
Transportation'         238       572        84         1         895         4.2      5.0      54.4      37.5       3.1       3.6
Others?                   5         2         0         a           7          0      31.6      60.4       0.0       0.0       0.0
Total                 10319      8269      2690       113       21391        100       6.6      27.5      48.7      17.3       100.0
*Small: Employ < 200, Medium: 200< -Employ < 1000, Large: 1000< -Employ c 10,000. Large: 10,000< -Employ.
'Non-manufacturing branches.
able prepared by A. Brown, 0. Ickae. and R. Ryrerman.



Table 8. Characteristics of the Largest Firms by Employment in Russia in 1989
Characteristic   Statistic    To 10   Top 25   Top 50   ToP 100 All Ex-large-
Separate Oblasts   Number           7       14       26        37          39
Separata Branches Number            4        4         8       15          15
Separate 4-digit
Industries       Number             4        9       22        40          46
Employment       % of Total       4.6      8.0      11.6    ti6.3        17.3
Mean         62,649   43,966   31,958   22,421       21,027
Employment with   % of Total      4.0      7.1      10.2      14.3       20.8
est rimtedMlC    Mean          92,698   65,561   46,899   33,001       22,073
Employment Share
of Total Sample', Mean           15.8     17.3      17.5      15.5       15.8
Employment Share
of Regional Sample Mean          50.0     53.2      54.1     53.7        55.0
Output           % of Total       4.2      9.8      14.4     21.6        22.9
Mean         1860.7   1734.9   1273.3    952.9        895.2
Output Share of
Total Sample      Mean           16.9   .19.0       18.8      16.0       16.7
Output Share of
Regional Sample    Mean          51.3     52.6      55.1      54.5       55.9
'Exlarge refers to enterprises with greater than 10,000 employees, of which there are 1 13 in the PlanEcon
sample and 217 in the sample with the estimated MIC-.
See Table 2 and the text for explanation.
-The share statistics are the means over enterprises of each enterprise's market share within its 4-digit industry
as measured by the given variable far the given market.
Notes:
46 of the top 50 enterprises represent only four branches. The other four enterprises are each from
separate branches making the total branches represented eight.
In the top 50 there are only six enterprises which are dominant nationally, that is only six with greater than or
equal to 35% employment share of the sample. There are only eight which are dominant in terms of sales.
In the top 50, 35 enterprises are dominant in terms of employment in their regional market, and 33 are doaiinant
in terms of sales in their regional market.
The biggest drip in employment size is between the third and fouth enterprises going from 88969 to 58379.
Only one-third of the extra-large enterprises have employment between 20,000 and 100,605, while the other
two-thirds have employment between 10,000 and 20,000.
Kemerovskaya Oblast has the largest share of very big enterprses; it contains six of the top 25 and seven
uf the top 50..
Table prepared by A. Brown, B. Ickes, and R. Ryterman.
50



Table 9. Aggregate Concenration Shares in the U.S. and Russia
Shr of f100           Sere of 200
Larvgot               Larpxt
Manu_ming             Manufnuti'no
Nation           Siz Meumre                   Cepeaden  An %I    Cepuudu   tin%)
United Stat.s    Domdic vlw uedded               32                     43.2
tl9U21           1nDesde plant suln              31.1                   44.0
Enloymimt in the U.S           23.8                   32.
Ru#ia            Emnplynwnt in nmnufacturin      15.7                   23.4
(1919)
U.S. data are astablishment data.
Sources Scherer and Ross [1990, p. 591 and PlanEcon data.
Table prepared by A. Brown, B. Ickes, and IL RytenMan.
5L
51



Table 1 0. Aggregate Concentration in Russian and U.S. Manufacturing
Perent of manufacturing sales accounted for by-
lumber of   4 largest      8 largest      20 largest    50 largest
Nation           Companies   companies       companies    companies    companies
Russia              17,196          6              9            15            24
(1989)
United States       310,341          9            12            18            27
(1987)                   .v - .                           __. 
U.S. data is company data.
Sources U.S. Census Bureau 1987 Concentration Ratios in Manufacturing and PlanEcon data.
Table prepared by A. Brown, B. Ickes, and R. Ryterman.
52



Table 11. Measures of Industrial Concentration of Firms in Russia In 1909
Value of Statfstic for Industries * with the FollowingNumher of firms
Measure of                                                                                                                        mor
Concentration                Statlslti              1                   3         4       6to 10   it la20  2ItO50  51to 100 than 100
Number of Industiies                                       43        24         1   .:   21. .    80          75        63        34      - 472
Percent of all Industries  Frequency                      1'1.11     6.9       4.7       8.2       19.7      18.5      15.5       8.4:.
Cumulative                         10.6      1G.5      21.2    :25.         48.1      64.8      80.0      88.5    1EB.O.
Number of Firms                                            43        48         57'    .B4         598      1134      2014      2382 .lE3i.
Percent of all Firms  Frequency                            n.2       0.?       0- 3--  0.4         2.8        5.3       9.4      11.1 7.3.3
cumulative                   .       .2       0.4        0 . 0.71d:.         3.9       9.2      18.B      29.7      c.0 -
Labor Force           Percent of Labor Employed
Inumber of workers)    by Industries:
Frequency                       D.23      0.38      0.49? .   0.03      10.52     13.27     18.53     13.94     41982
Cumulative                      0 22;   0.60        10     9     2 1,92   12.44   25.71     44.24     58.18.  100iO.
Charecteristics of Employment
in Firms in Industries:
Mean                            720       1091      1181,., 1556        2420      1609      1265       805 s -
Median                          255       481        271'    .422        694       505       432       3011
Minimum             ~2                      10         7~1                                                 >
Maximum                        7.e0      9421      10278         4    100605    48905 ' 45904    20845 !. 546
Range                          7'E0      9411      10271 E            100597    48904    45898    20841  8454ts
Industries are measured at Ihe 4-digil SIC level.
Table prepared by A. Brown, B. Ickes, and R. Ryterman.



Table 12. Frequency of Firms by Size In Industrial Concentration Classes
Size by                                  Frequency of Firms in Industries with the Following Number of Firms
Employment                        1        2       3        4  5to10 11to20 21to5051tolIOU   >100    Total
Small
IE < 200)                       15       11       24      27      101      203      514      904    8460   10319
Medium
(200< -E< 10001                 1 7      25       16      30      258      489      048     1022    5468    8269
Large
11000< -E< 10,000)              11       12       1B      28      210      354      510      448    I0B7    2690
Extra Large
(E>-10,000)                      0        0                  1    20       28       35       10       18      113
Total                           43       48      67       84      598    1134    2014    2382   15031   21391
Un
size by                                Frequency of Firms in Industries with the Following Four*Firm Concentration Ratios
(mployment               91-100%  81-90%  71-80%  61-70%   51-60%  41-150%  31-40%  21-30%  11-20%  0 10%             Total
Small
I(E<200)                        97       58      g0      133      127      184      284      912    1i29    6699   10319
Medium
(200 < -E < 1000_  138                   87     187      220      277      338      484    1032    1946    3582    8269
Large
(1000< -E <10,0I0)              92       58     151      118      254      187      314      531      513     472    2690
Extra Large
(E> -1I0,IJO)                   8        9        8      14       25        6       32        a        5        0     113
Total                          333      212     422      485      883      713    1114    2483    4193   10753   21391
The four-firm concentration ratio Is the employment of the four largest flirms as a percent of total employment in the Industry.
Tablp prepared by A. Brown, B. Ickes, and R. Ryterman.



Table 13. U.S. and Russian Concentration Ratios Based on SALES
Twu.diglt     Fmur.finu        Eight.fin
Industry    U.S.    Rush       U.S.  Ru.i.
Industry                 _                  Code  (19971    (19138    (15975  il989
FODO ANO KINDRED PRODUCTS                    20       la        11        19       15
TOBACCO PROOUCTS                              21      92        42        97       61
TEXTILE MILL PRODUCTS                         22      17         a        29       13
APPAREL AND OTHER TEXTILE PRODUCTS            23       10        6        13       11
LUMBER AND WOOD PRODUCTS                      24      10         4        14        7
FURNITURE AND FIXTURES                        25      20        11        25       18
PAPER AND ALUED PRODUCTS                     25       25        34        39       53
PUBLISHING AND PRINTING                       21       a        16        14       23
CHEMICALS AND ALLIEO PRODUCTS                28       19        13        26       22
PETROLEUM ANO COAL PRODUCTS                   29      29        41        61       62
RUBBER AND MISCELLANEOUS PLASTCS PROD.        30      16        25        22       36
LEATHER ANO LEATHER PRODUCTS                  31      11        21        16       31
STONE. CLAY. AND GLASS PRODUCTS               32      17         3        26        6
PRIMARY METAL INDUSTRIES                      33      25        30        36       46
FABRICATED METAL PRODUCTS                     34       9        24        13       32
INDUSTRIAL MACHINERY AND EQUIPMENT            35      22        12        28       1a
ELECTRICAL AND OTHER ELECTRONIC EQUIP.        36       19       11        29       19
TRANSPORTATION EQUIPMENT                      37      46        46        60       55
INSTRUMENTS ANO RELATED PRODUCTS              38      29        24        44       34
MISCELLANEOUS MANUFACTURING INDUSTRIES       39        7        10O       10       17
Sources U.S. Bureau at die Caom  Company Summary (19871 and PhoEcan data
Table prepared by A. Brown, B. Icbas and R. Rylhmu
Tahle 14. U.S. and Russian Concentration Ratios Based on EMPLOYMENT
Tw.digit   Four-firm      E_ghft-irm
tndutuy     U.S.   Russ        U.S.  RuB h
Industry                                    Cade   (18971   (18389    (19871   (19891
FOOD AND KINDRED PRODUCTS                     20       9         4        1i        6
TOBACCO PRODUCTS                             21       92        33        91       53
TEXTILE MILL PRODUCTS                         22      19         5        29        9
APPAREL AND OTHER TEXTILE PRODUCTS            23      10         6        13       10
LUMBER AND WOOD PRODUCTS                      24       7         2        10        4
FURNITURE AND FIXTURES                        25      15         9        24       15
PAPER ANO AWED PRODUCTS                       26      20        18        31       29
PUBUSHING AND PRINTING                        27       7         9        12       15
CHEMICALS AND ALLIEO PROOUCTS                 28      19         7        27       13-
PETROLEUM AND COAL PRODUCTS                   29      38        34        66       53
RUBBER AND MISCELLANEOUS PLASTICS PROD.       39      13        17        Ia       28
LEATHER AND LEATHER PRODUCTS                 31       12        13        18       20
STONE. CLAY, AND GLASS PRODUCTS               32      19         2        27        4
PRIMARY METAL INDUSTRIES                     33       20        10        30       30
FABRICATED METAL PRODUCTS                    34        6        16        10       23
INDUSTRIAL MACHINERY AND EQUIPMENT            35      18         9        24       14
ELECTRICAL AND OTHER ELECTRONIC EQUIP.       36       19        11        27       19
TRANSPORTATION EQUIPMENT                      37      38        35        52       43
INSTRUMENTS AND RELATEO PROOUCTS              30      23        14        40       25
MISCELLANEOUS MANUFACTURING INOUSTRIES       39        4         7         7       12
SOurces U.S. Census Bureau Company Summary (19171 and PlanEcon datL
Table prepared by A. Brawn, S. Ickes. and R. Ryterman.
55



Table 15. Measures of Industrial Concentration of Finms in Russia fOr 1989
Value of Statistic for lndusides I in the Following Decilas of Faur-rirm Concantrarion Ratios of EA&PLOYMENT
Afeasure of
Concentraton          SOafista 1solS0  111020%  21 to30%  311o 40% 41 to 505 51 to 60% 61 to 70% 711io 60%  81 to 90%  St to 100% Total
Number el Industries                                      20         31        37        27         31        36        33         38        28        125       406
Percent of            Frequency                           4.9       7.8       L.1        6.7       7.6       8.9        9.1       9.4       6.9       30.8
all Industiles        Cumulative                          4.0      12.6      21.7      28.3       36.0      44.e       53.0      62.3      69.2      100.0
llumber of Firms                                       10153      4193       2483      1114        713       663       405       422        212        333    21319
Percent of            Frequienicy                       50.3       19.6      11.6        5.2       3.3       3.2        2.3       2.0        1.0       1.8
all Firms              Cumulative                       50.3       69.9      81.5 __ 13.7         90.0      93.2       95.5      97.5      96.4      100.0
Number of Employees                                  2944198  2232027  2033250  1868026   741955  1530517    628866   662479   666221    444240 13751839
Percent of             Frequency                         21.4      16.2       14.8      13.6       5.4       11.1       4.6       4.0       4.8        3.2
all Employees          Cumulative                        21.4      37.6      52.4       65.0      71.4      62.5       67.1      91.9      96.6      100.0
Percent of             Frequency                         11.2      19.6       7.7       27.9       6.0      10.2        4.0       6.1       4.2        2.9 441320.4
all Sales .            Cumulative                        11.2      30.8      30.5       66.4      72.4      82.6       86.6      92.7      96.9      100.0
LA
Value of Staltistic for Indusiries ' n the following Deciles ol Fourfirm Concentration Ratios of SALES
Aleasuy, of
Concentration          Stallstic                   0 to 10%   II to20% 21 o130% 311040% 41 to E0% 511060%  61 to 70% 71aO 80% 811090% Sito 100% Total
Number of lodusltles                                       14        19        34        29         30        28         31        41        43        137       406
Percent of             Frequency                          3.4       4.7       9.4        7.1       7.4        6.9       7.6      10.1       10.6      33.7
all Industries         Cumulative                         3.4       8.1      16.5       23.0      31.0      31.9       45.6      55'7      66.3      100.0
Number of Firnis                                        0576      4515       2691      1484       1419       603       516        726       442        419    21319
Percent of             Frequency                         40.1      21.1       12.6       6.9       6.8        2.8       2.4       3.4        2.1       2.0
all Firms              Cumulalive                        40.1      61.2      73.8       80.7      97.3      90.2       92.8      96.0      98.0       100.0
Number of Employees                                  2664951   1476631   1902195  1270028  2126020   65241l  1051921   901732  1005634    500268 13751839
Petcent of             Frequency                         19.4      10.7       13.8       9.Z      15.5        6.2       7.6       6.6        7.3       3.8
all Employees          Cumulative                        12.4      30.1      43.9       53.1      68.6      74.6       82.5      89.1      95.4      100.0
Percent of             Frequency                         21.3       5.0       6.4       21.0      16.4        5:3       7.1       5.4       8.9        3.1  441320.4
all Sales             Cumulative                        21.3       26.3      32.7      53.1       70.1      75.4       82.5      87.9      96.8      100.0
'Indusities are measured at the 4-diglt SIC level.
Four-lirm concentration altiOs ere the sum of the statistic for the four largest firms as a percent of the total of the stalislic for the industry.
Table prepared 4y A. Brown, B. Ickes. and n. Ryterman.



Table 16. Distribution of U.S. and Russian Manufacturing Industries by Four-Firm  Ratias
All Ratiosby Sales
Four *Firm
Concentration                             Percentage of all      Percentap of       Percentae       Perantp of
Ratio Range    Number of Industries            Industries      Tota value added       of Output      Employment
U.S.       Russia      U.S.     Russio           U.S.              Russi           Rusia
0.19                86          25         19.2       7.1            21.7               18.7            28.8
20-39               163         52         36.4      14.9            38.8               17.2            24.0
40459               120         46         26.8      13.1            19.7               33.3           222
60-79               56          68         12.5      1rq.4           14.9               16.3            13.2
80.100              23          159        5.1       45.4            4.9                14.6            11.8
Total               448         350        100        100            100                100              100
Bassian Ratios by Employment
Four -Firm
Concentration                             Pecenage of Ill       Percentage of        Percentage     Percentage of
Raio Range     Number of Industries            Industries      Tal value added        of Output      Employment
U.S.       Russia      U.S.     Russia           US.               Russia          Russia
0-19                86          39         19.2      11.1            21.7               24.1            37.0
20-39               163         56         36.4      16.0            38.8               35.7            30.5
40-59               120         56         20.8      16.0            19.7               lS.0            13.4
50.79               56           63        12.5      19.0            14.9               11.3             9.8
800100              23          136         5.1      38.9             4.9               10.9             9.2
TotaI               448         350        100        100             100               100              100
U.S. data are for 1982 and ara establishment data.
Sources: Scherer and Ross 11990, p.831 and PlanEcon data
Table prepared by A. Brown. 8. Ilckes, and R. Rytenman
57



Table 17. Frequency of Firms by Firm-Indusury and Concentration Ratlo
Number of Firms                          Incidence of Firms in Indus_ties with the Following Four-Firm Concentration Ratios of Employment
in industry           0lto 100%  81 to 90%  71 to 80%  B1 to 70%  61 to 80%  41 to 50%  31 to 40%   21 to 30%  11 to 20%    0 to OV%       Total
1                            43          a           a         0           0          U          a          0          0          0         43
2                            48           0          0          0          0          0          0          0          0           0        48
3                            57           0          0          0          O          O          O          .0         0           0        57
3       __________ ______________________ .                ____
4                            84           0          0          0          0          0          0          0          0           0         84
Ln _
5 to 10                     101         175        181         72         69          0          a          0          0           0       598
lto2O                        0         37        212        204         239        282         60          0          0          0       1134
21 to50                        0          0         29       .119        286        421        671        498         so           0      2014
51 to iOO-                     0          0          0          0         89          0        338       1335        528         96       2382
> 10.0                        0          0          0          0           0          0        147        650      3577       10657      15031
Total                       333        212        422        485         683        713       1114       2483       4193      10753      21391
1ndustries are measured at the 4.diglt SIC level.
Thealour*lirm concentration ratio is the employment In the four largest firms as a percont of total employment in the Industry.
Table prepared by A. Brown, B. Ickos, and R. Ryterman.



Table 18. Sales Concentration Ratios for Representative Industries
for the United States (1987) and Russia.(1989)
4&Firm bEe         8 irm Ratio      |Nubsr of Firm.
S1C Code   Industry Descriptien                  U.S.      Ruia   I U.S.       Susie  lU..         Rusio
2067          Chewing gum                            96                 10o                  8
33310         Ptimary copper'                        92        54       120       85         7         12
2111          Cigarenes                              92        30       IDI       59         9        24
3641          Electric lamps                         S1        77       94        94        93         12
3711          Passengercars                          90        84        95       90        352        10
2043          Cereal breakfast foods                 17                 99                  33
2082          BOr and malt beverages                 07        13        98       22        101       237
3632          Household frif eortors and freezers    i5                 98                  40
3211          Ratcans                                82        63       (DI       74        65        29
3511          Turbines and tumibe generators         90        91       35        99        68         10
3221          Glass containers                       7         33        u9       53        35        36
3334          Primry aluminum                        74        66        95       92        34         11
3721          Aircraft                               72                  92                 137
3011          Tres and inner tubes                   69        60        87       92        114        10
2841          Soap and detergents                   E5         76        76       95        683        11
3691          Storage batteries                      64        57        78       81        125        13
3562          Ball and roller bearings               59        53        69       84        113        19
3411          MetaI cans                            54        100        70       100       161        2
2822          Synthetic rubber                      50         65        76       92        55         10
3144          Women's footwear, excpt athledic       50                  61                 123
(3140) Footwear                                         25                 37                  111
3523          Farm machinery and equipment          45         42        52       59       1576       147
3312          BlasxfumaacesandsteelniUls            44         46        63       71        271       36
2041          Rour and other grain mils              44        14        63       23        237       235
n11           Cotton weaving niuls                   42        18        59       28        246       122
3574          Seniconductors                        40        100       59        100i      755        2
3651          Household audio and video equip.       39       1WO        59       100       360        1
3621          Motors nd generators                   36        30        49       49        349       48
2051          Bread. cake, and related products     34         8        47        11       1948       1467
3965          Fasteners buttons. etc.               33         80        43       92        247        16
2873          Nitrigenous ferlizer                  33         49        55       79        117        13
2911          Petroleum refining                     32        42        52       65        200      , 31
3541          Metal-cutting machine tools            31        28        41       46        381       51
2066          Bottled and canned soft drinks         30        33        40       45        646       76
3241          Portland cement                        28        24       47       **40       123       42
2851          Paints and allied products             27        74       40        84       1121       .61
2653          Corrugated and sold fiber boxes        26       10D       41        100       952        4
2711          Newspapers                            25         54       39        65       7473       32
2834          Pharmaceutical preparations           22         33       36        52        640       63
2026          Fluid ilk     *                        21        10        32       15        652       472
3552          Textile machinery                      20        57        30       81        475       17
3452          Screw machine products                 16        99       24        100       834        5
2421          Sawmills nd planning nils             is        210       21        32       5252       199
3273          Ready-mixed concrete                   8         28        11       45       3749       40
2335       _Wmensandmisses' dresses                  6        100        10       100      5398        3
U.S. da  are establishment dna.
IDI Withheld to avoid discDsing data for individual cnvt anies.
*U.S. statistics are for 1982 from Scherer and Ross
Sources Scherer and Ross 11990. p. 771. U.S. Burmau of Census 1987 Concentration Ratios in Manufacturing, and PlanEcon date
Table Oreoared by A. Brown. B. Ickes. and R. Rvternan    59



'Table 19. Frequency of Dominant Firms within Industies in Russia in 1989
#as % of  su  as % of
Measured in employmert             # of fimns  nat'l total  nat'l sum
National
Fimrs with > -35% of market in their industry    . 173       0.81        3.8
Of those, firms in industries where only
ons firm has more than 35%                135       0.63        3.5
Regional
Finns with > -35% of market in their industry    2122        9.92       34.9
Of those, firms in industries where only
one firm has more than 35%               1634        7.64      27.8
Measured in sales
National
Firms with > - 35% of market in their industry     203       0.95        7.6
Of those, finms in industries where only
one firm has mare than 35%                163        0.76        6.7
Regional
Firms with > -35% of market in their industry    2189       10.23       37.9
Of those, fims in industries where only
one firm has more than 35%               1751        8.19      30.8
*Industry is measured at the four-digit SIC level.
Table prepared by A. Brown, B. Ickes, and R. Ryteman.
60



Table 20. Concentration Characteristics of Firms by Industrial Branches in Russia In 1989
FIRMS with > - 35%; Employment Share In their 4dig1t Int,       INDUSTRIES wlth firms with > -35% Employment Shares
As ftarof      As Sharo1 of   In Ind       #of 4digit            As Shaerof  #firms In  AiShaerof    As Share of
Branch            #Fi        rms      Branch irs BaeEm    wt only I     Indusre   #lofInd    Bronch Ind   theso Ind   Branch Firms  Branch Em
Agricultuie'          5           0.3            0.4           3            8           4         50.0         is           1.2           0.5
Apparel               10           12            3,3           8           19           a         421          26           31            4.6
Chemicals             5            11            50            6           27          5           I           35           77           113
Construction'         0           0.0            0.0           a            I           a          0.0          0           0.0          0.0
Food                  7           011            0 5           B           35           B          17.1        10           0.2          0.6
Furniture             4            1.0           0.4           P            9          4          44.4          6           1.5          0.5
Ind M&E               Is           1.5           6.2           14          44          Is         34.1         93           8.5          11.2
Lumber                4           0.2            0.3           2           14          3          21.4          4           0.2          0.3
Mining'               12           1.8           10.9          12          27          12         44.4         94          14.2          22.9
Miscellanuour         8           2.2            5.1           a           16           a         50.0         31           8.7          8.7
Petruleum             2           2.4            1 2           2            5          2          40.0         11          13.4          2.5
Primary Metal         10          4.4            3.3           8           19          9          47.4         25          11.1          5.3
Printino              9           0.6            4.0           7           13          a          61.5         30           2.1          8.8
Rubber                4           2.3            1.1           2           10          3          30.0          a           3.4          1.2
Services'             0           0.0            0.0           0           11          0           0.0          0           0.0          0.0
Sltone C&G   ,                     0.3           1.5           4           26          5          19.2         40           2.0          3.5
Textile               10           1.7           2.5           6           21           6         39.1         27           4.5          4.3
Tobacco               2           7.1            3.4           2            3          2          69.7          4          14.3          4.5
Transpart Equip       3           0.9            1.7           3            9          3          33.3         10           2.9          3.2
Trinspattatian'        I          0.1            0.1           1            a           1         16.7          1           0.1          0.1
Others                3           42.9          77.9           3            3          3          100.0         7          100.0        100.0
'Non-manufaclur ing branches.
Ta bI prepared by A. Brown, 9. Ickes, and R. Rytermen.



Table 21. Concentratlon of Employment In Brenches Across Regions for Enterprlsns In flussla In 199
llow %I
Crlumn  %len
Share ol
irantec           Central     Chernotem   E Siberia    Far East    Kaliningrad  N Caucasus North       Northwest  Urtals        Velga       Vyatka      W Siberis   Total E
Agriolturst*       i0.9.        3.3          6.9        18.3         8.?         4.5         17.0         2.6        10.2         5.7         5.8         5.3
0.E         1.1         1.9         7.1        23.7          0.9         5.6         0.9         1.1          0.8         1.5         0.9       1.6
Apparel            '29.6        4.2          3.9         3.6         0.5         13.7    .   2.4          6.8        10.7        10.3         7.6         6.6
4.7         3.2         2.5         3.1          2.9         6.9         1.8         4.9         2.7          3.5         4.4         2.7       3.8
Chamicahs          21.6         5.9          5.5         0.D         0.0         1.7          1.4         4.8        13.0        19.5         9.4        10.5
4.5         6.6         4.5          1.0         0.2         4.3         1.4         4.3         4.3          8.51        7.1 __    5.5         4.9
Construction'      16.8         4.2          3.8         4.1         0.5         3.0         14.3         6.0         8.2         4.1         3.5        31.3
0.1         0.1         0.1         0.1          0.1         0.0         0.4         0.1         0.1         0.0          0.1         0.4       0.1
Eleelroriles       24.1         9.5          4.1         1.2         1.0         8.5         0.1          8,6        13.3         9.3        11.6         10.6
3.3         8.1         2.2         0.9         5.4          2.4         0.0         5.3         2.9          2.7         5.0         3.7       3.2
Fabricated Metal    21.6        7,2         3.9          3.2         0.0        10.1          0.9        12.4        15.4         9,2         9.4         6.8
_______ __  2.6      4.0         1.9         2.0         0.0 _        3.2         0.5         6.6         2.9         2.3         4.0          2.0  *   2.8
Food               17.3         8.8          4.7         8.7         0.8        14.5          3.6         4.6        10.5        11.3         5.3         9.6
_______________          7.1        16.9         7.6        19.3         12.6        15.9         7.4         8.4         6.7         9.7          7.9         9.5       9.6
Furniture          25.1         3.6          5.1         3.7         0.6        17.6         3.3          5.1         8.4         9.0         6.6         8.8
1.8         1.2         1.4         1.4         2.1         3.4          1.1         2.8         0.9       . 1.4         1.7         1.6       1.7
Ind M&E   .        27.3         6.7          2.8         1.6         0.4         10.5         2.1         5.9        15.5        13.4         5.4         8.4
15.8        18.2         6.4         4.9         6.2         16.3         5.8        15.3        14.1        16.3        11.3        12.2      13.6
Instruments        43.0          1.8         1.7         0.5         0.6         8.8          0.1         8.5         8.9        14.3         6.2         5.71
3.7         0.7         0.0         0.2          1.9         2.0         0.0         3.3         1.2         2.6          1.9         1.2      2.0
Leathar            2S,8         5.6          2.8         2.3      0 .4          13.3          1.0         9.3        13.86       11.6         6.4         6.0
1.EI        1.9         0.6         0.9         1.2  _      2.0          0.3         3.0         1.6         1.0         2.2          1.1       I.7
Lumbar             11.7         0.9         16.7         6.1         0.3         2.6         19.2         4.5        14.6         4.4         7.7         9.1
3.4         1.3        19.1        12.5         3.7          2.0        26.1         5.9         6.6         2.7          5.0         6.6       6.8
Mining              6.8         3.4          9.5         7.6         0.2        IIA         10.6         2.2        18.3         3.7         0.8        25.6
2.3         5.4        12.8        14.1         3.0         10.5        17.1         3.4         9.9         2.61         1.0        22.1       8.0
Miscellaneous      37.1         2.2          2.1         1.0         0.0         10.3         4.1        10.0        11.4         4.5        14.2         3.1
2.1         0.6         0.5         0.3         0.0          1.5         1.1 !       2.5          1.0        0.5          2.8         0.4       1.3
a.,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1.3



Table 21 Cant.
Shari o1
Otanvh           Cenlral     Chernozem   E Sibetla    Far East    Kallningrad  N Caucasus North       Northwest  Urals        Vel         Vyatka      W Siheria   Total E
Paper              14.1         0.7        11.4         6.9         4.6          1.8       25.0         12.5        11.8        4.4         6.2         0.6  --
__________O.9                  _0.2     _2.7            2.1        10.9         0.3         7.2         3.4          1.1        0.6          1.4        0.1   _1.4|
Petroleum          14.2.        0.1        14.5          1.8        0.0          8.3         0.8         3.2        28A4        14.8        5.1          8.8
0.l       -0.0          2.4         0.4         0.0         0.9         0.1         0.6         1.9         1.3         0.8         0.9       1.0
Primary Metal      15.0         6.8         4.3          1.5        0.0          3.1         5.8         1.5        43.3         5.1         4.3         9.4
___________           3.9         6.3         4,5       _2.1          0.4          2.2 _     _7.1     _ 1.;           17.7        2.1         4.1          6.2 _    6.1
Plinliiig          41.3         3.3         3.1         3.2         0.5          6.7         2.7         9.6        10.6         8.6         3.6         8.3
1.8        0.7         0.5          0.8         0.9         0.8         0.1         1.9         0.8         0.8         0.6         0.7       1.0
Rubber             29.0         8.1         2.8         0.5         0.0         5.9         0.2         68.8         .2        20.0         5.9        10.9
___________2.E                  3.2         0.9         0.2         0.0         1.3         0.1         3.4         1.1         3.6         1.8         2.3       2.0
Serwice.,          25.6         4.3         6.8         5.0         OA         .10.5        2.1         5.3         15.9       10.1:        4.6         9.4
3.1         2.4         3.3         3.3         1.8         3.4         1.2         2.9         3.0 _       2.6         2.0         2.8       2.8
Slone C&C          26.5         5.7         8.0         4.8         0.3         9.1         3.3          6.7        14.5  .     11.6        4.6         8.9
__________ _8.1  8.5    7.8         6.3         4.0         7.e         4.9         8.1         7.3         7.8         5.3         7.2       7.5
TextIl e           55.2         3.2         4.0         0.7         0.3         7.3          1.3         5.3        6.5         7.8         4.1         4.2
.14.         3.9         4.2         1.1         3.2         5.2         1.8         6.3         2.7         4.3         4.0         2.8       8.2
Tobacco            23.7        10.5         2.2         0.0         0.0         18.2        0.0         13.5        8.1         7.8'        0.2         7.7
0.1         0.4         0.0         0.0         0.0         0.2         0.01        0.3         0.1         O.1         0.0         0.1      0.1
Transport Equip    25.5         1.9         3.6         3.7.        0.9         4.8         2.2          1.7       1OA 47.1                 15.7        2.6
7.9         2.7         4.4         6.1        11,3         4.0         3.2         2.4         5.0        17.6        17.5         2.0       7.2
Transportation     15.2         4.3        11.9         9.1         0.4          6.7         7.4       '3.5         14.6       10.2         5.1         11.7
2.3         3.1         7.2         7.5         2.2         2.6.        5.4         2.4         3.5         3.3         2.8         4.5       3.6
Othets             38.1      0 .0        0.                0.       0.0 .      30.2         0.0         .0  0       68.        13.8         0.0         8.5
___________0.0                  0.0         0.0         0.0         0.a         0 .0    _ 0.0           D.0         0.O         0.0         0.0         0.0       0.0
Share of total E   23.4         5.0         6.9         4.3         0.6        897          5.0         5.2         14.9       11.2         6.5         9.3           IOO.D
'Non-manulacturlng branches.
Table prapated by A. Brown, S. Ickes, and R. Rytorman.
a'~~
w



Table 22. Measures of Industrial Concentration Across Regions In Russia for 1989
Firm-lndustries'                                              Firm-lndustries
Region        StalTistic               1       <-4                Region      Statistic               1       <-4
Central       % of Industiles           19.0      49A             North       % of IndustrIes          39.2      68.4
Yof Firms                 1.4        7.6                       % of Firms                4.7       14.2
% of Employment           1.7       19.9                       % of Employment          19.8      36.9
Chernozem     % of Indusities           38.8      71.3            Northwest  % of industries          37.0       77.2
% of Firms                5.9       19.1                       % of Firms                8.5      30.9
% of Employment          15.5       49.0                       % of Employment          17A       52.0
E. Siberia    % of Industries           40.1      73.1            UralS       % of Industries         27.1       61.0
X of Firms                4.R       15.1                       Y of Firms                3.0      13.2
% of Employment          10.7       38.7                       % of Employment           8.2      29.5
Fat East      X of Industiies           42.1      70.1            Volga       % of Industries*        27.5       63.0
X of Firms                5.6       15.0                       X of Firms                3.0      14.1
X of Enployment          12.4       24.3                       X of Employment           7.3      41.1
Keliningrad   % of Industries           62.1      84.8            V Vyatka    X of Industries         38.8       71.1
X of Firms               27.0       62.8                       X of Firms                5.3       18.6
X of Employment          26.3       70.5                       X of Employment          15.1      52.5
N. Caucasus   X of Industries           33.1      55.7            W. Siberia   % of Indusiries        32.2       65.2
X of Firms                3.9       14.1                       X of Firms                3.7      13.9
X of Employment           B.9      25.3                        % of Employment           8.5      29.5
Column lists the value of the statistic for industries with the given number oa firms in that industry in that region.
Industries are measured at the 4-digit SIC level.
Table prepared by A. Brown, B. Ickes, and R. Ryterman.



Table 23. Geographic Distribution of Firms and Industries in Russia fot 1989
Value of Statistic for OBLAST$ fn Each of the Foalowing Decfils'
Unit      Statistic     1         2         3         4         6         a         7          8         9        10       Total
Firma     Mean           67.6      129.3     177.0    203.5    232.3       263.7    304.8      379.9    439.4    655.3
Minimum          27        98       157       l8e       221       247       288       340       404       517        27
Maximum          97       155       187       217       244       277       328       402       507       898       898
Range            70        57        30        29        23        30        40        62        103      381       871
Industries  Mean         33.5      55.5       63.1      69.8      81.6      85.7      95.0     114.5    128.6    179.8
Minimum          18        44        60        07        77        89        91       106       125       135         18
Maximum          41        B0        B8        74        85        91        103       123       132      234       234
Range            23        16         a         7         a         .5        12        17         7       99       216
Value of Siatlstif forCIT/ESin Ehch of the Following oDed11s
Unlt   StatIstin     1         2         3          4         5         0         7         9         9         10      Total
Firma     Mean            1.0        1.0       1.0       1.0       1.3       2.0       3.0       4.5       7.2     105.0
Minimum           1         1         I         1         1         2         Z         4         5         9          1
Maximum           1         1         I         1         2         2         4         5         9       768       788
Range             0         0         0         0         1         0         2          1        4       759       767
Industries  Mean            1         1         1         1        1.2        2        2.9       4.4       0.9      54.1
Minlmum           1         1         1         1         1         2         2         4         5.        9          1
Maximum           1         1         1         1         2         2         4         5         9       234       234
Range             0         0         0         0         1         0         2.         1        4       225       233
Industries are measured at the 4-diglt SIC level.
Each decile contaTns 10 percent of Russian cities or ablasts ranging from smallest to largest based on the unit of observation being analysed.
For example, when analyzing the geographic distribution oa firms In cities, cities are ranked based an their totil number oa firms.
Table prepared by A. Brown, B. Iekes. and R. Ryterman.



Table 24. Characteristics of Firms in Russian Cities by Firm and Industry Concentration for 1989
Value of St tittic for Cilia. wiih the Foliowing lNmter of FIRMS
Attrliute        Statistil     1         2         3         4      6tol     11 ta 20  2Itc 90  61to 100 101 to 200   >200
Nwnmbr 4f Total Clik .  Number         2097       576       356       292       693      228        92        33         15         2
Poment of Total Citles    Frequercy     47.9      13.2       9.1      8.4      15.8       6.2       2.1       0.8        0.3        0
Cumulative      47.9      81.1      69.3      75.7      91.6      96.9      98.9     99.6        I00       100
Number ofloslFlmim     Numbef          2097      1152      1069      1128      4660     3246      2760      2334       1632      1114
Poecent of Toltl Firma  Frequency        9.9       6.4      6.0       5.3      21.8      15.2      12.9      I0.9       98.       5.2
Cuniuf liv.       9.8     16.2      20.2      25.5      47.2      82.4      75.3     96.2        94.9     tOO
EmploymentbyFrlms      Men            335.9    319.2    286.7    274.3    361.2    659.3   1030.8    940.0    1108.B   1051.2
(numbeor)              Median           179        90        64        62        76       122      148       128        213       645
Minimum            I         1         3         1         1         1        2         4          5         9
Meximum         7157      6511    17784    40960    25525    30t92    99980    33235    10605    59379
Reng            7156      8510    17781    40959    25524    30091    99959    33231          100600    58371
Varilane      209203   424810   959389  1919212  1014338  2890801 13736589  U95489  12880984  5450188
Caef Var       138.2    204,2    341.6    491.8    278.9    267.9    359.8    226.8             320.9    222.1
Ori                   Peceint of Total       Frequeney         6.1       2.7       2.2       22       12.2 i    15.6     20.7      IC.0       14.8       8.5
Employment             Cumulative        5.1       7.8     10.0      12.2      24.4      40.0      60.7      78.7       91.5    108.0
Value of Sfo tutu for Citl J H'th th Folo Wing Numberof INDUSTRIES
Attribute        Statitle      1        2         3         4          to 10   11to20  21to60 S1 to 100 OlOto200  >200
Numbor of Tatil Cilhs    Number        2125       573       359       291      690       228        el        36          1         1
Percent of Total Ciltie    Frequency   48.8        13       8.2       8.4      15.9       5.2       1.9      0.8          0         0
Cumulative       48.6     61.7      69.9      78.3      92.1      97.3      99.1      100        100       100
Number of TolFinmn    Number           215S      1169      1105      1151     4791      3499      2947      3470        348       769
Percent of Total Firms    Froquency     10.1      5.5       5.2       5.4      22.4      16.3      13.8     16.2         1.0      3.6
Cumulaltlv       10.1      15.5     20.7      28.1      49.5      64.9      73.8     94.9       96.4       100
Employment by Flims    Mean           336.7    338.4    283.7    276.0    370.0    693.6   101?.9   1055.0    1115.9   1022.1
(number)               Medlin           182        65        82        61        78      123       123       178       1047       242
Minimum            I         1         3       .1          1         1        2         4         10         8
Maximum         7157    16930    17784    40960    25525    32348    99960   100605            13354    59379
Ringo           7159    18929    17781    40959    25524    32347    99959   100601            13344    59371
Vauiance      205945   701343   638042  1805830  1013561  3289613 12741143  9147E80   2622892  8726288
Coalf Vr        134.8    247.5    302.9    489.9    272.9    261.9    350.7    286.6           145.1    253.7
Poercnt of Total      Frequency         6.3       2.9       2.1       2.3      12.9      17.8      21.8     26.6        2.9       5.7
Employment             Cumulative       5.3       8.2      10.3      12.6      26.5      43.1      64.9     91.6       94.3    100.0
Table prupared by A. Brown, B. Ickes, and R. Ryterman.



Table 25. Employment Size of Firms In Cities by Firm Concentration ir Russia for 1989
Value of tho Statistic for Cities with tho following Numher ofFirms
1        2         3         4      Sto 11  11 to20  21to50  51tol0 101tto200   >200                Total
Number of Cities               2097       578       356       282       693      228        92        33         15         2     4374
Number   Size by
of Firma Employment
Small                1105       737       802       962      3105      1465      986       728        447       204    10319
IE < 2001
Medium                866       340       215       217      1215      1335     1392      1129        968       592      8269
(200<-E< 10001
Large                 126        75        49        48       330      421       469       460        400       312      2690
11000< -E< 100001
Extra Large             0         0         2         1        10       25        33        19         17         6       113
(E < - 100001
Total                2097      1152      1066      1128      4660     3248      2760      2334       1932      1114    21391
Tabl prepared by A. Brown, B. Ickes, and R. Ryterman.



Table 26. Frequency of Firms by Industrial Concentration Classes in Cities By Firm Concentration Classes
Number of firms               Firm Frequency In Cities with the Following Number of Firms
in industry                 1        2       3       4   to 10 Ill to20  21 to 50 61 to 100 101 to 200  > 200    Total
1                           3       2        1       0       6        a        5        6         4      10      43
2                           4        1       0       2        1      11        6        3         7       13      48
3                           2        1       0       0        6       7       13        4         9       15      57
4                           7       0        3       2       16      13       14        9         6       14      84
5to lo                     43      14       15       9       67      83       89       85        78      115     598
11to20                     93      43       24      24     148      166      182      149       173     134    1134
21 to 50                  147      69       43      47     270      288      306      349       311      194    2014
51 to 100                 231      99       78      88     381      310     364       376       282      173    2382
> 100                   1687      933     904      958    3767    2362    1781       1353       962     445   15031
>   Total                2097    1152    1068    1128    4660    3248    2760        2334      1832    1114  21391
4*Firm Concen                 Firm frequency in Cities with the Following Number of Firms
tration Ratio               1       2        3       4  5tolO  llto20 21to50 Sito 100 lOlto200  >200    Total
91 to lOO%                 23       B        5       5      33       51       48       37        42      85      333
81 to90%                   17      12        B       6      28       24       32       30       t28      31      212
71 to 80%                  29      11        7       6      G8       55       68       51        55      72      422
G1 to 70%                  43      13       17      13      63       78       62       71        67      58     485
51 to60%                   50      24       18      19      79      108      117      110        87      73      683
41 to50%                   65      34       16      12      g0       87      112      105       125      68      713
31 to40%                  109      34       25      42      173     171      182      167       161      70    1114
21 to30%                  176      58       el      45     309      323     422       469       383     237    2483
11 to20%                 384      184      120     147     732     745      684       596       411     194   4193
Oto 10%                  1191     798      788     833    3087    160B    1055        698       473     226   10753
Total                    2097    1152    1068    1128    4660    3245    2760        2334      1832    1114  21391
the four-firm concentration ratio Is the employment in the four largest firms as a percent of total employment in the industry.
Table prepared by A. Brown, B. Ickes, and R. Ryterman.



Policy Research Working Paper Series
Contact
Title                           Author                 Dale              for paper
WPS1319 The Financial Systemn and Public  Astl Demirgflg-Kunt   July 1994          B. Moore
Enterprise Reform: Concepts and    Ross Levine                           35261
Cases
WPS1320 Capital Structures in Developlng  Asil DemirgOu-Kunt    July 1994          B. Moore
Countries: Evidence from Ten    Vojislav Maksimovic                      35261
Countries
WPS1321 Institutions and the East Asian  Jose Edgardo Campos    July 1994          B. Moore
Miracle: Asymmetric Information,    Donald Lien                          35261
Rent-Seeking, and the Deliberation
Council
WPS1322 Reducing Regulatory Barriers to  Barbara Richard        July 1994          M. Dhokai
Private-Sector Participation in Latin  Thelma Triche                     33970
America's Water and Sanitation Services
WPS1323 Energy Pricng and Air Pollution:  Gunnar S. Eskeland    July 1994          C. Jones
Econometric Evidence from       Emmanuel Jimenez                         37699
Manufacturing in Chile and Indonesia UU Liu
WP51324 Voucher Funds in Transitional    Robert E. Anderson     July 1994          F. Hatab
Economies: The Czech and Slovak                                          35835
Experience
WPS1325 The Economics of Research and    Anwar Shah             July 1994          C. Jones
Development: How Research and                                            37699
Development Capital Affects Production
and Markets and Is Affected by Tax
Incentives
WPS1326 Banks, Capital Markets, and      Gerhard Pohl           July 1994          L Hovsepian
Corporate Govemrance: Lessors    Stijn Claessens                         37297
from Russia for Eastem Europe
WPS1327 Is the Debt Crisis History? Recent   Michael Dooley     July 1994          S. King-Watson
Private Caphal Inflows to Developing  Eduardo Fernandez-Arias            31047
Countries                       Kenneth Kietzer
WPS1328 The Use of New York Cotton Futures  Panos Varangis      July 1994          D. Gustafson
Contracts to Hedge Cotton Price Risk Elton Thigpen                       33714
in Developing Countries        Sudhakar Satyanarayan
WPS1329 The Regulation and Supervision of   David H. Scott      August 1994        K. Waelti
Domestic Financial Conglomerates                                         37655
WPS1330 Revenue Uncertainty and the Choice DelHin S. Go         August 1994        C. Jones
of Tax Instrument during the Transition                                  37699
in Eastem Europe



Policy Research Working Paper Series
Contact
Title                           Author                  Date              for paper
WPS1331The Myth ol Monopoly: A New View    Annette N. Brown      August 1994        M. Berg
of Industrial Structure In Russia  Barry W. Ickes                         36969
Randl Ryterman