497912
research-article2013
RRP46210.1177/0486613413497912Review of Radical Political Economics XX(X)Perraudin et al.
Inter-firm Dependency and
Employment Inequalities:
Theoretical Hypotheses and
Empirical Tests on French
Subcontracting Relationships
Review of Radical Political Economics
2014, Vol. 46(2) 199­–220
© 2013 Union for Radical
Political Economics
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0486613413497912
rrpe.sagepub.com
Corinne Perraudin1, Héloïse Petit1, Nadine Thèvenot1,
Bruno Tinel1, and Julie Valentin1
Abstract
This article highlights the importance of power relations in subcontracting relationships and
analyzes their impact on firms’ employment management practices. We show that the use of
subcontracting creates a chain of inter-firm economic dependency because it leads the principal
contractor to plan and control the activities of the subcontractors. The hypothesis is that this
chain of dependency influences both the skill structure and wage levels. Empirical tests carried
out on French data confirm that firms that subcontract outsource execution tasks and that the
hierarchy of firms impacts employees’ wage levels.
JEL classifications: L24, J82, J31
Keywords
subcontracting, skills, wages, power relation
1. Introduction
Since the beginning of the 1980s, firms have been adopting external growth strategies combined
with vertical disintegration processes, whereby activities previously carried out within vertically
integrated firms are externalized (Feenstra and Hanson 1996). Among the variety of inter-firm
links, this paper focuses on subcontracting relationships. Their spread has led to a proliferation
of situations of economic and financial dependency and changed the nature of the employment
relationship. It is no longer possible to interpret it as a bilateral relationship between employers
and employees, since employers are themselves part of hierarchized networks of subcontracting
relations.
Given the globalization of economic exchanges over the past 30 years, recent analyses of
subcontracting relations and their impact on employment often take place in an international
1CES,
Université Paris 1 Panthéon Sorbonne, Paris, France
Date received: May 26, 2012
Date accepted: January 1, 2013
Corresponding Author:
Bruno Tinel, CES, Université Paris 1, Maison des Sciences Economiques, 106-112 Boulevard de l’Hôpital, 75647 Paris
Cedex 13, France.
Email: [email protected]
200
Review of Radical Political Economics 46(2)
perspective. This literature analyzes the impact of international outsourcing of production on
employment and wage levels (Feenstra and Hanson 1996; Epstein 2006; Houseman 2007).
Feenstra and Hanson (1996) for instance study the disintegration of production chains in the
global economy. They show how skilled and unskilled labor is used differentially along the
“value chain” between industries and analyze the implications on wages. Besides this “value
chain” stream, Akkermans (2011) focuses on subcontracting relations within only one industry
through segmentation theory. Compared to this literature, our paper shows that these links
between inter-firm relations and employment practices are not only attributable to the internationalization of production chains but are inherent to the nature of subcontracting relationships.
We then examine the impact of subcontracting relations on the management of employment
within a single country both within and between industries.
This study is based on a national survey with a broad representative coverage whereas most
of the existing studies, when they include an empirical analysis, draw on case studies. It describes
statistically dominant situations from data covering all sectors of the French economy. Such
approach implies we do not take into account how subcontracting relationships can result in a
variety of situations but we estimate what is predominant at the national level.
The aim of the article is to show that subcontracting relationships are reflected in hierarchical
relationships between firms and that the power relations thus formed impact on firms’ mode of
employment management. Exploration of the literature on governance structures in subcontracting relations leads us, first, to adopt a theoretical perspective that emphasizes the central role that
power relations play in this type of inter-firm relations (Sacchetti and Sugden 2003; Grimshaw
and Rubery 2005). Our subcontracting framework can be interpreted as extending the radical
“divide and rule” hypothesis of the division of labor (Marglin 1974; Stone 1974; Edwards 1979;
Edwards et al. 1982) to inter-firm relations. Second, we formulate two types of hypotheses linking the inter-firm dependency relationships to particular modes of labor mobilization within
firms.
The first hypothesis concerns the homogeneity (or otherwise) of skill structures across firms.
It is based on the notion of division of labor where increased use of subcontracting might be
expected to result in the skills associated with supervisory and conception tasks being concentrated in the principal contractors, and execution tasks being located in the subcontracting firms.
The second hypothesis concerns the link between inter-firm dependency and the level of the
wages paid by firms. The hierarchized representation of the productive system suggests that the
use of subcontracting can be regarded as an indirect form of labor mobilization that enables principal contractors to have a workforce at their disposal when required without having to commit
themselves to a long-term employment relationship. Our aim here is to consider the consequences
in terms of pay for the workers of subcontractors. The hypothesis tested will be that the wages
paid in subcontracting companies are lower than those paid by principal contractors, for each
level of skill.
These theoretical hypotheses are tested empirically using data for France, where the use of
subcontracting is now very common behavior. In 2004, 55 percent of establishments with 20 or
more employees in the non-agricultural market sector stated they had used one or more firms to
carry out part of their activity. We use employer-employee matched data that combine the 20042005 REPONSE survey carried out by DARES (Direction de l’Animation de la Recherche, des
Etudes et des Statistiques, a department of the Ministry of Labor) and the 2005 annual social data
declarations (Déclarations Annuelles de Données Sociales/DADS), which are collected by
INSEE, the French national statistical office. The REPONSE survey provides detailed information on establishments’ employment management practices and commercial strategies. It is one
of the few French data sources that provides information on subcontracting behavior regardless
of sector. It is based on a sample of 2,930 establishments with 20 or more employees that is representative of the non-agricultural market sector as a whole. The DADS are an administrative
Perraudin et al.
201
(and therefore exhaustive) source that provides detailed information, notably on skill structures
and pay levels by establishment. Thus pairing these two data sources provides us with information on subcontracting, pay policy, and on the characteristics of the work force in terms of skills,
as well as on various characteristics of establishments.
The article is structured as follows. We begin by characterizing the hierarchical nature of
inter-firm dependency relations and highlight its impact on employment relationships (section 2).
We then formulate a hypothesis on the consequences of these hierarchical relations for the skill
structure, which we subject to an empirical test (section 3). Finally, we analyze how this hierarchical productive system influences employees’ pay (section 4).
2. Subcontracting Relationships as a Hierarchical Productive
System
Having identified the sources of principal contractors’ dominance over their subcontractors and
illustrated the hierarchical structure of the productive system that results from this in France, we
will analyze the consequences for employment relationships.
2.1 Defining subcontracting relationships as dependency relationships
The question of the dependency relations inherent in subcontracting relationships was already
being discussed in some studies published in the 1950s to 1970s (Houssiaux 1957a; Blois 1972).
These studies describe situations in which “some firms are gaining the advantages of vertical
integration without assuming the risks or rigidity of ownership – situations which might be
described as ‘vertical quasi-integration,” (Blois 1972: 253).1 As early as the 1950s, notably in the
studies by Houssiaux, the use of subcontracting was being analyzed in the context of firms’ strategies of expansion through integration or diversification. Compared with internal growth,
whereby firms invest in new plant and equipment, and “direct” external growth, when firms take
over others, subcontracting was defined by Houssiaux (1957a) as an “indirect method of external
integration” that made it possible to limit investment costs and reduce the risks of excess production capacities, while at the same time strengthening “large firms’ power to dominate” (405).
This power to dominate subcontractors was similar to a monopsony, which gave dominant firms
the power to incorporate subcontractors (Bain 1956), to switch from one to another, or to enter
into competition with them by developing the activity in question in-house. Consequently, they
enjoyed considerable bargaining power when contracts came up for renewal, enabling them to
force down prices and increase their profits. Houssiaux urges that consideration be given to the
consequences of subcontracting for firms’ growth potential and the degree of concentration in the
industries affected even though he emphasizes that subcontracting can be used to outsource labor
as part of an anti-union strategy (see below). Growth through subcontracting enabled firms to
form “groups” and thereby extend their dominance, not simply financially but technologically as
well.
Sacchetti and Sugden (2003), contrary to the analytical framework based on transaction cost
theory in which power rests on the distribution of property rights and is regarded as a case of
market failure, consider power as a constituent element of network relations. Following
Richardson (1972), who highlighted the division of labor in production networks, the power in
inter-firm relationships is derived from the ability of certain firms to plan in advance the
1To
the best of our knowledge, the term quasi-integration was popularized by Blois in 1972, but it was first
the subject of intensive theorization by Houssiaux in 1957a. It had already been used, incidentally, by Bain
(1956: 342) in a paper on industrial concentration and, particularly, by Kohls and Wiley in 1955, who used
this word to describe the organization of the broiler industry in the United States.
202
Review of Radical Political Economics 46(2)
activities of other, legally independent firms. More specifically, in seeking to identify who takes
the strategic decisions, who determines the network’s objectives, and how conflicts are resolved,
Sacchetti and Sugden propose that subcontracting relationships should be represented as “networks of direction,” in which power is concentrated in the hands of the principal contractor.
Sacchetti and Sugden (2003) also consider another category of networks in which power is not
concentrated in the hands of only one actor but distributed, not necessarily equally, between
interdependent actors who participate in strategic decision making. These “networks of mutual
dependence” require proximity between firms which creates the condition for them to be engaged
in cooperation and interactive learning to reach common objectives (such as industrial districts).
This approach is close to Powell’s “mutual dependency networks” (Powell 1990), in which
power is distributed among the firms in a network, each of which is able to exert some influence
over the development of the network’s policies and objectives. Yet, Sacchetti and Sugden (2003)
stress that subcontracting in the automotive industry is not based on this type of network. This is
also the conclusion that can be drawn from Gorgeu and Mathieu’s study (2005) of the French
automotive industry. More generally, Sacchetti and Sugden (2003) underline that in practice
“there may be no real agreement of actors to refrain from exerting their power at the expense of
others.” The use of subcontracting implies, by definition, an asymmetry of position, regardless of
the mutual benefits that both parties may enjoy, since a subcontracting relationship allocates
specific rights of control to the principal contractor.
In the end, the representation proposed by Sacchetti and Sugden (2003) is pyramid-shaped,
with the summit of the hierarchy represented by a core firm in which power over strategic and
conception questions is concentrated. This firm enters into long-term contracts with a small number of large subcontractors that may have some bargaining power, in view of their market positions and production technologies. The role of intellectual property, patents, and innovation can
be crucial here in gaining market power. The further down the pyramid one goes, the more the
number of subcontractors increases; this enables the core firm to make them compete with each
other in order to exploit profit sources. Thus the bottom of the pyramid is made up of production
units specialized in execution that are very dependent on the principal contractor (sweatshops)
and in which working conditions and pay are particularly poor. This hierarchical representation
of production relations helps to explain inter-firm forms of dependency.
Following Williamson (1985), another strain of the literature focuses on the definition of
hybrid modes of governance. Diamantopoulos (1987) for instance differentiates relative and
absolute dependency. More recently, such approach developed through the analysis of modes of
governance in international production chains (Altenburg 2006; Appelbaum 2008; Gereffi,
Humphrey, and Sturgeon 2005). Gereffi, Humphrey, and Sturgeon (2005) propose a theoretical
framework distinguishing five governance modes spanning from pure dependency to relationships based on equality. Our approach first differs in that we do not consider all forms of interfirm relationships in parallel but we directly focus on subcontracting relationships as a particular
form of organizing production. Second, in contrast to contractual theories of the firm, the starting
point of our analysis of subcontracting is the identification of the decision process among actors.
Consequently subcontracting relationships require to be analyzed within the analytical framework of power and hierarchy. Such approach echoes Cowling and Sugden’s (1998) analysis of
the essence of the modern large corporation. Assessing the existing contractual approach on
theory of the firm, the authors suggest focusing on economic planning and strategic decision
making is essential in understanding the functioning of modern corporations.
In a hierarchized structure, first-rank subcontractors may have a certain degree of bargaining
power (derived in particular from their technology or market position), which they can use to
negotiate long-term contracts. However, the influence these subcontractors can exert is limited to
solving particular problems linked to the production and technological specifications of the
goods they are to produce. What gives the networks their hierarchical structure is the
Perraudin et al.
203
concentration of managerial and controlling power over their strategic decisions and objectives,
which ultimately rests in the hands of the core firm. Thus subcontracting is a “segmented form of
productive organization,” to quote Vennin (1975), that involves dominance relations since, in
contrast to intermediate consumption, the goods produced as part of a subcontracting relationship
contribute to the production of a collective product which the principal contractor controls and
for which it therefore has the ultimate economic responsibility.
The dependency between subcontractors and principal contractors is, moreover, very evident
in publications issued by the public authorities. As early as the mid-1970s in France, a report by
the economic and social council defined subcontracting as follows: “a transaction in which one
firm entrusts another with the task of carrying out on its behalf and in accordance with a preestablished set of specifications part of the work required to produce the goods or services for
which it retains ultimate economic responsibility” (Conseil économique et social, report of 21
March 1973). Incidentally, it is this definition that serves as a basis for the Agence Française de
normalisation (AFNOR), the French national standards agency.
The economic or monetary dependency constituent of the subcontracting relationship is created by two complementary aspects:
- Upstream of the production process, the subcontractor’s activities are defined (“planned”)
both qualitatively and quantitatively by the principal contractor. As a result, the subcontractor
becomes economically or financially dependent on the principal contractor, since it does not
begin production without an order from the principal.
- Downstream of the production process, the subcontractor is in possession of a semi-finished
product that cannot be redeployed on the market independently of the principal contractor. To put
it another way, the subcontractor does not have the ultimate economic responsibility for the fragmented product, which is controlled by the principal contractor. Thus this aspect also gives rise to
economic or monetary dependency on the part of the subcontractor, since the fragmented product
is intended solely for the principal contractor, who will combine it with others in order to create a
marketable product. As a result, the subcontractor has no control over the sale of its product.2
These two aspects of the dependency are inherent in subcontracting relationships and distinguish them from other supply relationships in which suppliers conceive the product as a whole or
only part of it and can sell the semi-finished product to another buyer. In contrast, subcontractors
do not design the product but meet the command and the specifications given by one principal
contractor. In the French accounting system, subcontracting spending is specifically registered
and distinguished from other expenditures like supplies, raw materials, and other provisions
which denotes the difference of nature between subcontracting and supply relationships.
Thus inter-firm networks are structured hierarchically on several levels and more or less
loosely depending on the forms of inter-firm dependency, a subcontractor firm being all the more
dependent on its client the greater the share of its turnover that client accounts for. Since organizations differ in status and, consequently, in bargaining power, the gains are not distributed
equally within the network of organizations. As mentioned earlier, the bargaining power of the
subcontractor can increase if, for instance, it possesses a technology which is both uncommon
among its competitors and essential for the principal contractor. This illustrates the notion of
“cascade” subcontracting relationships where high-ranking subcontractors are relatively better
equipped for negotiation than those of the lower ranks. This situation is illustrated by the study
by Perraudin, Thèvenot, Tinel, and Valentin (2006), which highlights the differences in firms’
profitability depending on their position in the hierarchized structure.
2This
is similar to the old putting-out system, in which a principal contractor interposes between the market
and the producer in order to combine the latter’s output with others and sell the “collective product” in the
market (Marglin 1974).
204
Review of Radical Political Economics 46(2)
Table 1. Proportion of establishments, by dependency relations.
Being a subcontractor (SC)
Being a principal
contractor (PC)
Yes
No
Yes
No
PCSC: 15.2%
SC: 4.4%
PC: 39.1%
Not SC or PC: 41.3%
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Percents are sample weighted to be representative of establishments. There are 2,732 observations.
2.2 An empirical presentation of the hierarchy of the French productive system
based on the 2004-2005 REPONSE survey
The hierarchical nature of inter-firm relations can be characterized in terms of the economic
dependency resulting from the fact of being either a subcontractor or principal contractor. The
database on which the present article draws provides information on inter-firm economic dependency through the use of two declarative questions: (i) does the establishment declare itself to be
a subcontractor? (ii) does the establishment declare that it engages in subcontracting?
The answers to these two questions can be used to identify not only inter-firm dependency
resulting from the position of dominance exploited by principal contractors but also “cascade”
subcontracting (Appay 1998), a term used to describe situations in which subcontractor companies that are economically dependent on the principal contractor can offload some of the constraints to which they are subject on to other firms.
It is on this basis that a hierarchized chain of subcontracting relations is described (see Table 1).
The “pure principal contractors” (PC), i.e. establishments that act only as principal contractors
and are not themselves dependent on another principal contractor, constitute the upper level of
this inter-firm hierarchy (39.1 percent of establishments). The intermediate level is made up of
establishments which, while being subcontractors, are at the same time principal contractors
(PCSC) and are thus in a position to transfer some of the constraints resulting from their commercial relations with their principal contractors on to their subcontractors (15.2 percent of establishments). The final level consists of establishments that are solely subcontractors (SC) (4.4
percent of establishments). These three categories of establishments are investigated in parallel
to establishments that are not involved in any subcontracting relationship; they are neither principal contractors nor subcontractors.
Table A1 in the appendix shows the differences of distribution of establishments between sector and size by their dependency relations. PC establishments are over-represented in construction, automotive industry, energy, and consumer goods industry; PCSC in transports and
intermediate goods industry; and SC in consumer goods industry, transports, and agro-food
industry. Nevertheless, the relative position in the chain of contractors is not only linked to the
sector. Table A1 in the appendix shows indeed that both PC and SC exist in all sectors and for all
sizes of establishments.
2.3 The consequences of the hierarchized productive system for management of
the employment relationship
In his pioneering economic analysis, Houssiaux (1957b) noted that subcontracting constituted a
risk for workers, since subcontracting firms frequently do not fully comply with the legislation:
Perraudin et al.
205
non-compliance with production and safety standards, significant use of clandestine labor, etc.
Furthermore, Houssiaux emphasizes the fact that the workers hired by subcontractor firms are
generally less unionized and are subject to greater anti-union pressures than employees in the
large firms. The legally instituted absence of employee representatives and of works councils in
French firms with fewer than 50 employees (which is still true today) is an incentive to maintain
small firms, in which the employer’s authority can go unchallenged. Thus workers in subcontracting firms are less well protected and the employer’s authority is less disputed. Even in this
period, Houssiaux mentions the concerns of the major trade union federations about the risks of
social dumping associated with the development of subcontracting. Putting forward a hypothesis
that differs from the technico-economic arguments deployed hitherto in his analysis, Houssiaux
writes these significant lines, which undoubtedly reflect more what was to happen on a larger
scale two or three decades later: “This difference [between the social guarantees in subcontractor
firms and large integrated firms] explains why some firms can pursue quasi-integration: a large
firm wants to replace such an integrated workshop that is particularly difficult to manage with a
subcontractor; the ringleaders are dismissed and the other workers are divided up among other
workshops” (Houssiaux 1957b: 406). Eventually, such an outsourcing strategy can be interpreted
as a transposition at the level of the inter-firm relations of the “divide and rule” strategy analyzed
for instance by Stone (1974) in the radical tradition of segmented labor markets.3
Some recent studies also highlight the impact of inter-firm relationships on the management
of employment. Grimshaw and Rubery (2005: 1,038) observe that analysis of the employment
relationship has to take account of inter-firm relationships: “To understand the tensions and pressures on the internal employment relationship there is a clear need to locate this relationship
within the web of business-to-business production relationships as well as in the system of intercapitalist competition.”
Grimshaw and Rubery (2005: 1,030) note that the outsourcing of certain activities “may
reflect a desire to escape internal constraints that arise out of the very high trust relations on
which the internal compromise is based.” In cases in which normal cooperative practices prove
insufficient to call into question the previous arrangements, employers may resort to threatening
to use external labor in order to reconfigure the internal compromises. According to these authors,
the processes of externalization and fragmentation serve as instruments to be used by employers
seeking to change the power relations within their firms. The use of external workers in itself
enables employers to mobilize labor at reduced cost.
Grimshaw and Rubery stress that an employer can exert control over a workforce located
outside the immediate boundaries of its own organization, which requires inter-firm cooperative
networks to be structured and hierarchized to that end. A firm can seek to control and motivate
the workers who, in legal terms, are employed by its suppliers. In other words, some elements of
the organization of the production process that used to be analyzed solely in terms of the internal
employment relationship extend in reality beyond the firm’s legal boundaries.
Consequently, commercial contracts signed by principal contractors should not be considered
as only implying an exchange of goods or services but also as involving command over the subcontractor’s workforce. This interlocking of organizations, with one firm exerting more or less
direct control over the workforce of another, is not limited to a single level, since a firm that is
dependent to a greater or lesser extent on its client will also be a client of one or more other firms.
Ultimately, the chain of subcontracting relationships is constructed around a hierarchy of
inter-firm relationships, in which the dominant firms are able to influence the management of
3Tinel
(2013) shows that the “divide and rule” argument is not based on knowledge monopoly, contrary to
what is suggested by Marglin (1974), but is in fact a single element of the broader Babbage-Marx analysis
of the division of labor.
206
Review of Radical Political Economics 46(2)
employment beyond their own boundaries. The principal contractor’s dominance has many consequences for the management of employment in the subcontractor firms. We will examine two
of them, which are directly linked to the inter-firm dependency relationship, namely the power to
structure the inter-firm division of labor and hence to determine the categories of workers hired
by all the firms involved, and the power to influence pay levels in the dependent firms.
3. The Consequences of the Hierarchized Productive System for
the Skill Structure
Subcontracting relationships influence the distribution of tasks between firms and consequently
may alter the skill structure in individual firms. In order to clarify this link between the inter-firm
division of labor and the skill structure, it would seem advantageous to examine the issues at
stake in the current trend towards increased use of subcontracting with reference to the processes
of integration and disintegration that have punctuated the history of advanced capitalist economies over the past century. An empirical test is proposed to highlight the impact of the use of
subcontracting on the inter-firm division of labor and hence on the skill structure.
3.1 Integration and skill structure
Following the classic studies by Chandler, the prototype of the modern industrial firm was
defined by the large-scale industrial enterprises that contributed to the development of managerial capitalism.4 Multifunctional firms began to emerge from the end of the 19th century onwards;
as they grew considerably in size, they began to adopt a multidivisional structure. In order to
achieve much lower unit costs than those of the smaller firms that had existed hitherto, they made
extensive use of the most advanced production technologies, with which they were able to
achieve very high production volumes. These large firms proceeded to invest in large-scale production plants that enabled them to exploit to hitherto unequalled levels the economies of scale
and scope offered by the technology. Considerable investment was also channeled into national
and international distribution and marketing networks in order to bring sales volumes into line
with output. These giant firms were able to function and grow because such investment was
accompanied by the rapid expansion of managerial personnel, who were specially recruited and
trained to administer the enlarged production facilities and the increase in personnel working in
production and distribution, as well as to monitor and coordinate these two basic functional
activities and to plan and allocate resources for future activities.
As far as work organization and the division of labor were concerned, managerial firms put
into practice systematically and on a large scale Taylor’s principles of scientific management.
Conception was separated from execution and execution was itself broken down into basic tasks
depending on the requirements imposed by the technology in use. These principles of task fragmentation and specialization were applied not only to production on the shop floor but, as managerial firms grew, to distribution functions, R&D, and even managerial functions themselves.
This gave rise in turn to an extremely hierarchized organization. The most vital conception functions, such as strategic planning, finance and accounts, some R&D, activities and marketing,
were answerable directly to the top of the hierarchy. On the other hand, the most routine execution functions were allocated to relatively autonomous operational units that reproduced the general
organizational structure on a miniature scale.5
As far as the organizational structure by skill in large-scale integrated firms is concerned, the
most highly skilled jobs were concentrated in the departments specializing in conception
4Cf.
Chandler (1990), or even Simon (1947) and Penrose (1959).
had their own accounting, production, purchasing, sales, personnel management departments, etc.
5They
Perraudin et al.
207
functions, which were closest to the managerial department. At the other extreme, the least skilled
jobs, which were increasingly broken down into an ever smaller number of execution tasks, were
concentrated in the operational units; these units also contained a certain number of skilled jobs,
which were required for managerial and supervisory purposes. Although they were relatively
independent in accounting and managerial terms, these operational units remained dependent on
decisions taken at the top of the company; they did not act on their own account but rather put
into practice plans drawn up by senior management.
This business history written by Chandler is acknowledged by a broad range of the production
theories proponents à la Richardson: from the Marxian tradition (Braverman 1974 or Duménil
and Lévy 2004) and the radical school (Stone 1974) to the competence or capability-based theory
of the firm (for example Dosi, Nelson, and Winter 2000). The first insists on the shaping of the
division of labor by social relations of production whereas the second focuses on the technological aspects of the division of labor. This business history is also acknowledged by economists of
history like Lazonick (1991) who tries to find the right balance between the two previous
traditions.
3.2 Disintegration and the skill structure: Theoretical hypotheses
This long period of integration tendency ended several decades ago and the reverse movement
towards disintegration and reduction in company size started (see Thèvenot and Valentin 2005,
for France; Dey, Houseman, and Polivka 2006, for the United States). This swing back of the
pendulum certainly cannot be reduced to the rise of subcontracting observed in most of the
advanced capitalist countries.6 Nevertheless, the rise in subcontracting relations has to be considered as one of the factors that have been driving the disintegration of the integrated managerial
firm. But why did capitalists start to disintegrate production thirty years ago after having pursued
integration during more than a century? The general movement of disintegration can be interpreted as a reaction to the profit squeeze crisis of the 1970s which has resulted from the success
of the Keynesian social-democrat regulation of capitalism to achieve full employment (see
Duménil and Lévy 2004). At this time, the power and social rights gained by the workers inside
the big integrated corporation were undermining profits. This system had to be transformed and
rules changed in order to restore profitability, notably by a renewed grip of capital on the labor
process. The rise of subcontracting relations can be interpreted as an element of this class strategy
to improve and even to rebuild the subjection of labor to capital (Tinel et al. 2007). Milberg
(2008) proposes a complementary explanation where subcontracting is driven by the opportunity
to free up financial resources to increase dividend payments and share buybacks in order to satisfy shareholders. His analysis is focused on global value chains applied to relations between the
United States and China but it could apply as well to the French productive structure. Whatever
the main driving force towards industrial disintegration, we will discuss here the consequences
of the rising share of subcontracted output for the skill structures of firms.
In a division of labor perspective, planning and engineering tasks are separated from operational tasks and (highly) skilled employees tend to be in charge of conception tasks while the
unskilled workers are restricted to execution tasks. In this framework, the workers who specialize
in execution tasks, however fragmented they may be, have to be supervised. This supervisory and
control work is carried out by skilled workers and is supplementary to the execution work. Thus
we can make the hypothesis that certain organizations specialize in conception tasks and employ
6In
reality, subcontracting already existed in the managerial era, but it seemed to be so closely linked to the
rapid development of the large, multi-divisional firm that, in the 1950s, some authors made subcontracting one of the methods by which large integrated firms expanded by referring to it as “quasi-integration”
(Houssiaux 1957b). Cf. above.
208
Review of Radical Political Economics 46(2)
a higher share of skilled workers, whereas the externalized production units will, as a result of the
subcontracting, specialize in execution tasks and will therefore employ a higher share of unskilled
workers (although there will still be some skilled workers because of the need for supervision
imposed by the fragmented specialization).7 In formal terms, the emergence of subcontracting
from an initial situation in which firms are integrated leads to the establishment of several legally
separate production units, which specialize in different segments of the production process. In
reality, these firms are neither independent of each other nor operating on equal terms: some
specialize in implementing a set of specifications drawn up by the principal contractor and are
dependent on it. Consequently, in this approach to disintegration through subcontracting, in
which firms are structured into a hierarchy, the outsourcing has an impact on the skill structure
that is not neutral: the subcontracting firms employ higher shares of unskilled workers whereas
principal contractors retain more conception and management activities.
3.3 Inter-firm dependency and skill structure: Empirical test
At this point, the empirical test consists in examining the homogeneity (or otherwise) of the skill
structure in firms depending on their position in the inter-firm hierarchy. Pairing the REPONSE
survey with data from the 2005 annual social data declarations (DADS, INSEE) provides information on firms’ position in the subcontracting relationship and on the share of employees in
each firm in each of three skill levels: manual workers and unskilled white-collar employees,
skilled manual and white-collar workers, and highly skilled employees (managers and supervisors, technicians and professionals). The classification of manual workers as “unskilled manual
workers” and “skilled manual workers” exists in the French nomenclature. The same distinction
between unskilled and skilled workers does not exist directly in the nomenclature for whitecollars but it is possible to make it visible through the method of Burnod and Chenu (2001) which
is widely used in French publications. For white-collars, the notion of skill is based on the occupation of the salaried employee as defined by the French standard classification of occupations.
This method enables us to distinguish occupations according to complexity and control. Finally,
in 2005, 21 percent of the employees of the establishments of 20 people and more were unskilled,
37 percent were low skilled, and 42 percent were high-skilled workers (see Table 2).
The distribution of workers by skill according to the position of the establishments in the
subcontracting chain shows clearly a differentiation of skill utilization between establishments.
Nearly half of the workforce of the PC is highly skilled whereas less than a third of the SC workforce is highly skilled. A quarter of the SC establishments, workforce is unskilled and 45 percent
is low skilled while for PC it gives respectively 17 percent and 40 percent. This link between the
position of the establishment in the subcontracting chain and the workforce skills will be clarified
through an estimation below by controlling for some important variables such as sectors and
establishment sizes.
The empirical test assesses whether the share of employees by skill level in an establishment
varies depending on the firm’s position in the chain of dependency relations defined above. Since
some establishments do not have any employees in one of the 3 categories of skill levels, the
share of employees (for each of the three categories of skill levels) is estimated with the following Tobit model:
7This
mechanism has certainly an impact at the international level on the less advanced industrialized countries to which some of the subcontracting is directed, which may in turn give rise to increased demand for
unskilled labor.
209
Perraudin et al.
Table 2. Workforce distribution by skills, by dependency relations.
in %
% of … in the workforce
PC
Unskilled workers
Low skilled workers
Highly skilled workers
Total
16.6
39.6
43.7
100.0
PCSC
15.5
46.1
38.5
100.0
SC
25.5
45.5
29.0
100.0
Not SC or PC
Total
30.2
28.5
41.3
100.0
21.2
37.2
41.7
100.0
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Percents are sample weighted to be representative of establishments. There are 2,732 observations.
Y * = α0 + α1 PC j + α 2 PCSC j + α 3 SC j + βX j + ε j
Y jk =  jk
0

if Y jk* > 0
otherwise
where Yjk is the share of employees for skill level k = {unskilled, low skilled and highly skilled}
in establishment j, Xj is a vector of control variables including the sector of activity, the size and
age of establishment, company turnover and the workforce composition by gender and age of
employees (see Table A1 in appendix for the distribution of control variables). PCj, PCSCj, and
SCj are dummy variables that take the value 1 if the establishment j is respectively PC, PCSC or
SC and 0 otherwise. The impact of being a principal contractor, a subcontractor or both is estimated by comparison with the situation of a firm with no specific commercial relation. The aim
of this estimation is to exhibit a specific qualification structure according to the position in the
dependency chain by taking care of the composition effects which could bias this link. The direction of the estimation does not intend displaying a causality.8 For instance, the SC establishments
may have a share of unskilled workers which is significantly bigger because they are restricted to
execution tasks. But it can be argued as well that those establishments became SC because they
had not the critical skills which could have put them on another position in the chain. Both arguments are coherent with the present framework of analysis.
The results of the estimations are shown in Table A2 in the appendix. They illustrate that sector of activity and firm size play a crucial role in the structure of skills among establishments. The
workforce composition by gender and age is also determinant. Beyond these elements, the skill
structure is significantly different all along the inter-firm hierarchy, compared to firms without
commercial relation. Equality tests of coefficients have been calculated two by two, based on the
estimation results, in order to give an account of the differences associated with the specific position of establishments in the subcontracting relationship. Table 3 reports differences in the effects
(coefficients) and the significance of the differences (based on Student’s t tests).
As expected, the share of unskilled workers is all the greater the lower a firm is positioned in
the inter-firm hierarchy (i.e. the more it operates as a subcontractor) and the share of highly
skilled workers is all the greater the more a firm dominates the others in its network. In other
words, the share of unskilled workers is significantly higher for subcontractors than for both
principal contractors (8.6 points higher in SC than in PC) and subcontractor and principal contractors (7 points higher in SC than in PCSC). The situation of establishments in intermediate
8We
use the information of 2005 for the share of employees by skill whereas the position of the establishments in the dependency relation is available for 2004. This time delay does not serve as a proof of the
direction of the causal relation.
210
Review of Radical Political Economics 46(2)
Table 3. Differences and significance in effects of the inter-firm dependency on skill structure.
% of unskilled workers in % of low skilled workers
% of highly skilled
the workforce
in the workforce
workers in the workforce
PC − PCSC
PC − SC
PCSC − SC
–1.57
–8.59***
–7.02**
–2.20*
–3.98*
–1.78
3.33**
12.04***
8.71***
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Note: Table reports differences in coefficients (reported in Table A2 in appendix) and significance level based on
Student’s t tests for equality of the two coefficients (*, **, *** for 10%, 5%, 1%). For example, PC establishments
have a % of highly skilled workers in the workforce 3.33 higher than the PCSC establishments and this difference is
significant at the 5% level.
situations is not significantly different from principal contractors. Yet, the share of skilled manual
and white collar workers is particularly high in these establishments. PCSC establishments have
significantly more low skilled workers than principal contractors (2.2 points higher) but do not
differentiate themselves from “pure” subcontractors. The share of highly skilled employees
stands out in “pure” subcontractor firms: it is significantly lower for subcontractors than for others, and more precisely, it is significantly lower (of 8.7 points) for SC than for PCSC and lower
(of 3.3 points) for PCSC than for PC. In this regard, firms in intermediate positions or at the top
of the hierarchy are not significantly different from those that have no involvement in subcontracting. These results bring support to the present approach based on the division of labor.
Thus the inter-firm dependency that arises out of the subcontracting relationship is related to
firms’ skill structures. Our results are coherent with the hypothesis that the use of subcontracting
leads to a transfer of the least skilled workers to the most dependent firms in the productive system hierarchy. The inter-firm division of labor appears to be linked to inter-firm relationships. We
turn now to an examination of the wage practices to which this externalized workforce is
subject.
4. The Consequences of Productive System Hierarchization for
Wage Practices
The purpose of this section is to examine the links between the hierarchization of the productive
system and the establishments’ wage practices. The theoretical framework is first exposed and
then an empirical test is realized in the second subsection.
4.1 Subcontracting and wages: Theoretical hypotheses
Savings in labor costs in terms of wages, health insurance, and pension benefits are one of the
most frequently mentioned factors determining the use of subcontracting (Abraham and Taylor
1996; Houseman 2001; Gramm and Schnell 2001). This hypothesis is usually couched in terms
of efficiency wage, in the version of the theory based on norms of fairness (Akerlof and Yellen
1990). These norms prevent firms from paying different rates to permanent and temporary workers, and subcontracting is interpreted as a means of circumventing them. In all these studies, the
econometric analysis, which draws on American data, involves investigating whether pay levels
in the firms in question explain the use of flexible staffing arrangements. The expected link is
positive: the higher the wages firms offer their employees (those they employ directly), the
Perraudin et al.
211
greater incentive they will have to use subcontracting. The results obtained partially corroborate
this hypothesis. The study by Abraham and Taylor (1996), which draws on data derived from a
supplementary questionnaire in the Industry Wage Survey (IWS) carried out in the manufacturing sector in 1986 and 1987, shows that contracting for janitorial services (i.e. the least skilled
service activities) is primarily a response to a need to reduce hourly labor costs. Houseman
(2001), who draws on the 1995, 1997, and 1999 Current Population Survey Supplement on
Contingent and Alternative Work Arrangements, finds that benefits variables have a positive and
significant influence on the use of flexible staffing arrangements, except for contract workers.
Gramm and Schnell (2001), using data from their own 1994-96 survey of human resource managers in Alabama establishments, find that the probability of using subcontracting is positively
linked to the wages of core employees relative to those of other similar workers in the industry.
Those different studies find an influence of the wage paid over the subcontracting resort
behavior. The direction of causality can be discussed. Do firms resort to subcontracting because
the wage they pay to their workers is high or is it the relative position in the subcontracting chain
which leads to different payment schemes? In other words, what are the practices of firms according to their relative position in the chain and what are the elements which influence the wage
usual procedures of subcontracting firms? In their studies, Houssiaux (1957b) and Grimshaw and
Rubery (2005) emphasize that the use of subcontracting can be regarded as a means of mobilizing a cheaper workforce. The theoretical research mentioned in the first section of this article
suggest that hierarchies between firms reflect in the wage level because they contribute to “divide
and rule” and hence to increased competition between workers. This way of analyzing the link
between wages and subcontracting is used in a recent article by Berlinski (2008). This article
examines the pay differential between workers in a firm and workers employed by a subcontracting firm who carry out the same task (two activities are investigated, namely cleaning and security). The pay differential that remains unexplained by the characteristics of the workers and
firms turns out to be positive (more than 16 percent), which the author interprets as an indication
that the use of subcontracting may serve to reduce the wages and benefit costs associated with the
labor mobilized.9 The present paper is also based on this type of argument initiated by radical
political economy about segmented labor markets. But rather than examining the wage differentials between employees depending on the nature of the employment relationship linking them to
the beneficiary of their labor (Berlinski 2008), the econometric study presented below examines
the wage differentials for all employees with the same level of skill in an establishment depending on its position in the economic dependency chain described above.10 In other words, the
analysis focuses on the effect on workers’ wages to be in a specific establishment rather than in
another one which would be characterized by another position of the subcontracting chain. This
enables us to see whether economic dependency implied by the subcontracting chain is reflected
on the level of wages paid by firms.
4.2 Subcontracting and wages: An empirical test
The approach adopted involves assessing whether the hierarchy of subcontracting relations is
reflected in wage levels. In other words, we evaluate whether the wage levels in establishments
differ according to establishment’s positions in the dependency relationship, controlling for different composition effects. Considering the difference in workforce composition illustrated in the
previous section, the link between subcontracting relations and wage levels has to be examined
separately for each level of skills. This analysis is possible thanks to our database which provides
9The
author notes that this differential disappears, or is even reversed, for wages paid to workers in the highskill categories (engineers, computer scientists, technicians, etc.).
10It is not possible to have more details about skills decomposition within the database.
212
Review of Radical Political Economics 46(2)
Table 4. Distribution (quartiles and mean) of net hourly median wages (in euros) in establishments for
each level of skill, by dependency relations.
Unskilled workers median wage
PC
PCSC
SC
Lower Quartile
Median
Upper Quartile
Mean
7.2
6.9
6.7
7.8
7.7
7.1
8.7
8.7
8.0
8.2
8.1
7.6
Low skilled workers median wage
Not SC or PC
PC
PCSC
SC
Not SC or PC
Lower Quartile
Median
Upper Quartile
Mean
6.9
8.2
8.0
7.5
7.7
7.5
9.0
9.1
8.3
8.5
8.2
10.2
10.1
9.6
9.5
7.8
9.4
9.3
8.7
8.8
Highly skilled workers median wage
PC
PCSC
SC
Not SC or PC
Lower Quartile
Median
Upper Quartile
Mean
11.9
11.5
11.3
11.0
13.9
13.5
12.9
12.7
16.3
16.7
16.0
15.9
14.8
21.4
14.1
14.1
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Percents are sample weighted to be representative of establishments. There are 2,732 observations.
us, for each establishment, the average and quartiles wages paid for each skill level. The wages
in question are the net hourly wages for the three skill levels investigated in this study: unskilled
manual and white-collar workers, skilled manual and white-collar workers, and highly skilled
workers.
The summary of the distribution of the median hourly wages shows clearly a differentiation
between establishments (see Table 4). But this requires us to control the possible effects of the
characteristics of the establishments through an estimation.
We estimate the following linear model (with ordinary least square method):
W jk = α0 + α1 PC j + α 2 PCSC j + α 3 SC j + βX j + ε j
where Wjk is the median hourly wage (in logarithm) in the establishment j for the skill levels k =
{unskilled, low skilled and highly skilled}.11 Xj is a vector of control variables including the
establishment characteristics (such as sector of activity, size and age of establishment), size of
company as measured by turnover, the workforce characteristics (such as the share of women and
of employees under 30). We introduce also the state of the market for the last 3 years (growing,
declining, or stable) and the difficulty in predicting demand as declared by the employer
11The
results are qualitatively the same if the variable to be explained is the average hourly rate rather than
the median. The results are not reported here, but are available from the authors.
213
Perraudin et al.
Table 5. Differences and significance in effects of the inter-firm dependency on median net hourly
wage.
Unskilled workers
median wage
PC − PCSC
PC − SC
PCSC − SC
6.72***
9.48***
2.58
Low skilled workers
median wage
4.26***
7.49***
3.10
Highly skilled workers
median wage
5.56***
6.55*
0.94
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Note: Table reports (exponential) differences (see note 12) in coefficients (reported in Table A3 in appendix) and
significance level based on Student’s t tests for equality of the two coefficients (*, **, *** for 10%, 5%, 1%). For
example, PC establishments have median wage for unskilled workers 6.72% higher than PCSC establishments and this
difference is significant at the 1% level.
representative (in order to control the fluctuations of activity). PCj, PCSCj, and SCj are dummy
variables that take the value 1 if the establishment j is respectively PC, PCSC, or SC and 0 otherwise. Table A3 in the appendix shows the results of the whole estimation. As in the previous
section, Student’s t tests of coefficient differences for SC, PCSC, and PC examine the different
effects on wages of the position of the establishment. Table 5 reports values of the differences in
coefficients and significance of the differences based on the complete estimation.12
The results show that the median hourly wage varies considerably depending on the establishment’s position in the hierarchy of subcontracting relationships. Employees’ pay reflects the
power of principal contractors on subcontractors: it is significantly lower in establishments which
are both subcontractors and principal contractors than those which are pure principal contractors
and, also, in pure subcontractors than in pure principal contractors. Moreover, the findings are
significant for all three skill levels once establishment and workforce characteristics have been
controlled for. We can further note that the relation is particularly significant for unskilled workers whereas it is weaker for the highly skilled. More precisely, the median of the net hourly wages
for unskilled workers is 9.4 percent higher in the pure principal contractors than in the pure subcontractor firms; it is also 7.5 percent higher for low skilled workers and 6.4 percent higher for
the very highly skilled. One can note finally that the differences between SC and PCSC are not
significant.
5. Conclusion
Subcontracting relationships are dominated by power relations. Since it leads principal contractors to plan subcontractors’ activities and control the sale of their fragmented products, the use of
subcontracting creates a hierarchical inter-firm division of labor and a chain of economic dependency. Firms may shift on to others the economic constraints to which they are subject by becoming principal contractors in their turn. The hierarchical inter-firm division of labor is part of the
dominant firms’ growth processes, which are linked to particular forms of the intra-firm division
of labor.
The empirical tests carried out have shown, first, that firms that subcontract externalize execution tasks and, second, that wages fall as a firm’s dependency increases. It might be thought that
the field in which the present study was carried out has led to our results being underestimated.
the dependent variable is the logarithm of wages, we calculate, for example for a1, a2, the difference
100 × (exp(a1 – a2) –1) as the measure of the effect in percent of wage growth.
12Since
214
Review of Radical Political Economics 46(2)
First, it excludes small establishments with fewer than 20 employees, even though subcontracting affects them to an even greater degree. Second, it excludes subcontractor establishments
located abroad, whereas studies of international subcontracting clearly reveal labor exploitation
as a driving force (Epstein 2006). Finally, it takes no account of informal work, whereas sociological studies of subcontracting (Appay 1998; Jounin 2008) describe it as a key factor in the
construction of inter-firm dependency relations.
These results provide empirical evidence to support the hypothesis that firms’ position in subcontracting networks and their distance from the “pure” principal contractors reflects the extent
of their economic dependency. Our estimations do not constitute a test of the dependency hypothesis but they are coherent with the interpretation of the productive system in terms of the power
relations we propose. This framework is coherent with the twin neo-Marxist tradition of labor
process analysis and radical approach of segmented labor markets which both develop a theoretical and a historical content. They both insist on the prominent role of social relations of production – especially the conflict between labor and capital – in the analysis of the organization of
production and its transformations. They both consider that the division of labor instead of the
“firm” as such is the relevant core concept to understand not only the organization of production
but also the way labor markets are structured and the way firms are related and opposed to each
other. In this framework, it is hence natural to grasp jointly the issues of labor relations and of
industrial strategy instead of doing as if they were independent questions. As shown by this
paper, subcontracting relations typify particularly well this double sided aspect of production.
This research on subcontracting relations can be furthered in different directions. First, the
hierarchy between subcontracting firms has an effect on profitability. A reduced profitability for
subcontractors should translate onto wages and/or labor productivity. Besides, it would be interesting to have a more accurate view of the intensity of competition along the inter-firm hierarchy
in relation to several profit indicators. Another possible direction for further research is to try to
specify the differences in labor organization according to the position of the establishment in the
inter-firm hierarchy through items such as Taylorism, lean production/demand oriented production, learning organization, size of internal hierarchy, work (relative) autonomy, and work
intensification.
215
Perraudin et al.
Appendix
Table A1. Distribution of establishments for characteristics of establishments, by level of inter-firm
dependency.
PCSC
(in %)
SC
(in %)
Not SC
or PC
(in %)
% of
establishments
in the sample
26.4
55.0
56.8
46.4
64.7
38.1
31.5
34.2
34.7
24.3
31.8
5.8
12.
25.0
32.3
22.1
37.3
3.3
4.9
16.9
4.7
3.6
6.5
12.2
6.1
5.0
3.9
6.9
2.0
2.1
4.8
0.3
5.2
61.3
20.6
12.0
16.4
9.3
17.8
63.2
58.8
43.6
70.8
59.3
3.2
4.1
6.1
12.3
8.7
7.5
19.5
4.5
16.7
7.6
10.0
35.3
41.5
47.7
55.6
66.0
15.5
15.1
13.8
16.5
12.6
4.8
4.1
3.7
2.4
2.5
44.5
39.4
34.9
25.5
18.9
63.9
20.1
9.4
5.2
1.4
37.6
37.2
47.1
12.5
16.7
12.0
5.8
4.9
1.7
44.1
41.3
39.2
12.8
67.7
19.5
30.1
35.1
43.0
49.4
7.9
25.1
13.3
9.1
7.3
6.0
2.7
1.1
54.7
33.8
41.1
40.3
19.4
32.1
29.7
18.8
46.6
41.5
28.4
28.1
13.9
6.1
3.8
4.5
4.6
21.5
40.2
61.0
24.6
48.2
27.2
36.4
42.7
34.0
17.5
16.7
9.8
5.1
5.0
2.2
41.1
35.5
54.1
23.7
52.8
23.5
39.7
39.3
37.9
39.4
39.2
14.7
19.6
14.7
17.5
15.3
5.1
4.2
3.2
4.9
4.4
40.6
36.9
44.2
38.3
41.2
56.4
14.5
29.2
70.5
100.0
PC
(in %)
Sector
Agro-food industry
Consumer goods industry
Capital goods and automotive industries
Intermediate goods industry / energy
Construction
Transports
Commerce
Financial and real-estate activities
Business services
Personal and domestic services
Education, health, social services
Establishment size
Less than 50 employees
From 50 to 99
From 100 to 199
From 200 to 499
500 and more
Age of establishment
Less than 10 years
10 to 50 years
More than 50 years
Company turnover in € million
Less than 5
5 to 10
10 to 100
More than 100
Proportion of women
Less than 15%
From 15% to 65%
More than 65%
Proportion of employees aged under 30
Less than 12%
From 12 to 40%
More than 40%
State of the market
Growing
Declining
Stable
Difficulty in predicting demand
Total
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Percents are sample weighted to be representative of establishments. There are 2,732 observations, except for “State
of the market” and “Difficulty in predicting demand” (2,678 observations) due to nonresponse. Among establishments
of Agro-food industry, 26.4% are PC, 5.8% are PCSC, 6.5% are SC, and 61.3% are not SC or PC. Establishments of
Agro-food industry represent 3.2% of all establishments.
216
Review of Radical Political Economics 46(2)
Table A2. Skill structure and inter-firm dependency (results of Tobit estimation).
% of unskilled
workers in the
workforce
Commercial relations
PC
–0.028
PCSC
–0.012
SC
0.058
No commercial relation
ref
Sector
Agro-food industry
–0.073
Consumer goods industry
–0.183
Capital goods and automotive
–0.191
industries
Intermediate goods industry / energy –0.152
Construction
–0.199
Transports
–0.262
Commerce
ref
Financial and real-estate activities
–0.391
Business services
–0.207
Personal and domestic services
–0.002
Education, health, social services
–0.032
Establishment size
Less than 50 employees
ref
From 50 to 99
0.046
From 100 to 199
0.061
From 200 to 499
0.109
500 and more
0.066
Less than 10 years
–0.023
10 to 50 years
ref
More than 50 years
–0.019
Company turnover in € million
Less than 5
ref
5 to 10
–0.018
10 to 100
–0.021
More than 100
–0.047
Proportion of women
Less than 15%
–0.022
From 15% to 65%
ref
More than 65%
0.179
Proportion of employees aged under 30
Less than 12%
–0.031
From 12 to 40%
ref
More than 40%
0.156
Constant
0.263
Scale
0.270
Number of observations
2732
Number of censured observations
546
Log Likelihood
–686.65
% of low skilled
workers in the
workforce
% of highly skilled
workers in the
workforce
(0.013)**
(0.018)
(0.032)*
0.012
0.034
0.051
ref
(0.010)
(0.014)***
(0.025)**
0.015
–0.018
–0.105
ref
(0.012)
(0.016)
(0.028)***
(0.029)**
(0.027)***
(0.024)***
0.192
0.132
0.045
(0.024)***
(0.021)***
(0.019)***
–0.114
0.036
0.121
(0.027)***
(0.025)
(0.022)***
(0.021)***
(0.029)***
(0.028)***
0.115
0.132
0.335
ref
–0.071
–0.086
–0.135
–0.144
(0.017)***
(0.022)***
(0.022)***
0.021
0.043
–0.117
ref
0.350
0.222
0.110
0.171
(0.019)
(0.026)*
(0.025)***
(0.024)***
(0.018)***
(0.026)***
(0.026)***
(0.016)
(0.016)
(0.018)
(0.017)***
(0.016)
(0.012)
(0.029)***
(0.020)***
(0.029)
(0.028)
(0.021)***
(0.016)***
(0.023)***
(0.0220)***
(0.013)
ref
–0.009
–0.031
–0.045
–0.078
–0.009
ref
0.040
(0.010)***
ref
–0.015
–0.005
–0.021
0.055
0.022
ref
–0.009
(0.023)
(0.022)
(0.023)**
ref
0.037
0.013
–0.027
(0.018)**
(0.017)
(0.018)
ref
–0.006
0.013
0.055
(0.021)
(0.020)
(0.021)***
–0.107
ref
–0.099
(0.013)***
(0.014)***
0.042
ref
(0.013)***
–0.094
(0.021)***
0.349
(0.003)
0.248
2732
26
–103.16
(0.012)***
(0.015)***
(0.024)***
(0.003)
(0.018)**
(0.018)***
(0.020)***
(0.019)***
(0.019)
(0.015)
(0.015)***
(0.013)**
0.127
ref
–0.069
–0.031
ref
(0.017)*** –0.059
(0.027)***
0.351
(0.004)
0.215
2732
79
221.08
(0.014)
(0.014)**
(0.015)***
(0.015)***
(0.014)
(0.012)***
(0.012)***
(0.010)***
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Note: Table reports parameter estimates, standard errors (in parenthesis), and significant level (*, *, **, *** for 1%,
5%, 10%).
217
Perraudin et al.
Table A3. Median net hourly wage and inter-firm dependency (results of OLS estimation).
Unskilled
workers
median wage
Commercial relations
PC
0.01933
PCSC
–0.04572
SC
–0.07124
No commercial relation
ref
Sector
Agro-food industry
0.04810
Consumer goods industry
0.04089
Capital goods and automotive
0.04873
industries
Intermediate goods industry / energy 0.05354
Construction
0.01228
Transports
0.03922
Commerce
ref
Financial and real-estate activities
0.06494
Business services
0.01113
Personal and domestic services
0.01810
Education, health, social services
0.01295
Establishment size
Less than 50 employees
ref
From 50 to 100
–0.00383
From 100 to 200
–0.00336
From 200 to 500
0.04263
500 and more
0.10383
Age of establishment
Less than 10 years
0.00195
10 to 50 years
ref
More than 50 years
0.02192
Company turnover in € million
Less than 5
ref
5 to 10
0.01850
10 to 100
0.02367
More than 100
0.07488
State of the market
Growing
–0.00024
Stable
ref
Declining
0.00278
Difficulty in predicting demand –0.03745
Proportion of women
Less than 15%
0.03108
From 15% to 65%
ref
More than 65%
–0.02794
Proportion of employees aged under 30
Less than 12%
0.05297
From 12 to 40%
ref
More than 40%
–0.05428
Constant
2.03915
Number of observations
2143
Adj R2
0.2413
Low skilled
workers
median wage
Highly skilled
workers
median wage
(0.010)*
(0.014)***
(0.024)***
0.02567
–0.01609
–0.04660
ref
(0.008)***
(0.011)
(0.019)**
0.01406
–0.03827
–0.04812
ref
(0.012)
(0.016)**
(0.029)*
(0.021)**
(0.020)**
(0.018)***
0.06429
0.10827
0.09560
(0.018)***
(0.017)***
(0.015)***
–0.02156
0.04079
0.00946
(0.027)
(0.025)
(0.022)
(0.016)***
(0.021)
(0.021)*
(0.013)***
(0.018)***
(0.017)***
(0.024)***
(0.016)
(0.021)
(0.021)
0.12068
0.12696
0.05782
ref
0.09118
0.07208
0.06047
0.01699
(0.017)***
(0.012)***
(0.019)***
(0.018)
0.01255
0.13216
0.03791
ref
0.07554
0.10116
–0.09963
–0.06703
(0.020)
(0.026)***
(0.025)
(0.025)***
(0.018)***
(0.027)***
(0.026)**
(0.014)
(0.014)
(0.015)***
(0.015)***
ref
–0.01075
–0.00968
0.02457
0.09042
(0.011)
(0.011)
(0.012)**
(0.012)***
ref
–0.05003
–0.02086
–0.01913
0.09230
(0.016)***
(0.016)
(0.018)
(0.018)***
(0.017)
(0.012)
(0.014)
(0.009)**
–0.00291
ref
0.01887
(0.008)**
–0.02238
ref
0.00153
(0.017)
(0.017)
(0.017)***
ref
–0.00753
0.01151
0.07736
(0.014)
(0.014)
(0.014)***
ref
0.05008
0.06408
0.09623
(0.021)**
(0.020)***
(0.021)***
–0.04426
ref
–0.01647
–0.00841
(0.011)***
(0.015)
(0.011)
–0.02377
ref
–0.03127
(0.014)*
(0.014)**
0.06303
ref
–0.02565
2.60275
2653
0.1539
(0.012)***
(0.015)*
(0.027)***
(0.009)
(0.012)
(0.009)***
(0.011)***
(0.011)**
(0.010)***
(0.012)
(0.022)
(0.011)
–0.00441
ref
–0.00019
–0.03738
(0.007)
0.01252
ref
–0.01969
(0.009)
0.05147
ref
–0.05225
2.13618
2599
0.3117
(0.008)***
(0.010)
(0.007)***
(0.009)**
(0.010)
(0.018)
Source: 2004-2005 REPONSE survey (DARES), DADS 2005 (INSEE).
Field: Establishments of 20 and more employees in the private sector (excluding agricultural sector)
Note: Table reports parameter estimates, standard errors (in parenthesis), and significant level (*, *, **, *** for 1%,
5%, 10%).
218
Review of Radical Political Economics 46(2)
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or
publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research has been funded by the DARES (French Ministry of Employment and
Labor Research Unit) for the analysis of the 2004-2005 REPONSE survey.
References
Abraham, K. G., and S. K. Taylor. 1996. Firms’ use of outside contractors: Theory and evidence. Journal
of Labor Economics 14(3): 394-424.
Akerlof, G. A., and J. L. Yellen. 1990. The fair wage-effort hypothesis and unemployment. The Quarterly
Journal of Economics 105(2): 255-283.
Akkermans, D. H. M. 2011. Subcontracting and the creation of inequality in the Dutch clothing industry
1980-1992. Review of Radical Political Economics 43(2): 230-252.
Altenburg, T. 2006. Governance patterns in value chains and their development impact. The European
Journal of Development Research 18(4): 498-521.
Appay, B. 1998. Economic concentration and the externalization of labour. Economic and Industrial
Democracy 19: 161-184.
Bain, J. S. 1956. Advantages of the large firm: Production, distribution, and sales promotion. Journal of
Marketing 20(4): 336-346.
Berlinski, S. 2008. Wages and contracting out: Does the law of one price hold? British Journal of Industrial
Relations 46(1): 59-75.
Blois, K. J. 1972. Vertical quasi-integration. Journal of Industrial Economics 20(3): 253-272.
Braverman, H. 1974. Labor and monopoly capital. New York: Monthly Review Press (new ed. 1998).
Burnod, G., and A. Chenu. 2001. Employés qualifies et non qualifies: Une proposition d’aménagement de
la nomenclature des catégories socioprofessionnelles. Travail et Emploi 86: 87-105.
Chandler, A. D. 1990. Scale and scope, the dynamics of industrial capitalism. Cambridge, MA: Harvard
University Press.
Cowlin, K., and R. Sugden. 1998. The essence of modern corporations: Markets, strategic decision-making
and the theory of the firm. The Manchester School 66(1): 59-86.
Dey, M., S. Houseman, and A. Polivka. 2006. Manufactures’ outsourcing to employment services. Staff
Working Paper No 07-132, Kalamazoo, W.E. Upjohn Institute for Employment Research.
Diamantopoulos, A. 1987. Vertical quasi-integration revisited: The role of power. Managerial and Decision
Economics 8: 185-194.
Dosi, G., R. Nelson, and S. Winter, eds. 2000. The nature and the dynamics of organizational capabilities.
Oxford University Press.
Duménil, G., and D. Lévy. 2004. Capital resurgent. Roots of the neoliberal revolution. Cambridge, MA:
Harvard University Press.
Edwards, R. 1979. Contested terrain: The transformation of the workplace in the twentieth century. New
York: Basic Books.
Edwards, R., D. M. Gordon, and M. Reich. 1982. Segmented work, divided workers. Cambridge University
Press.
Epstein, G. 2006. Employment oriented macroeconomic policy and the challenge of outsourcing. PERI
Working Paper, Feb. Amherst, MA: University of Massachusetts.
Feenstra, R. C., and G. H. Hanson. 1996. Globalization, outsourcing, and wage inequality. American
Economic Review 86(2): 240-245.
Gereffi, G., J. Humphrey, and T. Sturgeon. 2005. The governance of global value chains. Review of
International Political Economy 12(1): 78-104.
Gorgeu, A., and R. Mathieu. 2005. L’action contrainte des directions d’usines. Analyse de la gestion de
l’emploi dans la filière automobile. Economies et Sociétés, Série Socio-Economie du Travail 26(8):
1,503-1,530.
Perraudin et al.
219
Gramm, C. L., and J. F. Schnell. 2001. The use of flexible staffing arrangements in core production jobs.
Industrial and Labor Relations Review 54(2): 245-258.
Grimshaw, D., and J. Rubery. 2005. Inter-capital relations and the network organisation: Redefining the
work and employment nexus. Cambridge Journal of Economics 29(6): 1,027-1,051.
Houseman, S. 2001. Why employers use flexible staffing arrangements: Evidence from an establishment
survey. Industrial and Labor Relations Review 55(1): 149-170.
Houssiaux, J. 1957a. Le concept de “quasi-intégration” et le rôle des sous-traitants dans l’industrie. Revue
Economique 8(2): 221-247.
Houssiaux, J. 1957b. Quasi-intégration, croissance des firmes et structures industrielles. Revue Economique
8(3): 385-411.
Jounin, N. 2008. Chantier interdit au public. Enquête parmi les travailleurs du bâtiment. Paris: Editions
La Découverte.
Kohls, R. L., and J. W. Wiley. 1955. Aspects of multiple-owner integration in the broiler industry. Journal
of Farm Economics 37(1): 81-89.
Lazonick, W. 1991. Business organization and the myth of the market economy. Cambridge University
Press.
Marglin, S. A. 1974. What do bosses do? The origins and functions of hierarchy in capitalist production.
Review of Radical Political Economics 6(2): 60-112.
Milberg, W. 2008. Shifting sources and uses of profits: Sustaining US financialization with global value
chains. Economy and Society 37(3): 420-451.
Stone, K. 1974. The origins of job structures in the steel industry. Review of Radical Political Economics
6(2) (summer): 113-73.
Penrose, E. T. 1959. The theory of the growth of the firm. Oxford: Basil Blackwell.
Perraudin, C., N. Thèvenot, B. Tinel, and J. Valentin. 2006. Sous-traitance dans l’industrie et ineffectivité
du droit du travail: Une analyse économique. Economie et Institutions 9: 35-55.
Powell, W. W. 1990. Neither market nor hierarchy: Network forms of organization. Research in
Organizational Behavior 12: 295-336.
Richardson, G. B. 1972. The organisation of industry. The Economic Journal 82(327): 883-896.
Sacchetti, S., and R. Sugden. 2003. The governance of networks and economic power: The nature and
impact of subcontracting relationships. Journal of Economic Surveys 17(5): 669-691.
Simon, H. 1947. Administrative behavior. A study of decision-making processes in administrative organizations. New York: The Free Press.
Thèvenot, N., and J. Valentin. 2005. La sous-traitance comme alternative au contrat de travail: Une evaluation empirique pour la France, 1984-2000. Economie Appliquée LVIII(3): 51-79.
Tinel, B., C. Perraudin, N. Thèvenot, and J. Valentin. 2007. La sous-traitance comme moyen de subordination réelle de la force de travail. Actuel Marx 41: 153-164.
Tinel, B. 2013. Why and how do capitalists divide labour? From Marglin and back again through Babbage
and Marx. Review of Political Economy 25(2), 254-272.
Vennin, B. 1975. Pratiques et signification de la sous-traitance dans l’industrie automobile en France. Revue
Economique 26(2): 280-306.
Williamson, O. 1985. The economic institutions of capitalism. Firms, markets, relational contracting. New
York: Free Press.
Author Biographies
Corinne Perraudin is associate professor at the University Paris 1 Panthéon Sorbonne and member of the
Centre d’économie de la Sorbonne. Her recent research is focused on labor relations, subcontracting, and
work organization.
Héloïse Petit is associate professor at the University Paris 1 Panthéon Sorbonne, member of the Centre
d’économie de la Sorbonne, and associate researcher at the Centre d’Etudes de l’Emploi. Her recent research
is focused on the link between firm characteristics (being quoted, a subcontractor or part of a business
group) and employment relations.
220
Review of Radical Political Economics 46(2)
Nadine Thèvenot is associate professor at the University Paris 1 Panthéon Sorbonne and member of the
Centre d’économie de la Sorbonne. Her recent research is focused on labor relations and subcontracting.
Bruno Tinel is associate professor at the University Paris 1 Panthéon Sorbonne and member of the Centre
d’économie de la Sorbonne. His recent research is focused on labor relations (labor process theory, subcontracting relations) and on the dynamics of public debt.
Julie Valentin is associate professor at the University Paris 1 Panthéon Sorbonne and member of the Centre
d’économie de la Sorbonne. Her recent research is focused on the link between employment legislation and
employers’ practices (subcontracting, use of temporary contracts, and layoffs).