1More than seventy years have passed since the 1937 seminal paper by Coase on the nature of the firm, in which he raised the question of the boundaries of the firm and thus of the market. This question remains topical even though the institutional and industrial environment has considerably changed. In the spirit of Coase, neo-institutionalist scholars have argued that vertical integration is the best governance structure when assets become more specific and that more appropriable quasi-rents are created because of the possibility of post-contractual opportunistic behavior and high transaction costs (Williamson 1975). However, even in an “industrial know-how-based view of the firm”, economists such as Monteverde and Teece (1982) have shown that “transaction cost considerations surrounding the development and deepening of human skills appear to have important ramifications for vertical integration in the automobile industry, thereby supporting the transactions cost paradigm advanced by Williamson”(ibid., p. 212). Previously, Mancke (1972) had shown that the iron ore industry “was transformed from an unconcentrated industry composed almost entirely of independent iron ore producers who sold their product over the open market into a concentrated industry” (ibid., p. 222). Thus, the de jure vertical integration  strategy was ‛paradigmatic’ in the organizational literature until the beginning of the 1990s, when the international industrial landscape changed significantly.
2The industrial dynamics of capitalism are significantly linked to transformations in organizational forms of production. In short, we have moved from the integrated traditional firm, as described by Chandler (1959, 1977), to the “disintegrated” modern firm based on highly idiosyncratic resources. These new trends in the industrial world, which occurred in the 1980s and 1990s, have led firms to engage in inter-firm cooperation and other forms of vertical quasi-integration . These changes have redefined the nature of the relationships between the modern firm and its suppliers by replacing simple arm’s-length dealings with new long-term network arrangements (e.g. Sturgeon 2002). The best evidence of this tendency is the emergence of the so-called global network-firm that unifies a set of legally independent firms positioned along a single supply chain to design and manufacture complex and specific products (see Chassagnon 2011a, 2011b, 2013; Baudry and Chassagnon 2012b). In this arrangement, one firm’s output is the input of another firm. The firms are linked to each other and form a network. A firm that we call “the hub-firm” organizes this network. These changes have led researchers to criticize the make-or-buy paradigm, which characterized the vertically integrated firm (e.g. Holmtröm and Roberts 1998). However, does the emergence of complex inter-firm organizations really constitute a new and significant theoretical improvement from an organizational economic perspective? What are the implications of vertical disintegration strategy and industrial dynamics for the boundaries of the capitalist firm? What are the nature and the main regularities of these new forms of governance structure? How have influential academic works analyzed and approached these complex organizational forms in recent decades?
3The aim of the paper is to address these questions by focusing on vertical quasi-integration forms of production. We propose to approach the research from the point of view of economic analysis methodology, by giving attention to three authors who developed their own approach: Williamson, Houssiaux and Blois. The theory of Williamson has three limitations related to the roles he gives to hybrids’ stability, to asset specificity and to inter-firm power relationships. It seems very interesting to shed light on the manner by which Houssiaux and Blois (who wrote on this subject before Williamson did) analyzed these three points similar to empirical evidence from the present-day situation – notably the global network-firm – to complete the Williamsonian philosophical and methodological approach. We thus think that this paper is one of the first to develop such an organizational economics view on this crucial question in organizational and industrial research.
4The academic literature generally considers Williamson (1991) to be one of the first organizational scholars to analyze the implications of industrial and institutional changes for the hybrids category. In contrast, we argue that this theoretical argument was in fact introduced in the 1950s by an influential French economist who is seldom quoted in academic work, namely Houssiaux (1957a, 1957b). Indeed, Houssiaux was the first to introduce the notion of “quasi-integration” to designate a middle zone composed of supplier-customer relationships between firms in successive production stages. According to Houssiaux, quasi-integration is a type of economic integration that emerges not within a firm, but within a stable and specific entity composed of a large firm and its supplier-subcontractors. A few years later, Blois (1972), who does not quote Houssiaux, also used the term “quasi-integration” to describe situations where some firms gain the advantages of vertical integration without assuming its risks or rigidity of ownership in an effort to understand how the repercussions of financial collapse could extend beyond the legal entity of the corporation. Thus, Blois uses power and economic dependency based on asset specificity to explain the quasi-integration strategy. Without using the rhetoric of transaction cost economics, these two authors thus preceded Williamson  in introducing a new governance structure that could match the industrial dynamics and the blurring of the boundaries of the firm. Furthermore, the industrial logic that underlies the works of Houssiaux and Blois seems to be closer to the view of Richardson (1972, 2003), who defined the third organizational form as a dense network of cooperation and affiliation by which firms are related, than to the “economizing view” of Williamson. In this sense, their work remains useful to appreciate and study the nature of the modern multinational firms, which are often based on a dense vertical network of inter-organizational relationships.
5The paper is organized in three main sections. Section 1 analyzes Williamson’s influential model of vertical integration and of hybrids and presents its main limitations. Section 2 presents the seminal thoughts of Houssiaux, who developed a theory of quasi-integration as a new stable organizational form. Section 3 rests on the work of Blois, who introduced the idea of inter-firm economic power to the study of vertical quasi-integration. Section 4 concludes the paper.
Williamson, asset specificity and the realm of hybrids as the dominant organizational economic thought
6The transaction cost theory of Williamson was, above all, a theory of vertical integration. As Williamson (1999) reminds us, his first transaction cost article dates to 1971 (The Vertical Integration of Production: Market Failure Considerations, published in the American Economic Review) and deals with vertical integration and make-or-buy decisions (see notably Williamson 1975, 1981). Williamson’s aim is to propose an analytical framework of the governance structures of capitalism based on transaction cost economizing. For him, market, hierarchy and hybrids are the main governance structures to be found in the institutions of capitalism. A governance structure is an institutional arrangement that minimizes transaction costs for a given environment. Williamson argues that these structures must implement the contractual modes that best suit controlling parties’ behaviors and commitments during the economic transaction as a function of three parameters: (1) uncertainty; (2) frequency and (3) asset specificity. Williamson explains that asset specificity gives sense to the assumption of behavioral uncertainty of agents linked to strategic conjectures they make regarding the attitudes of their partners. Frequency potentially increases this behavioral uncertainty, but it is asset specificity that gives substance to transaction frequency. Thus, in fine, Williamson (1985, 1991) proposes to relate the three “discrete” governance structures to the degree of asset specificity for a given level of uncertainty.
7We understand that asset specificity is at the center of arbitration of governance structures for Williamson. This idea is not really new because we find a trace of it in the works of Marshall (1920) and Polanyi (1962). Asset specificity “has reference to the degree to which an asset can be redeployed to alternative uses and by alternative users without sacrifice of productive value” (Williamson 1991, p. 281). The rationale of the existence of firms is hence contained in asset specificity (see Riordan and Williamson 1995). The main problem with such specific assets is hold-up situations (see also on this point Klein, Crawford and Alchian 1978). Williamson explains that “asset specificity creates bilateral dependency and poses added contracting hazards” (Williamson 1991, p. 282). When some economic agents conclude a contract for specific investments, a situation of strong economic dependence emerges. Because economic agents are supposed to be rationally limited and opportunistic, some of them can selfishly use lock-in to generate a quasi-rent. The risk of defection that comes from opportunism leads to a loss of efficiency because contractors, who are potentially inclined to hold up, can choose to not invest in specific assets. Thus, they avoid being locked into an economic relationship and bearing sunk costs that would result from the inability to redeploy specific assets.
8The presence of specific assets in transactions requires a different form of contract law to support the governance structure. From this perspective, Williamson uses the typology proposed by Macneil (1974). The argument can be explained by considering that transaction costs increase when the degree of asset specificity increases. When assets in a transaction are unspecific, transaction costs are low, and the “classical contract law” applies independently of frequency. This is the ideal market transaction. When assets are highly specific and transactions are recurrent, co-contractors have to protect against opportunistic behaviors and hold-up risks by implementing a “protective governance structure”. Hence, de jure vertical integration should be used, and the form of contract law that supports the integrated firm is that of forbearance doctrine. What interests us here is the third and last scenario that concerns inter-firm structures. When the degree of the specificity of assets in the transaction increases and the frequency of transactions is not very regular, it is the trilateral governance structure and the “neoclassical contract law” that should be implemented because, Williamson argues, it better facilitates continuity and encourages appropriated adaptation.
9The firms involved in such hybrids can maintain their autonomy by using flexible contracting mechanisms. As Williamson (1991) notes, “long-term, incomplete contracts require special adaptive mechanisms to effect realignment and restore efficiency when beset by unanticipated disturbances” (ibid., p. 272). However, in his 1975 book, Williamson did not consider hybrids because he considered “intermediate forms of contracting” (ibid., p. 109) to be very unstable and to appear as epiphenomena. In his heuristic demonstration of 1991, Williamson places markets and hierarchies as polar modes and locates “hybrid forms – various forms of long-term contracting, reciprocal trading, regulation, franchising, and the like – in relation to these polar modes”, considering hence that “the hybrid mode displays intermediate values in all features” (Williamson 1991, p. 280). Indeed, further in the text, he writes: “the hybrid mode is characterized by semi-strong incentives, an intermediate degree of administrative apparatus, displays semi-strong adaptations” (ibid., p. 283) compared to markets and hierarchies. This is the efficiency zone of hybrids. In this view, hybrids differ from the two polar opposites by degree of attributes, even though he argues that “the hybrid form of organization is not a loose amalgam of market and hierarchy but possesses its own disciplined rationale” (ibid., p. 294).
10However, it is also very detrimental to locate very different institutional arrangements at the same level. It is impossible to characterize, define and differentiate hybrids accurately based on Williamson’s three criteria. For example, in his 1991 paper, Williamson argues that joint ventures can be described as hybrids (like franchising or networks), but adds that these are “temporary forms of organization” (Williamson 1991, p. 293). Similarly, he evokes the famous “quasi-firms” (similar to vertical quasi-integration reasoning) of Eccles (1981) as hybrids, but says nothing more about how they can be an efficient solution except that “the details matter” (Williamson 1991, p. 293). In short, it seems very confusing to locate temporary and stable governance structures at the same level, except if, as some arguments seem to indicate, Williamson really considers hybrids to be an unstable form of organization. On this point, the theory of Williamson is unclear. Although he was very clear regarding the treatment of vertical integration, his analysis of hybrids sometimes appears confusing. Is it pertinent to locate different contractual arrangements at the same level along the same continuum – “plural forms”, to quote Bradach and Eccles (1989) – like strategic alliances, partnership, networks or franchising?
11Hodgson (2002) notes that “by adopting the obfuscatory concept of a firm-market hybrid, Williamson muddled his former claim that the firm is essentially different from the markets” (ibid., p. 50). Indeed, following his reasoning, high asset specificity (especially when it is coupled with uncertainty and frequency) necessarily leads to vertical integration. More precisely, he establishes the following basic regularity in a very recent paper: “as bilateral dependency builds up, the efficient governance of contractual relations progressively moves from simple market exchange to hybrid contracting (with credibility supports)” (Williamson 2008, p. 5). However, some organizational scholars have shown that asset specificity in hybrids is high. Holmström and Roberts (1998) argue that only a small part of inter-firm organizational changes can easily be understood in terms of traditional transaction cost theory, in which hold-up problems are resolved by integration. They write:
Many of the hybrid organizations that are emerging are characterized by high degrees of uncertainty, frequency and asset specificity, yet they do not lead to integration. In fact, high degrees of frequency and mutual dependency seem to support, rather than hinder, ongoing cooperation across firm boundaries.
13What is very surprising from a theoretical perspective is that the dynamics of Williamson indicate that hybrids are not stable and cohesive governance structures in the sense that the more bilateral dependency increases, the stronger the necessity to recourse to de jure vertical integration (hierarchy) becomes.
14In fact, Williamson explains vertical integration with the combination of a behavioral hypothesis (bounded rationality) and the characteristics of transactions. The internalization of transactions in a firm alters the sequence of transactions because the hierarchy substitutes an employment relationship, which provides a zone of acceptance for a commercial relationship. This relationship removes the risk of post-contractual supplier opportunism, and the employer can exercise authority over its employees. Additionally, the firm is unable to maintain high-powered market incentives. In other words, it is necessary to compare the benefits of vertical integration, which result from the removal of supplier hold-up risks and improved flexibility, with its costs, which result from weak incentives and bureaucratic costs. Thus, he writes:
given the burdens of bureaucracy to which internal organization is subject, the use of hybrid contracting – as a means by which to support continuity while preserving incentive intensity – can be thought of as a means by which to hold the burdens of bureaucracy (hierarchy) in abeyance.
16The introduction of the hybrid form (Williamson 1991) between the firm and the market does not affect general reasoning on holdup risks because only the unified ownership of assets – vertical integration – can resolve lock-in. In terms of the boundaries of the firm, the analysis of the quasi-integration forms’ functioning rules sheds light on a strong limitation of Williamson’s theory. There is an overvaluation of hold-up risks, which do not sufficiently explain the boundaries of the firm and the impact of ownership on economic agents’ incentives. As Dyer (1997) has shown from a comparison of automobile manufacturers, high asset specificity does not necessarily lead to vertical integration and formal hierarchy. Although asset specificity increases, the recourse to hierarchy is not an inevitable solution. The empirical evidence shows, for example, that the large global vertical network-firms that are described as hybrids have grown worldwide throughout the last three decades (see Lynn 2005; Berger 2006).
17Williamson’s theory is especially pertinent in terms of clarity and operationalization from a make-or-buy perspective. However, it is less convincing regarding the analysis of modern multinational firms that have changed greatly since the mid-1970s. In a 1981 paper, Williamson focuses on the large multidivisional integrated firm as the modern organizational form without giving attention to new modes of vertical subcontracting. He notes that multinational firms grew but only deals with the issues of direct foreign investments that this growth implies. Although he introduced hybrids in his 1985 book, Williamson has never proposed a solid analysis of vertical quasi-integration forms that is not based on asset property rights concentration (the “make” in the make-or-buy dichotomy). This is surprising because Williamson (2008) is very critical towards the supply chain management academic field and does not consider it to be successful.
18The theory of Williamson is based on a positive point of view, but from it he proposes normative recommendations and does not hesitate to say that empirical investigations largely corroborate his theory. He writes that transaction cost economics is “an empirical success story” (1999, p. 1092). Despite the difficulty of empirically measuring transaction cost approach’s theoretical assumptions, numerous attempts to do so have been successful in the sense mentioned by Williamson (see Shelanski and Klein 1995; Masten and Saussier 2002). However, other investigations have criticized this “empirical success story” (David and Han 2004; Hodgson and Carter 2006). We think it is useless to comment on these investigations because empirical arguments both for and against them are valid, making it impossible to reach a general conclusion. However, what matters is the adequacy of the Williamsonian theory for real-world industrial practices. What is evident is that special forms of vertical quasi-integration have never really been the object of precise investigations. This is particularly the case for the global network-firm.
19Organizational scholars have observed that the economic boundaries of the modern firm enclose a productive collective entity in which the organization remains vertical, but not linked by capital relationships. In brief, vertical inter-firm cooperation processes have profoundly affected the relationships between legally independent economic entities, and the so-called network-firm is the best evidence of this tendency. In the 1980 and 1990s, powerful multinational firms, from the automotive and aeronautics industries to textiles and the microcomputer industry, chose to form network-firms. In this context, the network-firm refers to a productive entity (e.g., Nike, Toyota, Airbus, IBM) that unifies a set of legally independent firms that are vertically integrated and coordinated by a main firm called the hub-firm, “which is the firm that, in fact, sets up the network and takes a proactive attitude in the care of it” (Jarillo 1988, p. 32), to produce a specific and complex good or service. Cooperating firms, often of different nationalities, produce a specific good through vertically fragmented production. Firms are vertically integrated without recourse to equity ownership (hierarchy).
20Empirical evidence shows that the network-firm, due to its original and specific nature, is composed of numerous strong specific assets (Dyer 1996, 1997; Powell 1998; Dyer and Nobeoka 2000) that can be human, physical, immaterial, temporal, dedicated or sited (Williamson 1991). The creation, accumulation and development of complementary resources specific to the network make the network-firm a singular and inimitable entity and dynamically structure its competitive advantages. This is an important limitation to the general reasoning of Williamson.
21Additionally, the question of coordination is increasingly important because of the higher degree of organizational complexity in the new economy. The functioning rules of the network-firm show that the distribution of power between actors is a crucial coordinating mechanism in a vertical network of production (Chassagnon 2011a, 2013). Due to the dependence between the members of a network, power is more widely dispersed and accrues to key actors in the network. The complementarity effects are significant in the network-firm. Each legal entity of the network is in a situation of economic dependence with the others, which requires compliance between the actors. The hub-firm allows each individual entity to acquire a portion of the power. The exploitation of de facto power is at the origin of the emergence of cooperation in the network-firm and strengthens its integrity. There is not an exclusive power, but rather several powers within the network-firm, whereas management power in the traditional capitalist firm is unilateral. Several organizational scholars have shown that power is at the center of vertical quasi-integrations like the network-firm (see later and, e.g., Rajan and Zingales 2000; Sacchetti and Sugden 2003; Caniëls and Roeleveld 2009; Baudry and Chassagnon 2012b; Chassagnon 2013).
22However, Williamson is very critical of the concept of power and considers that “the main problem with power is that the concept is so poorly defined that power can be and is invoked to explain virtually anything” (Williamson 1985, p. 237-238). To reinforce his position, Williamson does not hesitate to quote March (1966), who thinks that “power is a disappointing concept” (March 1966, p. 70). Recently, Williamson has written that power is a “myopic trap” in supply chain analysis (Williamson 2008, p. 10).
23After the confusing argument on hybrids as stable and specific governance structures, this is for us the second strong limitation of Williamson’s transaction cost economics, and it could result from the fact that Williamson does not give enough attention to relational factors and relational modes of contractualization in hybrids analysis (see Powell 1990; Uzzi 1997). The network-firm succeeds in maintaining strong incentives without making opportunistic behaviors emerge by encouraging different forms of inter-firm trust (see Kilduff and Krackhardt 2008). As Jones, Hesterly and Borgatti (1997, p. 911) explain, “many industries increasingly are using network governance-coordination characterized by informal social systems rather than by bureaucratic structures within firms and formal contractual between them”. For them, network governance “is composed of autonomous firms that operate like a single entity in these tasks requiring joint activity” (ibid., p. 916).
24In conclusion, Williamson’s transaction cost economics is a very solid theory of vertical integration and of make-or-buy decisions, but Williamson never really introduces vertical quasi-integration forms in his analysis because (1) he does not consider them to be specific, durable and stable governance structures; (2) he overestimates the role of asset specificity and hold-up risks in the legitimating of hierarchies compared to hybrid structures like network-firms; and (3) he does not consider power in the analysis of the governance structures of capitalism. In this view, the ?realistic character’ of the Williamsonian theory should be reconsidered regarding real-world industrial practices. However, it is interesting to observe that two organizational scholars – Houssiaux and Blois – addressed earlier and more substantially the emergence and development of vertical quasi-integration forms from a very different point of view. Their work will be the object of the next two sections.
Houssiaux as the founding father of the notion of quasi-integration
25Houssiaux is undoubtedly one of the leading representatives of French thought on industrial and business economics. He was interested in European industrial policies and, above all, by the nature of emergent inter-firm cooperative organizations. Notably, he proposed in 1957 – i.e., more than a decade before Williamson – a significant analysis of vertical integration and notably an accurate study of complex forms of vertical quasi-integration. However, it is important to underline the fact that Williamson never quoted and used the works of Houssiaux – he may have not known about them, as the works of Houssiaux have never been translated into English. This is very regrettable because we find in the works of Houssiaux some of the canonical arguments used later by Williamson – notably on hybrids.
26Houssiaux starts from the fact that economic theory did not propose a precise definition of the firm (which was not clearly explained by Williamson), but reduced it to a function of production, transformation and distribution of final products, to question the nature of organizational forms that do not correspond to the traditional market or to the vertical integration of productive activities. Thus, he noted that since the 1920s, and notably during the development of automotive industry, industrial firms have tried to find subcontractors to produce specific parts for new products without using vertical integration. Thus, Houssiaux (Houssiaux 1957a) argued that new organizational forms different from both the market and the firm had emerged:
this intermediate zone, that includes the usual customer relationships between firms acting at different and successive stages of production, constitutes the domain of quasi-integration. […]
Quasi-integration is a process of economic integration not based within the firm but in a group constituted by the big firm and its subcontractors.
28Houssiaux is hence the first industrial economist to use the notion of quasi-integration and to propose a definition of it. Nevertheless, it is also important to note that contrary to Williamson Houssiaux did not analyze different forms of hybrids based on vertical contractual relationships. This methodological argument should be kept in mind in the rest of the paper because it can bring some complementary answers to the Williamsonian theoretical model.
29Houssiaux, like Williamson, considered that quasi-integration could not be reduced to standard market subcontracting, even though this organizational form safeguards the legal and financial autonomy of subcontractors. Indeed, after observing the industrial dynamics of Europe in the 1950s, he concluded that these subcontracting relationships could be appreciated as integration relationships because “they do not rest on prices, quantities and other markets attributes of subcontracted products but on the technical aspects of relationships, methods and ranges of fabrication, delivery rates and times” (ibid.). For Houssiaux, quasi-integration emerges for two main reasons: (1) the difficulties encountered by the large integrated firm in responding to changes in demand (lack of flexibility) and thereby growing; and (2) the unproductive and poor organization of the markets for equipment. In this view, Houssiaux (1959, p. 838) defines quasi-integration as “the set of soft methods of concentration and integration of firms”. In this same article, he argued that he could affirm after several years of observation that quasi-integration, which goes hand in hand with the transformation of industrial structures, presents two features. First, it is characterized by the existence of a “center of coordination” (ibid.) that could be (but is not necessarily) the decentralized large firm. On the other hand, quasi-integration implies the “principle of the independence of the management of the firms grouped in the quasi-integrated whole” (ibid., p. 839).
30The analysis of the “vertical networks of subcontractors” (ibid., p. 223) also shows that quasi-integration covers the old domains of the diversified and integrated firm. In this sense, we find in the work of Houssiaux the idea that firms should concentrate on the activities for which they are best suited and outsource the activities that are far from their core competences (e.g., Bettis, Bradley and Hamel 1992). However, even though Houssiaux never used the notion of “hybrid of market and firm”, he affirmed that the firms involved in quasi-integration will be oriented towards the market or towards the firm as a function of their degree of specialization; this argument echoes back to that of “contractual continuum” developed by Williamson (1991) and explained previously. Additionally, and in strong contrast with Williamson, who bases his analysis on allocation considerations (e.g., Hodgson 1998), Houssiaux (1957a, 1957b) explained that it is the field of the “theory of the firm” (Houssiaux 1957a, p. 225) that should analyze the nature of quasi-integration by shedding light on “the conditions of production” (ibid., p. 226). In this view, inter-firm cooperation is highly important in the process of quasi-integration, notably because there are no legal and financial links between the firms in the network of production.
31According to Houssiaux, we can admit that a relationship between firms becomes a quasi-integration when two conditions are met: (1) the transfer of responsibility from the quasi-integrated firm and quasi-integrated subcontractors; and (2) durable relationships involving cooperation and interdependence between the customers and the subcontractors. The first condition means that the larger firm delegates a part of its responsibility to the subcontractor, which has to produce a part of the final complex product. In this view, the nature of subcontracting is different from that of the activities of a workshop chief in a decentralized firm and from that of the simple supplier (Adelman, 1949). Regarding the second condition, we note that Houssiaux (1957a) considers that all of the subcontracting firms in quasi-integration benefit from special and privileged relationships with the hub-firm, characterized by strong collaborations compared to normal market transactions. He adds that this strong inter-firm cooperation very often goes with the fact that a large part of the sales of the subcontractor is dedicated to the economic relationship. The argument of responsibility has been forgotten in the work of Williamson.
32In quasi-integration, “special suppliers” (ibid., p. 238) play a central role. Houssiaux takes the example of the automotive industry to show that the hub-firm requires that special suppliers produce a well-defined part of the final product, or even sometimes a complete product. For this reason, a special supplier in quasi-integration is also a “complementary supplier” (ibid., p. 241). However, the nature of this specific relationship between the hub-firm and special subcontractors is important to note. Houssiaux reminded us that the subcontracting contract is the main legal element of quasi-integration (this idea echoes back to the “Williamsonian work” of Masten 1988). This contract includes technical clauses containing precise contractual specifications defined by the hub-firm. These specifications are often very complex and detailed in inter-firm contracts between the hub-firm and its special subcontractors.
33Nevertheless, Houssiaux (1957a) also explained that quasi-integration does not necessarily imply formal contracts because substantial “de facto subcontracting” (ibid., p. 242) – notably among very small suppliers – is implemented without contracts but rests on collaborative and durable relationships, meaning that there is quasi-integration. This argument, which has sometimes been raised in the managerial literature on inter-firm networks (e.g., Powell 1990; Uzzi 1997), has not been fully developed and analyzed by Williamson, who focuses on the “contractual agent” (see Williamson 1985). However, it appears that the process of establishing vertical relationships in the global network-firm is often based on informal agreements and de facto power relationships (Baudry and Chassagnon 2012b). In this spirit, the work of Houssiaux seems to be close to the real-world practices embodied in the emergence and development of global network-firms. Another element characterizing these special subcontracting contracts is their length. These are long-term contracts between the hub-firm and its special subcontractors that will be broken only if the hub-firm decides to take recourse to de jure vertical integration. Otherwise, the loyalty of the hub-firm towards its specialized subcontractors is obvious, so the stability of the complementary quasi-integration is preserved from opportunistic behaviors. This argument strongly differs from the conclusion reached by Williamson about the instability of hybrid forms being due notably to actors’ opportunism – even though Houssiaux does not fully reject (in some special cases) the possibility of opportunistic unilateral behaviors.
34On the contrary, Houssiaux (1957a) shed light on the informal cooperative relationships between the hub-firm and its special subcontractors. For example, he noted that quasi-integration relationships include different forms of productive guidance and aids. These could be technical, based on raw materials provided by the hub-firm, or financial (notably through easy terms and overdraft facility). Additionally, Houssiaux explained that control of fabrication is only seldom imposed on special subcontractors in practice. Indeed, as Houssiaux wrote: “the internal control exercises a dangerous pressure on the subcontractor because it leads to the presence of a controller chosen by the big firm (the hub-firm) at the subcontractor, which is a source of frequent conflicts” (ibid., p. 246). We understand here that the importance of organizational atmosphere in economic exchange relationships is crucial for the coordination of inter-firm networks. However, if Williamson (1975) takes into account organizational atmosphere in the analysis of intra-firm cooperation, he never uses it for studying the conditions of emergence of inter-firm cooperation, whereas Houssiaux insists on the “sociological criteria of the quasi-integration” (Houssiaux 1957a, p. 244).
35At the end of the 1950s, Houssiaux understood that the boundaries of the modern firm are blurred in comparison to those of the traditional vertically integrated firm (see also Houssiaux 1960). Houssiaux (1957b) used the example of the notion of “groups”, which refers to an ownership structure of firms composed of the hub-firm and its subsidiaries, to recall that quasi-integration is not based on financial instruments, but on technical cooperation between legally independent firms aiming to produce a common product. He added the important idea that the notion of responsibility no longer applied to shareholders (as in groups), but to users. Clearly, this argument tends to moderate the role of ownership in the coordination of the productive activities of firms. This argument led Houssiaux (1957a) to note that quasi-integration transforms the interdependence relationships between firms in dependence relationships influenced by the “dominating power” (ibid., p. 244) of one of the parties. In this spirit, he wrote that “the domination effect grows with the degree of exclusivity of quasi-integration. It also grows with the degree of intervention of the big firm in the management of smaller firms” (ibid.). Thus, the view of Houssiaux is compatible with the analysis in terms of power and differs from the point of view of Williamson, who rejects power from the economics of the firm (see later). Additionally, Houssiaux reminded us that domination in quasi-integration depends on the stability and durability of inter-firm cooperation. For him, power relationships in quasi-integration are not unilateral because it is very damaging for the hub-firm to break up its collaborative relationship with its special subcontractors.
36Even though quasi-integration includes legally autonomous firms, it is evident for Houssiaux (1957b) that the strong links of cooperation between firms also imply strong forms of economic dependence and power. Hence, he made the distinction between “direct external integration” (financial vertical integration) and “indirect external integration” (quasi-integration) based on complex multilateral power relationships. As Houssiaux explained, it is crucial to reconsider the fact that because quasi-integration implies strong forms of power, it could lead to important social risks. Nevertheless, the empirical investigation by Houssiaux in Europe showed that the large hub-firms do not use their strong power on subcontractors, so “quasi-integration is based on best intentions and the collective interests of parties. So we can predict the generalization of quasi-integration during the next decades” (ibid., p. 411). Clearly, the emergence and development of global network-firms during the three last decades confirms the relevance of Houssiaux’s work and sheds light on an empirical limit of the Williamson’s theory. These complex economic organizations have brought power relationships into play such that it is very urgent to reconsider power in the analysis of the functioning rules of quasi-integration forms. The seminal work of Blois (1972) is very informative in this sense and should be reconsidered in light of real-world practices.
Blois and the integration of power and economic dependency in the analysis of quasi-integration
37A major development in vertical quasi-integration was proposed by the business scholar Keith Blois in 1972. In his seminal paper, Blois does not quote the works of Williamson (he uses his works in his more recent papers) or Houssiaux. Contrary to Williamsonian theory, Blois analyzes the implications of the blurring boundaries of the firm beyond the “legal entity of the firm” from a power perspective based on economic dependency and asset specificity. In a 1979 paper, he again argues that the majority of studies of organizations give little attention to their interaction with what he called “the rest of the world”. Thus, Blois writes that “one form of inter-organizational relationship which appears to be growing in importance and which shows that legally defined organization boundaries do not always bear much relation to the existing boundaries” (ibid., p. 55) .
38Blois’s (1972) view on quasi-integration is not based on the Williamson’s intermediate zone between the firm and market, but on vertical inter-firm networks that take advantage of the benefits of vertical integration without suffering from its drawbacks . He is interested in organizational forms “where some firms are gaining the advantages of vertical integration without assuming the risks of rigidity of ownership – a situation which might be described as vertical quasi-integration” (ibid., p. 253). In this view, the global network-firm is clearly a form of what Blois calls vertical quasi-integration. Interestingly, he concludes his paper by writing that “the influence of some organizations spreads much further and affects supplier’s suppliers, e.g., by specification of particular raw materials or sub-components” (ibid., p. 270). Unlike Williamson (who uses the argument of moving-hybrids) and in the spirit of Houssiaux (who speaks about a third stable economic organization), Blois underlined the durable and specific aspect of vertical quasi-integration forms beginning in the early 1970s.
39As such, Blois (1972) sheds light on the power relationships that characterize the process of cooperation between firms and is interested in the nature of these relationships. He considers that one of the more important characteristics of vertical quasi-integration is the exercise of the power of the hub-firm over its suppliers. He writes, for example, that “it would seem that there is some evidence that many suppliers in a variety of industries are finding it very difficult to maintain their managerial independence from large customers” (ibid., p. 269). The hub-firm’s power results from the fact that suppliers are strongly dependent upon the hub-firm for output production. Blois (1972) thus takes into account the situation where a hub-firm decides to change its dedicated supplier to show that, in this case, specialized suppliers are in a very difficult position that could jeopardize their durability and survival. As Diamantopoulos (1987) notes, “according to Blois (1972), it is the underlying threat of possibly losing a considerable proportion of one’s business that puts the supplier at a disadvantage and confers to the customer an important bargaining weapon” (ibid., p. 185).
40One of the main objectives of Blois’s 1972 paper is to analyze the conditions leading suppliers to be economically dependent upon the hub-firm (see also Blois 1977, 1978). For him, the hub-firm has considerable bargaining power  over its suppliers, and he argues that the larger the hub-firm grows, the stronger its power on the subcontracting parties becomes. This economic power is often exerted to obtain special conditions and terms of trade. According to Blois (1972, p. 268), “this type of relationship is the type often found within vertically integrated organizations” so that “such a relationship between customer (hub-firm) and supplier can be considered as a type of vertical integration without legal firm”. He adds that such a conclusion implies the need to consider what is meant by a firm; in Penrose’s 1959 view of the firm based not on the legal concept but on the integration of planning, the boundaries of the firm are blurred. Vertical quasi-integration is indeed characterized by the integration of planning. Hence, Blois (1972) differs from Houssiaux and Williamson’s points of view in the sense that vertical quasi-integration is not located between the firm and the market but is close to being a firm as such.
41Even though, according to Blois (ibid.), the power of the hub-firm often seems to be unilateral, he describes some conditions for which inter-firm power relationships are more balanced and fruitful for suppliers that can be active. In a view close to the one advocated by Williamson (1985) concerning the risks of hold-up, Blois explains that when knowledge is not useful outside a particular inter-firm relationship, the supplier becomes tied into a position of dependence, but the hub-firm is also in a position of dependence because of the large economies of scale and the associated cost of building up a new organizational team with suppliers – concerning the latter, he uses the term “sunk costs” (Blois 1979, p. 61). This form of “power of resistance” (Blois 1972, p. 266) clearly lies in the specificity  of the supplier’s activities and could also be applied to the case of legal vertical integration:
this may seem to be in contradiction of the point made earlier – namely, that specificity makes a supplier vulnerable to its customers. However, it follows from the fact that skills, machinery, etc., are highly specific and not immediately as effectively applicable elsewhere that it would take time for a customer to collect together the necessary skills and machinery to produce the product itself.
43It is important to note here that Blois prefigured in 1972 the theory of inter-firm hold-up notably developed later by Klein, Crawford and Alchian (1978) and Williamson (1985). Additionally, Blois proposes some arguments to show that de jure vertical integration is not always economically possible for the hub-firm (even in the presence of opportunistic behaviors) and thus completes the analytical structure of Williamson. Contrary to Houssiaux, Blois does not neglect the Williamsonian notion of human opportunism. It seems evident that opportunism and power are closely linked to each other (Baudry and Chassagnon, 2010). Nevertheless, and despite the exercise of power relationships, Blois (1979) argues that the common element of vertical quasi-integration is cooperation and commitment by the parties involved in the inter-firm production process. For him, “the essence of co-operative arrangements seems to be that the parties involved accept some degree of obligation (or in other words give some assurances) about their future conduct” (ibid., p. 56). In other words, unlike Williamson and in the view of Houssiaux, Blois considers that vertical quasi-integration in which we can locate the global network firm is a specific and durable governance structure that will not necessarily be transformed in time in de jure vertical integration due to the high specificity of the supplier’s activity. From this perspective, Blois’ line of thought is close to the arguments developed by Richardson (1972) that saw behind hierarchy and cooperative organizational forms of production that are by nature very specific.
44Additionally, Blois (1990, 2001) has more recently commented on the transaction cost theory of Williamson. His very critical appraisal of Williamson’s work raises two important points. On the one hand, Blois affirms that, contrary to Williamson’s (1985) argument, vertical quasi-integration is a new form of ?administration’ in which the behavior of the involved parties cannot be perceived and observed from contractual arrangements. Blois (2001) takes the example of the firms Baird and Marks & Spencer to show that a subcontractor constituting a single and autonomous legal entity (Baird) is in fact a part of an integrated structure (with the name Marks & Spencer) that is not recognized by law. Chassagnon (2011b) explains that, in this case, we can speak about de facto vertical integration: “firms establish these privileged relationships through social interactions with each other because there is de facto vertical integration (that is, unspecified contracts, long-term goals and tasks structured by power) across the legal boundaries of the firms involved” (ibid., p. 120). On the other hand, Blois (2001) is very critical of the failure to include the concept of power in the analysis of inter-firm networks and quasi-integration. Blois affirms that the transaction cost theory should be completed by other theories based on power relationships and resource dependency. Indeed, he argues that power situations cannot be eliminated ex ante by farsighted contracting, as Williamson argues . For Blois, power relationships are inherent to the existence and growth of vertical quasi-integration forms, and it appears that often, as Emerson (1962) has shown, “the power of A over B is equal to and based upon the dependence of B upon A” (ibid., p. 32). However, this result does not imply a unilateral and hegemonic form of inter-firm power, as we will see now.
45Blois’s point of view on power in inter-firm networks was developed further by Diamantopoulos, who proposed revisiting vertical quasi-integration. In his 1987 paper, he in fact tries to complete the approach that Blois began. He starts with the fact that Blois bases his analysis strictly on economic power in a mainly coercive fashion to go beyond economic power and shed light on other sources of power. Indeed, according to Diamantopoulos, it is necessary to move in the direction of a more balanced analysis of power “since both economic and non-economic power sources are instrumental in determining the total power available to a firm” (ibid., p. 192). He agrees that vertically integrated systems are by nature very different from conventional markets and considers that integrated systems imply centralized authority and often a single decision center that coordinates the vertical supply chain. However, he goes beyond Blois’s analysis of inter-firm power and laments that “in the analysis of vertical quasi-integration undertaken by Blois (1972) the only form of power considered is economic power, the exercise of which forces the supplier to comply with the demands of the customer” (Diamantopoulos 1987, p. 187).
46He does not reduce power to economic power based on absolute dependence, instead adding to it five types of power: (1) reward power; (2) coercive power; (3) legitimate power; (4) referent power; and (5) expert power. One of the more important of his associated conclusions is that the central and traditional power in vertically quasi-integrated systems is informal – de facto in the sense used by Chassagnon (2011a, 2011b, 2013) – and based notably upon the hub-firm’s reputation and coordinating role. However, above all, this power is not exclusive but joint. He writes that:
a balanced power relationship may still obtain in a vertically quasi-integrated system even if power differentials exists with respect to particular decision-making areas, by trading spheres of influence, explicitly or implicitly, the transacting firms may be able to achieve an equilibrium situation in terms of the distribution of scopes of influence.
48Whereas Blois analyzes power in terms of direct and absolute economic dependence, Diamantopoulos (1987) clearly adds the hub-firm’s dependence on complementary suppliers. This means that the concept of power-dependence, when it is applied to hub-firm-supplier dyads, implies that both the hub-firm’s and the suppliers’ trading alternatives must be considered. Power is not unilateral but multilateral in cooperative inter-firm relationships because both hub-firms and suppliers have power in the exchange network .
49To sum up, for Diamantopoulos, inter-firm power “should be conceptualized in terms of relative rather than direct dependence and that mutual rather than unilateral influence is likely to be manifested in vertically quasi-integrated systems even in the presence of power asymmetry” (Diamantopoulos 1987, p. 191). We think that this argument should be reintroduced to the analysis of the global network-firm, where long-term relational arrangements are strongly based on de facto multilateral inter-firm powers. This conclusion on inter-firm power shows the relevance of Blois’s 1972 argument that should be used to better analyze the nature of complex economic organizations of the new capitalism, their boundaries and their associated dysfunctions. Additionally, following MacMillan and Farmer (1979), the view of vertical quasi-integration – which they define as “a confederation of firms in vertical relation to one another” (ibid., p. 278) – based on formal and informal power could be used to reconcile the positions of Williamson and Blois:
we believe that many of the benefits that Williamson would ascribe to internal hierarchical systems, and which Blois would ascribe to the existence of market (economic) power, can be gained by both buyer and seller through the development of mutual respect, understanding and trust in a more informal manner.
51To conclude, it is very important to note that Blois, in his 1979 paper, remained conscious not only of the crucial policy and regulatory implications of the definition of the boundaries of the firm but also of “industrial and economic policy and about understanding how modern industrial organizations interact” (Blois 1979, p. 62). More exactly, he writes that “the existence of quasi-integration indicates, in an economy where there are large organizations, the difficulty in defining what a firm is. This has important implications for government policy with regard to the control of industry and economic development for if the size of a firm is defined according to the available economic indicators or legal definition” (ibid., p. 60). This remark is worthy of being reconsidered in light of the “new” real world characterized by the demands of regulating a new form of capitalism made of ultra-dominating global network-firms that have strong power without collective legal responsibilities and boundaries.
52In this paper, we adopted an economic analysis methodology and in some part a business history perspective to analyze three fundamental contributions to the debate on the process of vertical quasi-integration. We conclude that Williamson’s theory of vertical integration operationalization and empirical scope constituted major contributions in organizational economics. However, the introduction of hybrids in his analysis presents some limitations due to the confusing analysis of their stability and specificity (the Williamsonian approach does not give an intrinsic existence to hybrids that exist only ?artificially’ in comparison to market and hierarchy, i.e., to two polar modes based on trading contracts and employment contracts, respectively), the overestimation of asset specificity and hold-up risks (actors’ opportunism) in the analysis of the institutions of capitalism and the rejection of power (and relational factors) from his analytical framework. It seems that we find behind these methodological commitments the aim of proposing a strong and global theory of vertical integration based on precise behavioral and functional hypotheses. His main theoretical objective is completed with the wish to advance toward a strong empirical and operational background on firms’ boundaries.
53On the contrary, and to propose some research directions, we think that what Houssiaux and Blois call vertical quasi-integration closely corresponds to the modern real-world industrial practices among which we find the global network-firm. This new organizational form is a stable and specific real entity based on intrinsic network governance mechanisms that is worthy of being better analyzed and described by taking into account different forms of inter-firm powers resulting from systemic interdependency. This conclusion can be explained by the fact they did not really want to advance toward a strong theoretical framework based on precise hypotheses; in fact, they propose a more inductive perspective on vertical organization forms. In this view, it seems fundamental to reconsider the seminal and little-known contribution of Houssiaux and the innovative developments of Blois that are still accurate at the present time. The argument on the specificity of vertical quasi-integration implies going beyond both the false simple dichotomy of the make-or-buy decision and the strange hybrid category and approaching the problem as a singular cooperative inter-firm governance structure, that is to say, as a third distinct organizational form based on relational vertical re-integration. We leave this suggestion for future studies in business and organizational economics.
54Additionally, it seems interesting to better understand why unlike to Houssiaux and especially to Blois Williamson underestimates both power relationships and informal social coordinating mechanisms in the building blocks of his theory. The position of Williamson concerning power is astonishing regarding the behavioral assumptions he uses, namely the bounded rationality and the opportunism of economic agents. These hypotheses theoretically lead to potential conflicts and thus to power relationships, which seem not to be far from Williamson’s analysis. In fact, for Williamson, power rests on ex post situations. On the contrary, he argues that, on the one hand, transaction should be studied in its entirety – i.e., both ex ante and ex post – and, on the other hand, that economic agents are not myopic but anticipate possible power positions. In other words, economic agents could protect themselves against power situations (Baudry and Chassagnon, 2010). Williamson consciously made the choice to not use power and its associated concepts (apart from authority) in order to focus on a discrete and comparative analysis of governance structures – giving no place to political forms of regulation of economic relationships. He has proposed what he calls a “pragmatic methodology” (see Williamson 2009) based on four precepts: (1) keep it simple; (2) get it right; (3) make it plausible; (4) and prediction and empirical testing. In this view, Williamson considers, for example, about the seminal work of Coase that “the missing ingredient was lack of operationalization” (Ibid., p. 149).
55According to Baudry and Chassagnon (2012a), the Williamson’s methodological framework can be invoked to justify his controversial position. For Williamson (2009), the theoretical success of transaction cost economics is based on the operationalization of both Coase (1937) and Commons (1931). He adds that his theory of the firm yields numerous refutable implications. Hence, a methodological concern of Williamson is surely about the operationalization of the insights and concepts he brings from other perspectives and disciplines. Needless to add that the concept of power such as it is used in organizational theory cannot be easily adapted with the Williamson’s methodological commitments. It would be unsuitable to introduce power in the economics of the firm because it would be impossible to operationalize this “vague” concept. This can explain the fact that Williamson does not benefit from the important organizational contributions to power analysis. Notably, a curious point in Williamson’s analysis of power rests on the poor consideration for resource dependency organizational theory. Williamson contents himself with quoting some important authors in organization theory such as Selznick (1957) or Perrow (1986) and conceding that it suffices for his purpose “to remark that power, for a long time, has been the congenial organization theory perspective” (Williamson 1990, p. 182). Arguably, it is interesting to understand more precisely why Williamson does not give great attention to the concept of power as it is used by resource dependency theory (see Baudry and Chassagnon 2010). If it seems possible to reproach Blois and especially Houssiaux for underestimating in some parts the risk of opportunism in inter-firm relational contracts (see Baker, Gibbons and Murphy 2008), the neglect of power issues and the focus on opportunism seem to show that Williamson’s theory “is vulnerable to the reproach that production is neglected in favor of allocation” (Hodgson 1998, p. 31). We think that by basing his theory on the contractual man, Williamson has chosen to liberate economics from non-orthodox political economy.
Directeur de la Recherche, Professeur d’économie-HDR en économie, ESDES-Université Catholique de Lyon. ESDES, 23, place Carnot 6902 Lyon. firstname.lastname@example.org.
De jure vertical integration refers here to formal hierarchy.
Different from formal hierarchies; see Dietrich (1994).
Williamson cited neither Houssiaux nor Blois.
Blois’s argument here is close to the analysis of the governance of modern firms led by Rajan and Zingales (2000).
Strong disadvantages such as: “(1) disparities between production capacities at various stages of production, (2) adverse public opinion and governmental pressure, (3) lack of specialization, (4) inflexibility of operations, (5) over-extension of management team, (6) lack of direct competitive pressure on the costs of intermediate products” (Blois 1979, p. 57).
Recently, Blois (2010) has questioned the legitimacy of such a power.
Blois (1979, p. 59) considers that “the degree of flexibility or the specificity of a supplier is a concept concerned with the idea that an organization is committed to a certain product range and/or number of markets in such a way as to make it difficult for it, in the short-run, to produce another range of products and/or enter another market while maintaining its current level of profitability. This being a situation described in the economics literature as the existence of Barriers to Exit”.
For Williamson (2009, p. 150), “all complex contracts are unavoidably incomplete (by reason of bounded rationality) yet human actors are assumed to have the capacity to look ahead, recognize hazards, work out the mechanisms, and, albeit imperfectly, factor the ramifications back into the ex ante contractual design (by reason of feasible foresight)”.
This result seems to be in part empirically supported by the work of Berger (2006).