1. The development of cooperative relations within and between firms

During the 1980s, a new paradigm emerged in the labour market and organizational literature to analyse current trends in organizational forms and employment relations. Inspired by the best practices of Japanese firms, as well as the most successful American and European firms, this approach contends that the bureaucratic employment system that came to dominate the organizational landscape during the post-World War II period is leaving the way to a new employment system based on strong mutual commitment between employers and employees.

The bureaucratic employment system is characterized by a high degree of work standardization and specialization. Key components of the employment relationship are governed by rules and procedures. Promotion and firing are based on seniority, and wages are tied to the job rather than the individual. Jobs are also linked between themselves by promotion ladders starting from entry-level positions, the "ports of entry" in the organization [Doeringer and Piore, 1971, p. 2]. Worker skills focus on the tasks related to a given position and are mainly developed through informal on-the-job training. Osterman [1988] highlights the internal logic of this employment system: while management gives up some of its prerogatives by negotiating formal rules of work organization, it retains a source of flexibility in its freedom to adjust the volume of employment through hiring and firing. Labour market theorists emphasize the role of unions in the development of workplace regulations [Edwards, 1979; Jacoby, 1985; Kochan et al, 1986], and organizational theorists stress that the bureaucracy fits a stable product-market environment, in which organizations focus on internal efficiency [Burns and Stalker, 1966; Miles and Snow, 1978]. Homogeneous and steadily growing markets after World War II supported this competitive strategy, and bureaucratic systems have been a dominant feature of large blue-collar work settings during this period.

Conditions of stable growth have now been replaced by mutually reinforcing trends toward diversified and unstable output demands, globalization, and increased competition [Campbell, 1991; Dicken, 1992]. In this new competitive environment, firms have to develop product quality and adaptability to market changes [Hayes, Wheelwright and Clark, 1988; Dertouzos, Lester and Solow, 1990; Porter, 1990]. The commitment employment system supports these strategic objectives through enlarged and flexible work assignments. Trust and loyalty are here the governing mechanisms of the employment relationship [Ouchi, 1980, 1981]. Employees' work ethics are carefully weighted in hiring decision, and wages are tied to individuals rather than job attributes. It is implicitly agreed that workers will produce extra-efforts to support the firm's strategic objectives, and that the firm will recognize their contribution through adequate compensation levels as well as employment stability. Employee skills development plays an essential role in commitment systems due to broad and changing work assignments. Learning is part of the job, and includes technical, interpersonal as well as general skills. Osterman [1988] defines the internal logic of this employment system: the numerical flexibility of employment is restrained by employment security policies, but functional flexibility is enhanced by the possibility to redeploy employees' tasks when needs arise from changes in the firm's products or processes. Commitment systems have traditionally been associated with white-collar employment [Jacoby, 1985; Osterman, 1988], but they develop in blue-collar settings due to changing environmental conditions [Osterman, 1988; Walton, 1985]. In addition, firms react to increased environmental complexity and uncertainty by developing long term relations with their close trading partners, based on trust and commitment and supportive of transaction skills development [Aoki, 1990; Best, 1990; Badaracco, 1991; Johnson and Lawrence, 1988; Piore and Sabel, 1984; Powell, 1990]. In this perspective, firms become tied together within productive networks in the same way that employers and employees are tied together in employment systems. While the bureaucratic firm had clearly defined boundaries, with the rule as inside regulating mechanism and the market as outside regulating mechanism, the frontiers of the flexible organization are blurred as the same device, informal commitment, binds the firm to its employees and close trading partners.

The theoretical foundations for a progressive replacement of the bureaucratic system by the commitment system can be found in contingency theory, that relates firms' environmental and organizational characteristics [for example, Burns and Stalker, 1966; Lawrence and Lorsh, 1967], and in labour market theory, that integrates firms' employment practices into coherent systems [Doeringer and Piore, 1971; Osterman, 1987; Piore, 1975]. Indeed, this highly stylized discussion summarizes a variety of findings from several overlapping fields: industrial relations and labour market research [Appelbaum and Batt, 1994; Brown, Reich and Stern, 1991; Kochan and Osterman, 1994], economics [Aoki, 1990; Best, 1990], organizational studies [Badaracco, 1991; Dertouzos et al., 1990; Hayes et al., 1988; Powell, 1990; Walton, 1985], as well as human resource management studies [Dyer et Holder, 1988; Fombrun, 1984; Schuler and Jackson, 1987; Ulrich, 1991].

Why do we find such a widespread interest for this new organizational model in the literature? One reason for its attractiveness is that it provides a way to unify the interests of the firm and those of the individual. While the principle of the rule suggests some forms of arbitration between opposed interests in the bureaucratic system, the principle of trust and commitment carries a sense of shared interests within the firm, or between the firm and its close trading partners. In the United States, influential publications also argued that the commitment system could provide an avenue for restoring both the country's competitive edge and living standards. If American firms remain stuck with bureaucratic systems, the argument goes, its productivity gap compared to competitors will aggravate, as will the decline in real average earnings [Commission on the Skills of the American Workforce, 1990; OTA, 1990]. The superiority of the new organizational model is demonstrated by numerous case studies that show firms' impressive performance achievements under the new system [for example, Adler, 1992; Berg et al., 1994; Cutcher-Gershenfeld, 1991; MacDuffie and Krafcik, 1992].

However, the question of its diffusion among American firms remains problematic. Large-scale surveys show some diffusion of new workplace practices [Osterman, 1994], but findings can be inconsistent [Lawler, Mohrman, and Ledgord, 1992] or contradictory [Bailey, 1994]. These mixed results can be explained in a variety of ways. First, the large-scale measurement of organizational forms is hampered by methodological problems inherent to the "bureaucratic versus commitment" thesis. Indeed, this approach relies on "ideal types" of organizations and employment systems, consistent with a rich tradition in the organizational and labour market literature. However, an ideal type is:

formed by the one-sided accentuation of one or more points of view and by the synthesis of a great many diffuse, discrete, more or less present and occasionally absent concrete individual phenomena, which are arranged according to this one-sidedly emphasized viewpoint into a unified analytical construct. In its conceptual purity, this mental construct cannot be found empirically anywhere in reality [Weber, 1949, p. 90].

The idea that the new model will become predominant among organizations is also problematic on a theoretical level. Indeed, its major limitation lies in the interpretation of current environmental changes and their organizational requirements. This view stresses the move toward a more complex and changing environment, but considers only the qualitative aspects of this evolution, such as constant technological innovation and the atomization of markets, that create a need for continuous organizational adaptation and long term investment in human resources. However, an equally important pressure stems from quantitative aspects of changes in firms' competitive environment. Conditions of stable growth no longer characterize developed economies, and firms must adapt to mature and erratic markets. Unpredictable changes in the level, and not only the nature, of market demand create a continuous pressure to reduce the fixed costs of the organization, including long-term employment relations. Thus, on one side, firms need to develop a stable pool of human resources to enhance their adaptability to qualitative changes, but on the other side, they need to reduce fixed costs linked to human resources. How can firms meet these contradictory requirements? The literature on the cooperative model does not answer this question, as it considers only part of the pressures exerted on employment practices. A closely related limitation of this model lies in its focus on long-term employment. Indeed, both the bureaucratic and commitment employment systems emphasize long-term relations between firms and their employees. The theoretical rationale for long term relations lies in the development of firm-specific skills by the individual, a principle first defined in human capital theory [Becker, 1964; Oi, 1962], and later integrated in internal labour market theory [Doeringer and Piore, 1971] as well as transaction cost theory [Williamson, 1975]. What about situations where individuals develop skills that are not firm-specific, and engage in employment relations on a short-term basis?

2. The variety of employment systems and organizational forms

A closer look at the labour market and organizational literature reveals a greater variety in firms' employment patterns, including the traditional and the professional employment systems, both based on employment instability and low firm-specific skills.

The traditional employment system is characterized by little employee skills development and employment stability. High turnover, arbitrary work rules, limited skills requirements, and reduced opportunities for upward mobility are defining characteristics of this "secondary labour market" [Doeringer and Piore, 1971]. The market appears as the central coordinating mechanism of the traditional system. As Williamson [1975] argues, prices can regulate the employment relationship when work is low skilled and easy to monitor, so that turnover is not costly to the firm. Although the traditional system has received little attention from the organizational and strategic management literature, one can infer from its characteristics that it fits a competitive strategy based on cost reduction. While the bureaucratic system emphasizes a reduction of cost per production unit, the traditional system also minimizes fixed employment costs by adjusting employment and earning levels to variations in output demand. This flexibility rests on unilateral hiring and firing decisions, as well as pay systems based on hourly or piece rate. The traditional employment system is associated with a highly unstable output demand, a situation typical of manufacturing activities in the late nineteenth and early twentieth centuries, and that persisted in a few isolated sectors during the post-war expansion [Doeringer and Piore, 1971; Jacoby, 1985]. On a conceptual level, it fits an organizational environment both unstable, due to fluctuations in output demand, and simple, due to low skills requirements.

In the professional system, employment is also unstable but employee skills development plays a major role. Skills are largely acquired outside the firm through education, and developed along the individual's career through frequent changes in work assignments and work settings. Consequently, they tend to be profession- rather than firm-specific, and firms tend to buy them through high wages rather than develop them. Professional positions are specialized, reflecting expert knowledge, but individuals enjoy a substantial degree of autonomy in their work. They tend to identify more with their profession than with their employer [Burns and Stalker, 1966; Piore, 1975], which creates a continuous ambiguity between the objectives of the firm and those of the individual. Initially associated with craft workers in the labour market literature [Doeringer and Piore, 1971], professional labour markets later enlarged to occupations with skills development processes located outside the firm [Osterman, 1984; Piore, 1975]. In the organizational literature, this employment system supports competitive strategies based on continuous innovation, in a complex and turbulent product-market environment [Burns and Stalker, 1966; Miles and Snow, 1978].

Figure A relates the bureaucratic, commitment, traditional and professional employment systems to a dominant type of competitive strategy and coordinating mechanism (see appendix). In this four-cells framework, employment systems are differentiated along the key dimensions of employee skills development and employment stability. Employee skills development are central to commitment and professional systems, but limited in bureaucratic and traditional systems. They are the responsibility of the firm in bureaucratic and commitment systems, but rest mainly with the individual in professional and traditional systems. The four employment systems are also associated with distinct environmental conditions. In bureaucratic and commitment systems, employment stability is supported by a stable competitive environment, while professional and traditional systems allow firms to deal with environmental instability. In complex environments, firms rely on employee skills through commitment and professional systems, while bureaucratic and traditional systems fit less sophisticated environments. Thus, an important quality of this framework is to stress that the adoption of the commitment system requires some environmental stability allowing for the development of long-term employment relations.

As shown in Table 1, this four-cells framework provides a synthesis of various typologies developed in the labour market, transaction cost, organizational and human resource management literature (see appendix). Based on congruent characteristics in labour market and organizational research, it defines key configurations of the employment relationship [Osterman, 1987]. It should be stressed, however, that frontiers between employment systems are neither clear nor stable. Indeed, in Weber's words, each of these employment systems is a "one-sided accentuation" of the employment relationship, stressing the need for arbitration and standardization in the bureaucracy, social bonding in the commitment system, innovation in the professional system, and labour cost flexibility in the traditional system.

3. The core-periphery model as an alternative conceptual framework

On a theoretical level, this typology of employment and contracting systems can help us solve the puzzle of conflicting environmental pressures on firms' organizational forms. In response to increased environmental complexity, the commitment system could replace the bureaucratic system for the core competencies of the organization, but in response to increased environmental instability, professional and traditional systems would develop to manage competencies that are peripheral to the organization. A complete picture of current organizational changes would thus include (1) a transformation of internal labour markets toward a more flexible work organization, and (2) a diminution of their relative importance together with the growth of external labour markets, based on traditional employment systems for low-skilled work and professional systems for high-skilled work. In that perspective, the adoption of the commitment system implies that firms use some complementary forms of unstable employment in order to buffer core workers from environmental instability. This core-periphery view also applies to inter-firm relations, as firms may develop stable links with their core partners, but meet sudden changes in output demand by resorting to short-term, unstable trading relations with peripheral partners.

The core-periphery thesis has been developed by Osterman [1988], Pfeffer and Baron [1988], Noyelle [1987] and Hirschhorn [1988] in the United States, Atkinson [1985] in the United Kingdom, as well as Piore [1980] and Standing [1986] in an international perspective. Noyelle [1987] emphasizes the training dimension of the core-periphery strategy, that translates into an increased transfer of training costs and implementation outside the firm. Pfeffer and Baron [1988] distinguish three dimensions of the "externalization of work": physical location (homework), employment duration (part time and short term employment), and administrative control (temporary help services, employee leasing, business services). They argue that:

because externalization enhances flexibility, it may actually be an essential concomitant of human resource policies that emphasize long term employment, commitment to the permanent workforce, and a shared vision or distinctive competence (p. 274).

Osterman [1988] considers that both the "industrial" (bureaucratic) and the "salaried" (commitment) models are under strong pressure due to the cost of providing employment security, so that firms would limit the size of their core workforce and resort to peripheral workers as a buffering mechanism. Yet he doubts the viability of core-periphery strategies on the ground that peripheral workers lack the commitment necessary for firms to achieve competitive advantages based on quality and differentiation. This would probably hold true if the same tasks and skills requirements applied to core and periphery workers. However, firms may limit the potentially disruptive impact of peripheral workers by organizing their production process so that tasks that are essential to their competitive strategies remain assigned to core workers, while peripheral workers are assigned to tasks that are either highly standardized, easy to monitor, or loosely coupled to the production process. Moreover, the higher risk associated with employment instability in periphery positions might stimulate workers' performance just as much as a security promise.

The inter-firm dimension of core-periphery strategies has been addressed by Miles and Snow [1986], and more recently Harrison [1994]. Miles and Snow [1986] discuss how their ideal types of firms might play complementary roles within a given industry: "analysers" (quality strategy) adapt and rationalize innovations produced by "prospectors" (innovation strategy), and leave mass market applications to "defenders" (cost reduction strategy). The various types of firms are linked together within a "dynamic network", that provides diversity and flexibility to the industry. An important implication of Miles and Snow's framework is that firms differentiate employment systems on the basis of tasks differentiation within a given industry, so that the functions performed in the traditional, professional and commitment systems are different and complementary. Harrison [1994] analyses the disintegration of the large bureaucracy and the emergence of network forms of organization based on smaller production units. He contends that this trend is accompanied by a growing segmentation of the labour force, the "dark side" of flexible specialization.

Thus, core-periphery strategies can develop at two levels: direct employment relations and inter-firm trading relations, a distinction increasingly blurred by the move away from rules within these two spheres. They can also be studied from several perspectives.

First, firms' employment and trading practices reveal core-periphery strategies from the perspective of a given firm. Along these lines, the buffering of a core workforce by peripheral workers is documented by case studies or small-scale surveys in the American automobile, electronics, finance and service industries [Abraham, 1988; Bertrand and Noyelle, 1987; Hirschhorn, 1988; Mangum, Mayall and Nelson, 1985; Osterman, 1988] as well as various industries in the United Kingdom [Atkinson, 1988; Hakim, 1990; Walsh, 1990].

Second, firms' employment characteristics highlight the impact of core-periphery strategies from the perspective of inter-firm differences in employment systems. Such differences appear in large-scale surveys of American manufacturing [Ichniowski, 1990], as well as French manufacturing [CEREQ, 1987] and electronic industries [Delattre and Eymard-Duvernay, 1983]. They are also documented by case studies of contracting systems in Italian industrial districts [Brusco, 1982; Taplin, 1989], Japan's manufacturing industries [Kumazawa and Yamada, 1989; Smitka, 1991], and in the production networks of multinational firms such as Nike [Donaghu and Barff, 1990] or Benetton [Harrison, 1994].

Third, labour market characteristics highlight the impact of core-periphery strategies from the perspective of the workforce. Large scale surveys reveal a segmentation of the American workforce that can be conceptually tied to firms' core-periphery strategies. This segmentation appears in the polarization of wages since the mid-1970s [Harrison, 1994; Mishel and Bernstein, 1994], the growth of contingent forms of employment such as part-time and temporary work [Callaghan and Hartman, 1991; Ehrenberg, Rosenberg and Li, 1988; Hartmann and Lapidus, 1989; Tilly, 1992], and the overall decline in employment stability [Marcotte, 1994; Swinnerton and Wial, 1995]. Other studies stress differences among occupational groups in terms of earnings and benefits, skills requirements, working conditions, and employment status [Gittelman and Howell, 1994], as well as access to training [BLS, 1994]. Finally, a number of industry studies combine several perspectives and highlight the segmentation of the workforce resulting from firms' core-periphery strategies in the American motion picture industry [Christopherson and Storper, 1989], and high tech industries [Colclough and Tolbert, 1992; Lozano, 1989].

The existence of core-periphery strategies in Japan's manufacturing, Italy's industrial districts, and America's high tech industries is of special significance because these industries have inspired much of the literature on the emergence of the commitment system as a new organizational model. Indeed, the "dark side" of flexible production systems has been consistently overlooked in mainstream organizational literature, so that additional research is needed to improve our understanding of how core-periphery systems develop, and what their implications are in terms of firms' competitiveness and employees' welfare.