Changing Regimes of Workplace Governance, Shifting Axes of Social Mobilization, and the Challenge to Industrial Relations Theory



    1. Department of Economics, MIT
    Search for more papers by this author

    1. Graduate School of Business, University of Chicago
    Search for more papers by this author
    • *

      The authors’ affiliations are, respectively, Department of Economics, MIT; Graduate School of Business, University of Chicago. E-mail: We have accumulated debts to a number of people, too great to be acknowledged here. But we were particularly influenced by conversations with the students and faculty of the IWER group at MIT, and with Frank Dobbin, Deborah Kolb, Victoria Hattam, Natasha Iskander, Chris Desan, Sarah Ashwin, Carola Frege, John Logan, Bob McKersie, and Andrew Schrank.


This article challenges prevailing views about the collapse of the New Deal industrial relations system and the role of the market. It argues that the old system has been replaced not by the market but by an employment rights regime, in which the rules of the workplace are imposed by law, judicial opinions, and administrative rulings, supplemented by mechanisms at the enterprise level that are responsive to the law but also are susceptible to employee pressures, both individual and collective. The emergence of this regime is the product of a shift in the axes of social and political mobilization from mobilization around economic identities rooted in class, industry, occupation, and enterprise to identities rooted in the society outside the workplace: sex, race, ethnicity, age, disability, and sexual orientation. The shift in the axes of mobilization in turn reflects the collapse of the underlying model of social and economic organization upon which the collective bargaining regime was built and more fundamentally a shift in our understanding of the nature of industrial society and its direction of evolution in history. This interpretation poses a challenge to the conceptual tools used in industrial relations to understand the issues of work and to frame the public policy debate. We conclude with some suggestions as to the direction in which we might move to provide an alternative conceptual framework.

It is commonplace among industrial relations scholars and practitioners that over the last twenty-five years the world of industrial relations and workplace governance has been fundamentally, probably irrevocably, altered. The system of collective bargaining that grew out of the New Deal reforms has collapsed. That collapse has been produced by the increasing encroachment of the competitive market and the growing hegemony of neo-liberal ideology that has championed market-oriented reforms not only in labor relations but also in public policy more broadly. These changes have led to increasing disparities in the distribution of income and social welfare in American society, and a fundamental imbalance in political power. This paper challenges this characterization and offers an alternative that suggests both that the United States has pursued a much more progressive social policy than it is generally given credit for and that the regulatory role of the state over this period has been underestimated.

We start from the accepted view that the New Deal collective bargaining system has collapsed. But our argument, laid out in the next section Changing Regimes of Workplace Governance, departs from that view in three critical respects. First we argue that the regime that has replaced collective bargaining is not a market regime at all but rather a regime of substantive employment rights specified in law, judicial opinions, and administrative rulings, supplemented by mechanisms at the enterprise level that are responsive to these rules and regulations but also susceptible to employee pressures. Second, we argue that the emergence of the new regime has been driven, not by neo-liberal ideology, but rather by a shift in the axes of social mobilization from mobilization around economic identities associated with class, industry, occupation, and enterprise to mobilization around identities rooted outside the workplace: sex, race, ethnicity, age, disability, and sexual orientation. Third, the shift in the axes of social mobilization reflects the collapse of the underlying model of social and economic organization upon which the New Deal collective bargaining regime was based. Indeed, the collapse of the New Deal model reflects an even more fundamental shift in our understanding of the nature of industrial society and its direction of evolution in history.

The paper goes on, in the next section, to argue that this alternative interpretation implies significant shortcomings in the conceptual tools we use to understand the issues at work and to frame public policy debates about employment; without new tools, it is difficult to explore the implications of the changes that have taken place. It does not actually provide an alternative conceptual framework, but it draws upon recent developments in economic sociology and organizational studies to suggest directions in which we might move.

Changing Regimes of Workplace Governance

The Regime Shift. Industrial Relations Systems was defined by John T. Dunlop (1958) in a book that shaped the self-conception of the field as a domain of scholarly inquiry and as a community of practice, as that social realm that produces the rules of the workplace. Both the New Deal era collective bargaining regime and the regime that has replaced it do this, but they do so in very different ways. In the collective bargaining regime, the rules were produced by negotiations between employers and representatives of their workers organized into trade unions and administered through practices and procedures specified in union contracts. In unorganized workplaces, employers were nominally free to establish rules unilaterally and change them at will, but large industrial employers were effectively constrained from doing so by the threat that if they departed radically from the practice in unionized shops, their workers would quickly turn to a union to represent them.

The nature of the regime envisaged by proponents of deregulation, and those who view the current regime as market oriented, is never fully specified. But its essential characteristic is the ability of employers to adjust work rules in response to changes in the economic environment and the market pressures those changes produce. Whether a productive system can ever operate in this way without social restraints is debatable. One can argue that economic processes are always mediated and constrained by rules and customs, even if those are “latent” or informal (Polanyi 1957; Kennedy 1997). But, however one might characterize the system of workplace governance implicit in the competitive model, it is not the system that has emerged in place of collective bargaining.

The rules in contemporary U.S. workplaces are as explicit as those under the old regime, yet they grow not out of collective bargaining but rather out of substantive regulation embodied in statutes, administrative rulings, and court decisions and are given coherence by the human resource practices of large corporations and their personnel handbooks and procedures. The main impetus behind the new regulatory framework has been equal employment opportunity legislation. Such regulation has a long history in the United States, but the effort was reinvigorated, and for the first time became serious and effective, when Title VII of the Civil Rights Act of 1964 was passed under pressure from the Black civil rights movement. This law created a new administrative agency that over the course of the next three decades defined what exactly constituted discriminatory practices and how they should be remedied. Through this and subsequent legislation, similar protections were extended to other socially stigmatized and disadvantaged groups, including women, other racial and ethnic minorities, the physically disabled, the aged, and, on local levels, to gays, lesbians, and transsexuals (Skrentny 2002; Wakefield and Uggen 2004). In the 1980s, legislation mandating family leave and advance notice of layoffs was passed. State courts in this period began to impose limits on the doctrine of employment-at-will that has historically governed individual contracts of employment in the United States (Morriss 1995; Autor, Donohue, and Schwab 2002; Edelman, Abraham, and Erlanger 1992). In the 1990s, there has been a proliferation of employment legislation at the state and local level, most notably mandating living wages for contractors of local government activities and, in 2004 in California, paid family leave (Kruse and Hale 2003).

In the attempt to negotiate the complex environment these regulations created, American businesses developed a system of employment relations that has come to be called human resource management (HRM) (Dobbin and Sutton 1998; Kelly and Dobbin 1998). The implications of HRM are far-reaching, but one of its important characteristics has been to generate a set of standard personnel policies and practices that apply to all workers, even those not directly covered by legislation. Pressures from groups similar to those that promoted government regulation but now acting in the shop or enterprise have further shaped the implementation of these policies and practices in the workplace. Their activities have taken a variety of forms ranging from activism and legal action to more subtle forms of internal pressure. Increasingly, these are later codified in the form of written personnel policies meant to protect the company against charges of arbitrary and unequal treatment. Such policies are not easily adjusted in response to changes in the business environment, and hence have some of the rigidities found in the old system.

Finally, beginning in the 1990s, employers developed a new strategy centered on private arbitration procedures for dealing with the growing burden of employment regulation. Employees were required as a condition of hire to sign an agreement to take disputes that might otherwise have been pursued through the courts or administrative agencies to these arbitration procedures instead. These arrangements initially seemed of dubious legality. But in 2001, in the Circuit City opinion, the Supreme Court upheld—indeed encouraged—the use of arbitration in such cases.1 The importance of this decision is parallel to the so-called Steelworkers Trilogy,2 which solidified the legitimacy of collective bargaining by creating an obligation to arbitrate disputes arising over the interpretation of collective bargaining agreements—a procedure that also was not anticipated in the law but which grew up in practice in the course of the early post-WWII period and was then mandated by the courts. If this earlier episode is any guide, the nature of the arbitration procedures under the new system will be further refined by subsequent court decisions. But, unless the Court overturns its ruling, Circuit City makes binding arbitration the capstone of the employment rights regime and incorporates into it essential features of the old.

Thus, the two regimes share an evolutionary pattern. But they nevertheless contrast sharply in the role of government regulation. Substantive legislation under collective bargaining basically sets the minimum labor standards for jobs at the bottom of the wage hierarchy (although there still remained a substantial “uncovered sector”). Above that floor, government regulations focused on procedural standards. These standards effectively created the rights of workers to form and join trade unions, and of those unions, once formed, to engage in collective bargaining. As interpreted by the courts, it came to make a distinction between contract negotiations and contract administration, and required binding arbitration in disputes over interpretation of the agreement. But parties were free to set the substantive terms and conditions of employment through that bargaining process. Contractual coverage was far from universal, but at its height, nearly a third of the private sector labor force belonged to a union. This, plus the threat of organization in the nonunionized sector, was such that union gains effectively spread throughout most of the industrial economy.

The demise of the collective bargaining regime is signaled by the decline in collective bargaining coverage in the private sector from a height of 35 percent of the labor force in 1955 to 8 percent in 2005. The decline was concentrated in the period from 1975 to 1990. Private sector union coverage as late as 1975 remained at 28 percent, but by 1990 was only 13 percent. It is reflected even more dramatically in measures of labor unrest and strike activity. These figures are extremely volatile; the percentage of work time lost in strikes of units of over 1000 employees, for example, varies from a high of 0.38 percent in 1949 to a low of 0.07 percent in 1957, 1963, and 1973. But, in 1980, it began a steep decline: falling to a low of 0.02 percent by 1990, after which it essentially stabilized.

That collective bargaining has been replaced by an employment rights regime is suggested by the rise in employment litigation over this period of time. Data collected and analyzed by the Dunlop Commission on the Future of Worker–Management Relations indicate that lawsuits filed in federal district court on employment law rose from 4331 cases in 1971 to 22,968 cases in 1991, or from 6 percent of all civil cases filed to 16 percent, (Commission on the Future of Worker-Management Relations 1994; Weil 2003). We have been unable to update these figures exactly; but data collected by the Cornell Law Project suggest that further increases in the 1990s were modest (one or two percentage points).3

The spread of the new regime is also suggested by the rapid expansion of private sector arbitration of the type sanctioned by Circuit City. According to one estimate, the number of private sector enterprises with such arrangements rose from 2 percent in 1991 to 7.6 percent in 1996, and 15 percent in 2003 (Colvin 2003). This compares to 8 percent of workers covered by collective bargaining in 2003 (collective bargaining coverage by enterprise is not available). Another marker is a study from the American Arbitration Association, which found that in 1996–1998, 46 percent of its members arbitrated one or more non-union employment disputes and 82 percent arbitrated a dispute within the last three years that required them to interpret a statute (Picher, Seeber, and Lipsky 2000).

Shifting Axes of Social Mobilization.  Whatever the role of the market in the decline of collective bargaining, the rise of the employment rights regime has been driven not by the market at all, but by a shift in the locus of social and political mobilization. The collective bargaining regime emerged under political pressures generated by the mobilization of industrial workers, and unions were organized around a set of identities rooted in craft, profession, industry, and enterprise. The more recent employment laws have been generated by political mobilization around social identities linked to race, to ethnicity, and to personal characteristics associated with social stigma and/or the focus of labor market discrimination such as sex, age, physical disability, and sexual orientation.

The fountainhead of the new regime was legislation responsive to the pressures of the Black civil rights movement, much of which involved extra-electoral politics, demonstrations, and riots, the most dramatic manifestation of which was the 1963 March on Washington for Jobs and Freedom where Martin Luther King delivered his “I Have a Dream” speech. The march led directly to the passage of the Civil Rights Act of 1964, Title VII of which formed the fulcrum around which the new employment law has been built. Not all of the laws and regulations of the new regime were generated by pressure from groups of this kind, but many are the product of a similar process (for example, the Family Leave Act) (Kelly and Dobbin 1999). Mobilization at the state and local level that produced the new living wage laws also developed around these social identities (Reynolds 2001; Fine 2003). Identity groups of this kind have emerged within many enterprises and most professional associations pressing for rights and benefits at this level in much the same way as unions pressed for employment benefits historically (Scully and Segal 2002). Domestic-partner benefits, for example, spread from one enterprise to another in part through grassroots organizations of gay and lesbian employees (Scully and Segal 2002; Briscoe and Safford 2005; Zald, Morrill, and Rao 2005).

While the roles of the new identity groups in the employment rights regime is analogous to those of unions under collective bargaining, the nature of the identity groups and the way in which they are constituted as actors is very different. Unions are formal organizations with a well-defined constituency and a dues-paying membership base that endows the organization with a certain amount of financial autonomy. In the United States, they are rooted in the shop at the point of production; the shop-level organization is embedded in a larger organizational structure that typically encompasses the whole of the enterprise of which the shop is a part and extends to the industry and/or craft or profession with which the workers are affiliated. The union organization extends regionally and nationally as well. The organization is structured to maintain a credible and continuous threat to exercise economic and political power, separately or in a coordinated fashion, at each of these levels.

The principle activities of unions are also well defined. They invariably include collective bargaining and, in some cases, intermediary functions such as the recruitment, training, and placement of labor (this last through union-managed hiring halls). Many unions also provide insurance and various kinds of “mutual aid” that create a separate income stream, giving them further financial autonomy. The law regulates each of these dimensions in a way that gives labor unions legal standing and ensures organizational continuity.

The literature on the groups that generate claims under the employment rights regime is limited (though see Friedman 1996; Stone 1996; Scully and Segal 2002). Our understanding of it draws heavily on our own research which has focused on identity groups in and around the engineering profession, on information gathered incidentally in business organizations that we were studying for other reasons, and on a study of the origin and diffusion of domestic-partner benefits (Piore and Safford 2003a,b; Briscoe and Safford 2005).

Relative to unions, these groups are much more diffuse with a more tenuous organizational structure. They exist at both the grassroots and national levels. The grassroots organizations offer a particularly sharp contrast to unions. They tend to operate as affinity groups offering social support or consciousness-raising. Here the groups in the workplace are not very different from neighborhood groups, sports clubs, student organizations, and alumni groups of the same identity with whom they may share membership or even meetings and activities. The specific goals of these groups—the demands that they make in the workplace—when they have them are extremely various: Some focus on the company's internal policies (diversity, domestic-partner benefits); others on employment-related policies (internships for Black students); still others on the company's role outside (especially charitable contributions, such as scholarships at local universities), or on external support of the groups in question in ways that are unrelated to the company's mission (an official group in a gay pride parade, the boycott of conventions, and sales activities in places with restrictive legislation or practices, for example, in South Carolina when the state offended Blacks by deciding to fly the Confederate flag over its state capital).

Since so much of grassroots organization is informal, it is difficult to quantify. In some companies, however, employee groups have been formally recognized and thus became the centerpiece of a firm's “diversity initiatives.” IBM and AT&T each have over fifty groups including, at AT&T, Asian workers, African American, Jewish, the disabled, Christians, former military personnel, women, and a group for “over 40s.” Some constituencies have more than one network in place at the company. JP Morgan boasts that over 6500 employees that are involved in a variety of affinity groups of Fortune 500 companies have or will soon have employee networks or affinity groups (Diversity Best Practices 2001).

A parallel structure of identity groups exists at the national level. Here the organizations are much more formal, with a paid professional staff. They are concerned with legislative lobbying, litigation, public education, and political campaigns around specific issues. They may have dues-paying members but they draw their principal support from foundation grants, corporate and individual contributions. Their links to the grassroots organization are tenuous and informal, often based on personal relationships. But they do provide technical support and sometimes loose coordination on particular issues, such as in corporate campaigns for domestic-partner benefits. It has been argued that these national organizations with professional staffs divorced from a mass membership base have become typical of American politics more broadly, symptomatic of a decline in social capital and civic engagement (Putnam 1993; Skocpol 1999). But our research suggests that the relationship to citizen involvement is much more complex and that the politics of these groups, at least in the workplace, involves wider participation and has greater local impact than the current political science literature suggests (also, Estlund 2003; but c.f., Carré and Heckscher 2005; Safford 2005).

This is certainly the case if one compares the role of these new groups in the employment rights regime to unions in the old regime. A union's effectiveness depends on its continuity as an organization and a very strong membership commitment, both of which are increasingly difficult to achieve. Identity groups, on the other hand, do not require members to make any financial commitment or any public declaration of support. The members can move freely among different types of groups—in the workplace, sports clubs, neighborhood groups, social and job networking groups—or even among groups representing different facets of their multiple identities, without jeopardizing the capacity of workplace organizations to reassert themselves in support of new demands in the future. Because the claims they make are largely moral and symbolic, they do not require a large financial commitment on the part of the companies against which they are made and they often appeal to divisions of the company concerned with diversity or public relations in contrast to unions whose demands attract the immediate attention of the production and financial departments. The diffuse and relatively chaotic nature of the organizational structure gives it many of the advantages of the new network structures that have grown up in the economy, and of course it circumvents the problem, which the blurring boundaries of productive organizations and enterprises pose for union organization under the NLRA.

Contrasting Measures of Relative Welfare.  The different structures of social and political mobilization offer dramatically different perspectives on American social policy in the period in which union organization and collective bargaining collapsed. Viewed from the perspective of economic identity, and measured in terms of the distribution of income by income class, there has been a dramatic deterioration of relative welfare, as the conventional view suggests. In the course of the 1980s, the median income in the lower deciles of the income distribution declined, and the upper deciles rose, relative to the median of the income distribution as a whole, to levels comparable to the years preceding the Great Depression and the passage of the New Deal welfare legislation (Levy 1998). But if one asks the question of how the various social identity groups that drove the employment rights regime have fared relative to the previous generation of the same group, there is substantial social progress.

Progress was most striking for women, for the aged, and for ethnic groups in which recent immigrants predominate. Median earnings of full-time, full-year women workers relative to men rose from 61 percent in 1980 to 74 percent in 1994 (Statistical Abstract of the U.S., 1996, p. 469). Real median income of a husband–wife family over age 65 increased by 8 percent between 1970 and 1980 (compared to 1 percent for all households) and a further 9 percent between 1980 and 1990 (compared to 4 percent for all households). Immigrants, who compare themselves to those left behind in the home country, experience an even more dramatic increase in their relative welfare.

For Blacks and for the disabled the picture is more nuanced. Expected gains in income levels and a narrowing of the gap with the rest of the population have not materialized. But employment legislation affecting these groups was part of a larger campaign to end their subordination and stigmatization, not only in the labor market but much more broadly in politics, in education, in access to government programs, and in public accommodation. Especially in public accommodation, the gains have been enormous and highly visible. Jim Crow in the South has been virtually eliminated and Blacks have come up from the back of the bus where their parents were once forced to sit. Public architecture has been entirely redesigned to accommodate the physically handicapped; over 85 percent of workplaces have been reconfigured to open access to the disabled (Acemoglu and Angrist 2001; Kruse and Hale 2003).

Underlying Model of Industrial Society.  A focus on the contrast between the new and old regime, and their different axes of social mobilization, suggests an understanding of a regime change that is very different from one centered on market forces and market-oriented public policies. The economic identities around which the collective bargaining system was built were rooted in a model of industrial society, a model that actually had very little to do with the market. It emerged gradually, beginning in the late 18th century, and was further developed and refined over subsequent decades. That model collapsed in the course of the last three decades. And it is difficult to see how the old regime of industrial relations could have survived the demise of the model upon which it was predicated. The new regime had very little to do directly with the collapse of the old. But it has grown up in a way that is consistent with the emergent model of postindustrial society and the axes of mobilization that have surfaced within it.

At the heart of that old model was the notion of a society composed of distinct realms of activity, each governed by its own standards and separated from each other by sharp boundaries. The most fundamental boundary as noted by Max Weber was the separation of the economy from other realms of social activity. Production and exchange were moved out of the home and into the factory or office, and the household was rendered as a separate space reserved for consumption and for private familial activity. The separation of the two realms involved a distinction between the standards and forms of judgments that governed within them. The standards of the economic realm in particular were “scientific” and “rational”; the standards governing in the household were “affective” and “personal” (Weber 1958, 1968). This “new” industrial model replaced a traditional structure in which production and consumption were intertwined, and production typically occurred within the household, often in the family's living quarters and household and work roles were one and the same. The finances of the household and the business were intertwined.

The characteristic institution of the social realm was the family; in the economic realm, it was the corporate enterprise. Each was taken to be stable, enduring, and well defined. The economic and social realms were not only sharply separated, but each was structured, and the two connected, in particular ways—ways that were reflected in the institutional composition of the era. A single dominant, male agent—the family breadwinner whose earnings are the family's main support—represented the family in the economy. Conflicts between the social and the economic realms could thus be resolved by adjustments in the breadwinner's wage and, since the workplace was also a social setting, the other terms and conditions of his employment. Finally, these breadwinners were organized into, and represented by, trade unions, which thereby came to mediate between the economic and the social structure through collective bargaining with corporate enterprises.

The position of the family in this construction was reinforced by and reflected in the bourgeois family model that emerged in the course the 19th and early 20th centuries. That model became a normative standard toward which Americans aspired and against which they judged themselves as individuals (Dell Brady 1991; Coontz 1992). Public policy was designed around this model at one and the same time, nudging people toward it and helping them to achieve it. The model prescribed distinct roles for women who managed the household realm and preserved the personal and affective values that prevailed within it, and for men who supported the household through outside work and cultivated the strength and competitive spirit required to effectively do so.

The whole of industrial society never, of course, conformed either to the bourgeois family model or to the broader separation of the economic and social realms in which it was embedded. But the model was thought to represent the direction in which the society was evolving, and legislation was structured both to be consistent with the movement in that direction and to facilitate it. Emblematic of the way this occurred was the outlawing of industrial homework in the United States as part of the effort to legislate and then enforce a minimum hourly wage in the 1930s. The movement toward this industrial model, however, came to a halt in the 1970s, and the model, in so far as it was embedded in its key institutions, disintegrated. In the United States, the collapse of the trade union movement in the early 1980s was the most dramatic break in the structure. The other institutions did not dissolve in so sudden and decisive a manner, but in the course of the last several decades, both the family and the enterprise have gradually lost the characteristics that were central to the old model of the relationship between the economic and social realms.

In the case of the family, the critical factor has been the progressive rise in female labor force participation, especially of women with small children. An additional influence has been the growth of income support associated with the welfare state itself. Together, these diversified the sources of family income, weakening dependence upon the male head as breadwinner, and complicating the connection between the economic and social realms. The increasing commitment of women to paid labor and the growth in other sources of income also led ancillary family members—the aged, youth, the disabled—to move out and form their own households (D’Emilio 1998; Costa 1999). The increase in the divorce rate, the rise of children born to single mothers, the growth of separate households headed by the aged, the emergence of gay and lesbian households, all reflected and reinforced these trends. Virtually all of the groups that have displaced trade unions as the locus of social and political mobilization and are not defined by race and ethnicity (i.e., women, the aged, the disabled, gays, and lesbians) are connected with these developments.

The declining integrity of the enterprise as an institution is largely the product of the 1980s and 1990s. As major corporations were threatened with bankruptcy, the traditional boundaries of industries were redrawn as the lines separating different markets shifted, blurred, or disappeared entirely. The most spectacular example is the way the separate industries of communications, office technology, information technology, printing, and photographic imaging have merged (Lester and Piore 2004). Cross-functional and interorganizational teams, just-in-time delivery, strategic alliances, and the like have all led to the interpenetration of once distinct enterprises (Sabel 1991). As the integrity of the corporation has been compromised in this way, the institutions of job security and social welfare that were attached to the enterprise have collapsed and the ability of the enterprise to serve as a locus of economic identity has been progressively compromised.

The organization of work is also shifting in ways that break down traditional occupational and professional distinctions. This shift has been widely noted (Osterman 1996; Capelli 1999; Kalleberg 2003; Marsden 2004). Its precise quantitative significance has not, however, been documented nor are its underlying causes fully understood. It is thus unclear how much the reality has changed. But qualitative studies make clear that the perception of reality certainly has. The new organization is most prominently associated with the project/team organization of work (Grabher 2002; Sydow and Staber 2002). Projects are of limited duration and require workers with a set of shifting and specialized skills that cooperate closely for a limited period of time. Workers are continually redeployed as projects come to an end and new projects are launched. Other key developments are the use of contingent workers (particularly at the higher-skilled end of the market), the increased reliance on strategic partnerships for services, and the increasing complexity of supply chains (Kunda, Barley, and Evans 2002; Bidwell 2005). Together, these developments undermine standard industrial, occupational and professional categories and the institutions built around them, and have led to, among other things, an increase in the need for labor market intermediaries who understand the particular characteristics of individual jobs and workers and can facilitate the matching processes (Houseman, Kalleberg, and Erickcek 2003; Fernandez-Mateo 2004).

A variety of different institutions and organizations have emerged to play the intermediary role, and both employers and workers typically draw on several of these at the same time when they are in the labor market. These range from temporary help services and professional headhunters and recruiters to university alumni associations and informal alumni networks, associations, and networks of former employers of prominent companies such as Citigroup in banking, IBM in information technology, and AT&T in communications. But also among the networks from which employers and workers draw are those organized around sex, race, ethnicity, sexual orientation, and the like—those, in other words, which have become the new locus of social mobilization.

The logic that has grown out of the sum total all of these changes in the composition of families and enterprises is suggested by contrasting the Supreme Court's reasoning in two key labor decisions: the 1960 steelworkers’ trilogy interpreting the National Labor Relations Act and the 2002 Toyota Motors versus Williams interpreting the Americans with Disabilities Act. The steelworkers’ trilogy is the last of a series of cases in which the Court fashioned the law governing arbitration under collective agreements in the new deal regime. “The collective bargaining agreement,” the Court argues, “. . . calls into being a new common law—the common law of a particular industry or of a particular plant [emphasis added].” Only the arbitrator, an insider chosen by the parties, is in a position to interpret that law, and hence the Court binds itself to defer to his opinion.

The Toyota case, decided forty years later, concerns the definition of disability in a case alleging discrimination in employment. It overturns a lower court decision that defines disability narrowly in relationship to the particular tasks of the job in the workplace in question. “This,” the court asserts,

“was an error. . . . The central inquiry must be whether the claimant is unable to perform the variety of tasks central to most people's daily lives, not whether the claimant is unable to perform the tasks associated with her specific jobs. . . . [not] by analyzing the effect of the impairment in the workplace. . . . [H]ousehold chores, bathing, and brushing one's teeth are among the types of manual tasks of central importance to people's daily lives, and should have been part of the assessment of whether respondent was substantially limited in performing manual tasks.”

Thus, the employment rights regime is based on an entirely new conception of how society and the economy relate to each other as the Toyota decision makes clear. The notion of a strict distinction between the two realms—a la Weber—is gone and, in this sense, the new regime is more universal. It incorporates elements of what had been reserved for the home sphere back into the economic. Distinctions that were relegated to the social sphere—race, religion, family role, life stage, etc.—now play an important role in the operation of the economy. Moreover, the breakdown of strict boundaries among industries and among organizations has brought with it a movement away from rationalized notions of economic interaction. In its place, ways of relating—trust, learning, affect—that had been shuttled out of the economy and into the household have come to occupy center stage. The old regime—based as it was on an outdated conceptualization of economy and society—simply could not withstand the change.

The Challenge to Industrial Relations Theory

The field of industrial relations has grappled with the decline of the New Deal era collective bargaining at least since the early 1980s. As Thomas Kochan—one of the field's leading proponents—has put it: “The central challenge facing our generation . . . is to update our ideas, policies, and practices so they can once again achieve their fundamental mission . . . the enhancement of efficiency and equity at work and in our societies” (Kochan 2003). In response, some have called for new models of worker representation (e.g., Kochan, Katz and McKersie 1986; Heckscher 1988; Osterman, Kochan, Locke, and Piore 2001); others suggested possibilities for the revival of collective bargaining as we know it (e.g., Moody 1987; Turner, Katz, and Hurd 2001). In either case, the vision is of a new system that is to come at some future point after the current period of neo-liberal hegemony has dissipated.

Our argument poses a strong challenge to that claim: the transition has already occurred.4 Yet the field has been blind to this development because the emergent system resembles neither the system we have left behind nor any of the reworked versions that have been proposed to replace it. In short, the new system is a total surprise. That surprise is ultimately a product of the conceptual apparatus that structures the field of industrial relations. That apparatus creates the expectations in terms of which we understand and edit reality. In order to recognize what has actually happened and think through its implications, for practice, policy, and social welfare, we need to identify what precisely it is about our conceptual framework that creates these false expectations and make the adjustments required to correct them. It is of course impossible to do this in the short compass of an essay of this kind. But it is possible to identify directions in which we might move.

The Challenge to Existing Industrial Relations Theory.  For these purposes, we turn again to John T. Dunlop's (1958) book, Industrial Relations Systems. Dunlop's framework is actually an extension of Parsonian sociology. In Parson's view, society is organized in a series of relatively autonomous realms of activity. Each of these realms plays a functional part in the operation of the society as a whole. The separate operations are coordinated and made to work in harmony by the way in which the outputs of one realm feed into and constrain the operation of the others. These distinct realms are in turn the domain of study of a different social science. Dunlop's basic claim was that industrial relations is to be understood as one of these realms, on a par with the society, with politics, and above all with the economy. Just as the economic system produces goods and services, the industrial relations system produces the rules of the workplaces.

At first blush, this approach seems well suited to our problem: Dunlop's framework was in fact designed to compare the rules, and the process through which they were produced, across industries and in different countries at different times. Differences are traced to variations in constraints; specifically, the constraints that other realms of the Parsonian social system impose. Presumably, changes in these constraints would also explain why a given system evolves in time. Indeed, the explanations that attribute the collapse of the New Deal collective bargaining regime to deregulation (a change in the political and market constraints), to globalization (a change in the economic constraint), to IT, to transportation, and to communication (changes in the technological constraint) all have this character.

But when confronted with the new system, Dunlop's framework has three principal problems. The first concerns the identity of the actors themselves. Conventional industrial relations theory assumes a set of well-defined categories and strict boundaries defining the actors of the system. But the received categories are compromised by the emergence of a new set of categories within the new system, which operates on a terrain that is no longer delimited along clear lines of organization and industry. The categories around which the theory is structured reflect, in turn, an outdated view of the structure of industrial society and how that structure would evolve. This is a general problem with Parsonian sociology, but it is no less a problem for industrial relations for the fact that it is more widely shared.

The second set of problems concerns the way in which the actors in the system behave. The industrial relations theory assumes that actors act purposively, generally in an economic environment where they are seeking to maximize some monetary value—the firm profits, the unions, the income of their members, or their membership base. But identity groups in the new system seem to be motivated as much by the desire for social recognition as by economic gain, and the pressures they exert seem to be largely moral and symbolic rather than economic. More difficult to understand still, the traditional actors in the system—workers, employers, and the government—respond to these pressures, despite the fact that, unlike unions, the identity groups lack economic power, lack a dues-paying base, do not wield an effective strike threat, and, indeed, do not even have a continuous organizational existence.

Finally, the framework is basically static, or more accurately, comparative static. It is designed to compare regimes once they exist, but it does not examine the process through which they emerge or change over time. It sees each regime as determined by and growing out of a series of constraints imposed by the other regimes (or subsystems) of the larger social system and implicitly assumes that given these constraints, the regime in question is stable and coherent. It does not examine the process through which a system adapts to changes in the constraints it faces, and hence how one system might transform into another, let alone the possibility that more than one regime might be consistent with the same set of constraints or that a regime, given the set of constraints in which it operates, may be neither stable nor coherent.

In large measure, each of these three shortcomings stems from Dunlop's embrace of postwar notions of behavioralism, which viewed institutions as a veil on more fundamental forces. That approach has been superceded by a variety of new institutionalisms in sociology, political science, and economics (Hall and Taylor 1996; North 1990; Thelen 1999). Many of these new institutionalists trace their roots to an earlier tradition in industrial relations associated with John R. Commons and the Wisconsin School of Economics. Commons and his students were concerned with elaborating a dynamic rather than equilibrium-oriented model: one in which actors are, as Hodgson (1994) summarizes it, “mutually entwined in a web of partially durable and self-reinforcing institutions.” Neo-institutionalism—particularly in sociology and political science—grew out of a recognition of Parson's (and therefore by extension Dunlop's) shortcomings and is actively engaged in building new theory to overcome them. It seems natural therefore to look to the modern-day inheritors of his ideas to address the limits of existing industrial relations theory.

Institutional Isomorphism.  In sociology, most directly relevant to the problem at hand is the concept of “institutional isomorphism.” The central idea is that economic behavior is determined, not simply by formal institutions’ or individuals’ rational calculations, but rather by a range of structural, cognitive, and behavioral factors that generate conformity (or at least similarity) among actors’ behaviors (DiMaggio and Powell 1991). Particularly in the absence of established rules and norms of behavior, practices diffuse via processes of imitation, competition, coercion, or professional norms rather than through a deliberate, rational, or strategic decision-making process, as envisaged in economics and management decision theory.5 The concept is not tightly linked to an alternative theory of decision making. But it is easy to see why it might be more likely to govern the diffusion of practice if they develop and spread in a sequence of small acts and local decisions.

That a neo-institutionalist approach captures a key component of the process through which the employment rights regime emerged is nothing new. The HRM component is the focus of a series of articles by Frank Dobbin and his coauthors (Dobbin, Sutton, Meyer, and Richard 1993; Dobbin and Sutton 1998; Kelly and Dobbin 1999) that have served as a vehicle for the development of the concept in the first place. But while institutional isomorphism might help us to understand the processes by which the new regime has taken shape, what it cannot tell us is why the old regime fell off in the first place.

For instance, one major focus of contemporary industrial relations has been on high-performance work systems that respond to the limits of the set of work rules that developed under the old regime in the new environment. But despite proven efficacy (MacDuffie 1995; Ichniowski, Kochan, Levine, Olson, and Strauss 1997; Ichniowski, Shaw, and Prennushi 1997), they have failed to diffuse; institutional isomorphism has been blocked. Multiple obstacles stand in the way. Inertia and the well-known forces behind Michels’ ([1915] 1962) iron law of oligarchy have certainly played their role (Voss and Sherman 2000). But so too have a set of deliberate, calculated decisions made by the principle actors in the light of economic interests; a fact that much current thinking in economic sociology—which takes an essentially passive view toward human action—has more difficulty addressing.

A salient example is Saturn (Rubinstein, Bennett, and Kochan 1993; Rubinstein and Kochan 2001). It is a whole new company created by General Motors in the 1980s. The organization of work and the collective bargaining contract that governs the plant were a deliberate attempt on the part of the company and its major union, the United Automobile Workers, to meet the challenge of the new business environment and the competitive pressures, which at that time were emanating from Japan. They constituted a totally new approach to labor–management relations, as distinctive from the regime growing out of the New Deal as the employment rights regime. That approach, moreover, was judged by indices of business performance a success and the Saturn Division continues in operation today. But the diffusion of the Saturn model has been forestalled. Automotive companies, which copy each other's practices in an isomorphic fashion along other dimensions, have not copied Saturn.

Saturn is perhaps a unique episode in the evolution and decline of the old regime in the degree to which it was the subject of collaboration and negotiation between labor and management. With relatively few exceptions, the other major episodes involve unilateral efforts of management to block union organization and limit collective bargaining. But those efforts and the union response have been equally the product of self-conscious and deliberate business decisions. On the part of business, these campaigns have been carefully crafted by outside consultants specializing in what is euphemistically called “union avoidance.” An important component of the union response has been “corporate campaigns,” designed by their own consultants, with equal deliberation, albeit with considerably less success. Institutional isomorphism might account at most for the legitimacy of these new tactics, which were almost unthinkable in the earlier post-war decades by the ethos of the old regime. Nevertheless, as it stands, it is a peripheral part of the basic story because it lacks a vocabulary for explaining how interested parties actively shape—or in this case, block—the diffusion of practices over the course of their institutionalization. If economic sociology offers an important lesson to industrial relations in the form of institutional isomorphism, industrial relations scholarship has its own lessons to teach about the limits of the new theory.

It may be possible, however, to merge the lessons of the two theories into a coherent explanation of regime change. Perhaps decisions about older institutions, precisely because they are so highly articulated and the regimes in which they are imbedded are so well understood, are made in a deliberate, calculated way. But the decisions associated with the emergence of the new regime pass under the radar screen of the powerful actors, are not subject to calculation, and develop mainly through the relatively passive processes of institutional isomorphism. The decisions about the old regime were made by the CEO, while the decisions affecting the new were made by the legal department, by lower levels of HRM, by “diversity” officers, and by people allocating the companies’ charitable contributions. This suggests that the old regime and the new develop in very different ways, through processes that are not directly related, perhaps not related at all. But combining industrial relations and institutional isomorphism is at best an incomplete way of addressing the issues confronting the field. It does not explain where the institutions that diffuse (or fail to diffuse) come from in the first place. And it does not address the question of identities, which seems so central to the transition between regimes, at least as we have characterized it.

Identity, Interest, and Institutional Change.  Identity does not achieve explicit recognition in either the old collective bargaining regime or in the conceptual apparatus that industrial relations developed to analyze it. But neither is it completely lacking. Under the National Labor Relations Act, the bargaining unit was defined by establishing boundaries around a “community of interest” that could well have been termed a common identity (although the term itself is completely anachronistic to the language of the era). But since motivation and behavior in conventional industrial relations is so closely tied to economic interest, and the power to impose material loss upon one's adversaries, any notion of identity is always attached to the individual's relationship to the means of production. The concept does not extend to workers who are organized around their place in the larger society or to claims of a moral and symbolic nature rather than material.

It has always been the case, of course, that there was some correlation between social identity and economic standing. Black workers were deliberately assigned to lower status jobs on the plant floor, skilled Northern European workers assumed and protect their elevated place in the production hierarchy, and construction unions were famously organized along ethnic lines within various geographic jurisdictions. But today, the connections directly linking economic role and social identity are blurred to the point of being almost unrecognizable. Where they persist (and of course they do persist), the distinctions become something to be eradicated and therefore an issue around which social identity groups in question focus their organizing energy.

Thus, to the extent that conventional industrial relations theory recognizes identity (or can be stretched to recognize identity), it starts from the point at which identity is already incorporated within the formal rule-making mechanisms. It skips over the prior processes in which the group first comes to recognize itself as a group with a set of claims to make within the context of the workplace, and then society—and most importantly employers—come to recognize and respond to the group's legitimacy (Kelly 1998).

Sabel (1980) offers one way of thinking about identity formation; one in which collective identities emerge from particular struggles at specific points in history. The classic story is that of a group of Southern Italians who migrate north to find work. Their own self-conception—one shared by their northern counterparts—was that of being peasants temporarily interloping on the prosperity of the industrializing north to accumulate funds that they would eventually take back with them and invest to enhance their status within the peasant community. They stayed so long in the northern industry that they effectively become industrial workers. Nevertheless, they continue to understand themselves as peasants. It was only after an incident provoked a strike in which both southern and northern workers joined forces that the southerners began to conceive of themselves—and are in turn recognized by others—as workers per se. Only once they have achieved entry to the identity is it possible to join forces to assert their mutual interests.

That approach is in some sense a complement to Commons’ view that institutions result from the resolution of specific—local—conflicts that are then aggregated and formalized over time. But while Commons had no theory to explain why particular groups of workers emerge to proclaim themselves a community with a common set of interests, Sabel's conceptualization suggests that this kind of identification materializes out of the amalgamation and transformation of existing identities. For instance, it was not until the Congress of Industrial Organizations movement of the 1930s that the concept of “blue collar” entered into the American lexicon. Previously, workers thought of themselves in separate terms—skilled and manual—not as a single group with linked interests. It was the industrial labor movement's struggle, subsequently strengthened by the Second World War—that workers identified with the unifying concept of being “blue collar” per se and it was around this particular configuration of working class identity that the National Labor Relations Act and its follow-on legislation were eventually institutionalized.

Another approach to identity formation is suggested by a body of work associated with the Chicago School of Sociology (Barley 1989; Evans 2000). Here, identity emerges not from particular struggles but from the process in which workers construct the stories of their everyday lives. These stories are contextualized by the ways in which workers interact with social institutions. The key concept here is career, but the term is understood not simply as a sequence of jobs but rather as any sequence of life events—work-related or otherwise—which can be strung together into a coherent plot. Identity can be thought of as the character within that storyline with which the individual is associated. But different stories can be spun out of the same events; the salience of a given identity depends on which story is spun. There is a strong parallel here with the notion of radical technological discontinuities in which established businesses miss the emergence of new technologies because their attention and analytical capabilities are focused elsewhere. Disruptive technologies are not disruptive because they are inherently superior to existing technologies—in fact generally, at least initially, they are not. Rather, they are disruptive because they are embedded within a different way of organizing the technology and the identities associated with it (Christensen 1997). Disruptive technologies, in other words, are associated with alternative storylines and alternative identities.

Under the New Deal era of collective bargaining, the broader storyline—the one with which workers ultimately identified—presumed the Weberian distinction between the rational economic realm and the irrational social realm. This gave rise to the familiar definition of career as the sequential series of economic roles a worker was likely to play out on his or her way toward retirement, devoid, in many ways, of any other social context. Today, though, the economic storyline is far more complicated, and certainly less predictable. It is impossible in today's world to imagine one's work career without incorporating one's social context into it: that is, such aspects of life as parenthood, health, and the social stigma that may attach to one's race, religion, or gender (Higgins 2001; Briscoe 2004). Social identity, in this context, serves as a ready alternative to a work-centered plot and, therefore, to work identities as an axis of mobilization (Arthur, Khapova, and Wilderom 2005).

Why one plot emerges and another takes its place is less clear. In the case of the shift from collective bargaining to employment rights, it seems to have involved a combination of purposive strategic action—on the part of some particular civil rights leaders at a critical moment in history—and a series of unintended consequences that eventually coalesced into the organizing logic of the new regime. The one thing that is clear is that the process of identity formation and change is ongoing; it does not stop once the identity has been incorporated into a given set of institutions.

This raises a separate conceptual issue: What is meant by a regime in industrial relations; how would we know that a regime shift has actually occurred? The identities on which a given regime is based are constantly evolving, but the formal institutions remain. Thus, new institutions and identities are difficult to quantify in part because our statistical categories and data collection were created to analyze the operation of the old regime, not the new one. As a result, we cannot provide a definitive answer to the regime question. Yet we think it is useful to distinguish between and contrast two approaches.

The conventional approach, suggested by the title of Dunlop's book, is systems theory according to which institutions and actors are not to be understood in isolation; rather, they exist relative to each other, and hence are to be found in constellations or systems. The presence of one institution facilitates—could even compel—the emergence of another institution. The collapse of a key element of an institution (the unions, for example, or the family) might bring down the whole system. The result is the production of many of the effects now stressed in contemporary institutional theories: lock-in, path dependence, punctuated equilibriums. Once a new regime gets established, the actors and institutions associated with the old regime will have trouble surviving because their complements are increasingly absent.

An alternative, which seems more consistent with the way in which employment rights has overtaken collective bargaining, is that industrial relations is constructed out of a vision of industrial society and how it is evolving in history. That vision may be expressed as a model or as a system but it is actually a story or a narrative (similar to the way a worker's identity is linked to the story they tell about their lives). The story recognizes not only institutions and actors associated with the emergent regime but also those left over from some earlier moment of history. When the story is translated into a conceptual apparatus or the categories become the basis of public policy, only the forms of the emergent regime count; the earlier forms go unrecognized or are actually suppressed.

This notion of historical narrative is suggestive of a view of human behavior very different from the deliberative, rational models underlying conventional industrial relations. In this alternative view, historical explanation then comes to focus on the question of why some stories or narratives emerge and gain hold on our imagination. The answers that emerge in the literature, which develop this view suggest that it has less to do with the measured significance of particular institutions and events than with the ability of certain stories to make sense out of individual lives and to link those lives to the groups and institutions around which the larger social narrative revolves. It is the themes in these stories about where industrial society is going—rather than the scholarly texts—which underlie our own convictions about the significance of contemporary developments.


This paper develops a contentious argument. It is widely accepted that the New Deal system of collective bargaining has collapsed. But we go considerably beyond the accepted wisdom when we argue that a new system has already emerged, reflecting a shift in the fundamental structures of industrial society and a change in the locus of social mobilization. To argue that the difficulty that the field has had recognizing this shift and coming to terms with it reflects the limits its analytical apparatus compounds that challenge to the point of questioning the field's continuing existence as a scholarly discipline. But industrial relations is not simply a field of scholarly endeavor. It is also a field committed to explicitly informing a community of practice. This has linked it to certain institutions and individuals whose place in the world is threatened by the changes we have been discussing. Those linkages are an additional factor, making it difficult for us to recognize and accept changes on the ground, let alone to think about them dispassionately.

To argue in this way is not to say that the changes we discuss are desirable. Indeed, nothing in our argument suggests that we should welcome the decline of the old system nor celebrate the emergence of the new. Still less do we wish to criticize either the institutional commitments that have made it difficult to recognize and accept the changes or the intellectual commitments that made it hard to understand them. The latter constitutes the scholarly heritage of the discipline and the foundations upon which a new and more powerful conceptual apparatus must be built.

The commitment to building a discipline that speaks not only to scholars but to practice and to practitioners is what distinguishes industrial relations most from other scholarly endeavors and is its claim to our allegiance. That commitment must be expressed through specific actors and institutions. But ultimately it is not any one set of commitments but a much broader and more general commitment that defines the field; a commitment to those actors and institutions that struggle to find a voice and provide a vehicle for the less powerful, the oppressed, the under-represented, and the socially stigmatized in industrial society. These people once found unique expression through trade unions organized around their economic identities and positions. But they have now come to understand themselves in different ways and seek expression through new institutions. The very concerns underlying the field's commitment to trade unions and collective bargaining now compel us to recognize those new forms and develop the analytical tools required to understand, evaluate, and help them find effective expression in practice and in action.


  • *

    The authors’ affiliations are, respectively, Department of Economics, MIT; Graduate School of Business, University of Chicago. E-mail: We have accumulated debts to a number of people, too great to be acknowledged here. But we were particularly influenced by conversations with the students and faculty of the IWER group at MIT, and with Frank Dobbin, Deborah Kolb, Victoria Hattam, Natasha Iskander, Chris Desan, Sarah Ashwin, Carola Frege, John Logan, Bob McKersie, and Andrew Schrank.

  • 1

    Circuit City Stores, Inc. v. Adams (99-1379) 532 U.S. 105 (2001) 194 F.3d 1070; see also Gilmer v. Interstate/Johnson Lane Corp., 500 U.S. 20, 111 S. Ct. 1647, 1655 (1991).

  • 2

    United Steelworkers v. American Manufacturing Co. (1960); United Steelworkers v. Warrior & Gulf Navigation Co. (1960); United Steelworkers v. Enterprise Wheel & Car Corp. (1960). See Stone (2005).

  • 3


  • 4

    The elements of the new system we describe here have been in place for some time: A set of employment laws supplemented by the formalization of internal human resource practices culminating in the arrival of a new system of binding arbitration with worker voice provided for through organization at the national and company level organized along lines of social identities rather than economic ones. The various pieces of this system have been discussed elsewhere [on mandatory arbitration, see Stone (1996, 2005) and Estlund (2003); on the rise of HRM, see Sutton, Dobbin, Meyer, and Scott (1994); on the role of identity-based affinity groups, see Friedman (1996), Scully and Segal (2002), and Carré and Heckscher (2005)]. But while the elements are known, their relationship and coherence as a new American industrial relations system has not yet been fully articulated.

  • 5

    This idea is prefigured in some of the lower-level theoretical constructs of industrial relations, ironically, even constructs developed by Dunlop himself (e.g., wage contours and pattern bargaining). But this earlier work generally lacked the theoretical clarity or focus on mechanisms that is evident in modern economic sociology's concept of institutional isomorphism.