PART ONE

Chapter One
Culture and the Emergence of Small-Group Activities

Over the past twenty-five years in the advanced market economies, and even among some of the command economies, a movement has grown for building small-group activities among lower-ranking production and office employees. The term small-group activities refers to workshop- or office-based groups that are given a greater opportunity to exercise direct control over everyday work decisions and the solving of workshop problems. In the United States the rise of quality circles (employee-involvement groups) and, of late, self-managing teams has come to symbolize this movement, which has been associated in varying degrees and ways with a broader effort to expand employee participation in managerial decision making. In many countries this movement arose as a response to a crisis in management confidence that in turn derived from the need to respond to a threat to or loss of a competitive ability. The idea was to find a more effective way to recruit and make better use of employees to achieve organizational goals while satisfying the needs of individual employees for control over their immediate work environment. There were significant national variations in this regard, and indeed such variations are central to this book's strategy of analysis.

In some nations powerful currents have sustained spring tide conditions. In others the ebb tide has come quickly, and only scattered debris is left to show small-group activities having been present. In still others a riptide churns the currents in opposite direc-

tions. These disturbed currents obscure underlying movements, which only the passage of time will clarify.

The task I set myself was to analyze and understand these different currents. I wanted to understand how these activities spread—diffused—in given nations. Why had they spread more rapidly in some countries than in others? What were the channels of diffusion? Through what mechanisms and with what effect? Analysis of these movements is more complex than might initially appear to be the case; a variety of environmental factors affect the course of development of small-group activities. Moreover, management commitment to these small-group activities often displays a faddish quality.

What Is So Important about Small-Group Activities?

There are compelling practical reasons for focusing on small-group activities. The failure of American business to fully use its human assets inhibits growth in quality and productivity, thereby undermining national living standards. There is good reason to believe that some competitor nations make better use of their human assets, thereby developing a competitive edge in the fierce struggle for international markets. Small-group activities allow individual workers to share their knowledge and skills and to develop them in ways that enhance economic success. This helps not only to assure the firm's success but to ensure employment security as well.

A more recent line of reasoning stresses the impact of the new technologies of flexible manufacturing and microelectronics, combined with growing worldwide competition in price and quality. In practice this means heightened pressures to produce more custom-made high-quality products and services to meet more volatile and sophisticated customer tastes. These developments are said to lead to further payoffs for increased teamwork, employee discretion, and participation to get the most out of the technology and to better respond to customer demands.

Some advocates take a different tack, starting from the proposition that small-group activities give individuals significant participation in organizational decision making. This in turn is said to be a strong democratizing force that finally brings the benefits of politi-

cal democracy into the workplace. Notions of self-governance and self-determination underlie this perspective (see Dachler and Wilpert 1978:5). This claim contrasts with the neo-Marxist view that small-group activities are just one more management technique to control workers. The neo-Marxists see these evolving practices as part of a new corporatist strategy that, by fostering a sense of participation in the organization on the part of employees, destroys collective solidarity among workers and makes them vulnerable to a higher level of management control (Edwards 1979; Burawoy 1983).

Still another view stressing the participative aspects of small-group activities emanates from the moral arena. Secular humanitarian and religious leaders focus on the significance of participation for increasing the control of individual employees over their own lives and providing an outlet for their creative talents, thereby elevating individual dignity and self-respect. The draft of the bishops' pastoral letter "Catholic Social Teaching and the U.S. Economy" (National Catholic News Service 1984), for example, states that economic activity should be an expression of the distinctive dignity of human beings and that work should enable people to find a significant measure of self-realization and creativity in their labor.

Theories of human growth and development, with their stress on self-actualization, are a related perspective deriving from social science. Theorists working in this tradition talk about the "humanization of work" and tend to see participation as a key strategy for improving employee motivation and allowing individuals thereby to achieve their true potential. The focus tends to be on the individual's relation to the work task itself rather than on questioning the broad distribution of power and decisions.

There is now significant evidence to show that Japanese firms in particular have made better use of their human resources than have the Americans; this has significant consequences for living standards and the quality of life at work. Furthermore, by effectively conducting small-group activities, Japanese firms achieve a variety of economic benefits. These involve flexibility, speed, economizing on resources, and the execution of decisions based on accurate information (see Aoki 1987). Just as it is naive to think that small-group activity inevitably leads to self-actualization and democratization, it is unduly cynical to think that it is always a corporatist strategy designed to destroy the collective solidarity of workers.

What consequences small-group activities have is above all an empirical question. In focusing on the spread of small-group activities across firms and the role of national infrastructures in promoting this, I concentrate more on management than on workers' behavior. The preceding discussion is designed simply to show the reader that the content of what is being diffused is important.

Why a Comparative Study of Small-Group Activities?

What is the contribution of a comparison of different national experiences with small-group activities? First, it is clear that there are distinctive national approaches to small-group activities. The American model stresses direct, informal participation and provides relatively limited scope for those at the bottom of the organization to participate in decision making. This is strikingly different from the more formal representational approach to participation taken in continental European nations (see Dachler and Wilpert 1978:23-25). To be sure, collective bargaining arrangements in unionized U.S. firms should rightly be seen as providing representative participation, but relatively few of these arrangements have been used explicitly to pursue individual participation in work decisions per se.

From a social science perspective, a comparison of different national experiences permits one to begin identifying the characteristic elements of successful diffusion processes. It also helps us identify those elements of the diffusion process that are unique to particular nations; such unique elements are hidden when one studies one nation's experiences. The role of culture provides a dramatic example. It is impossible to estimate its effects when examining only one culture, but in comparison with others, we have the opportunity to clarify its significance. Given the attention paid to cultural effects in the analysis of small-group activities, this leverage appears to be of considerable benefit in furthering more robust explanations. By our knowledge of unique national characteristics, we in turn enhance our understanding of contrasting social structures. As Reinhard Bendix (1956:445) has argued, if different societies over time confront and resolve a similar problem, then a comparative analysis of their divergent solutions will help us understand the divergent character of the respective social structures in a process of change.

Finally, from a policy perspective, it is instructive to see how other nations "do it," so that we may learn from their experience. There is ample evidence of cross-national flows of information about small-group activities. I shall examine how social actors in different nations have perceived and responded to such information flows. There are a great many misunderstandings, colored by national ideology and cultural traditions, about how the borrowing process operates.

At a minimum, the very fact that small-group activities in one form or another have landed on the agendas of all the market economies over the past twenty-five years is remarkable. How did this happen and why? How is it that in 1984 a draft OECD report on the worldwide automobile industry could conclude:

Although there seems to be a diversity of managerial technology development and practice in the automobile industry in various countries, the general trend is towards increased worker participation in production decision-making. The quality circle approach, involving more off-line input and cooperation between workers and management, is considered important and is being experimented with and implemented by practically all major automobile companies around the world.... There is also a trend towards giving more on-line responsibilities to the workers, ranging from preventive maintenance to quality control, and to multi-process handling…. The strongest general trend is the introduction of group work …. There is a trend away from "man-machine relations" towards "team technological system relations." This involves restructuring the production process by setting teams which often number from three to eight people. It represents a move away from assembly-line structures with their isolating and alienating effects.
(OECD 1984a: 37, 39)

The common trends described in the OECD automotive study are really quite extraordinary. In varying degrees, they have been reported for other industries as well. Are there common driving forces in each country, or is it simply a matter of "international faddism"?

A second question recognizes some of the variation internationally and leads us to ask why small-group activities ended up on the agenda earlier in one place and later in another. A focus on national differences also leads us to ask why the nature of the response has varied from one place to another. This kind of specific research is far more likely to yield useful information to makers of public policy and scholars than the recent vogue for writing vague popular

treatises on how one nation's management system is generally superior to another's.

My focus is on the evolution of small-group activities in the United States, Japan, and Sweden between the early 1960s and the mid 1980s. The stress is principally on why developments have taken the direction they have, but I also explore what paths have not been taken and why. The comparative strategy has been developed for a variety of personal and intellectual reasons. I have long believed that explicit national comparisons are to be preferred to specific national studies of work organization (which often contain implicit and unverified comparisons). As a long-term student of the Japanese situation, I noted that even when comparative studies were carried out by American scholars, they commonly compared Japan to the United States. An ethnocentric bias commonly led to the implicit judgment when explaining observed differences that we were "normal" and they were "different." Japan's recent economic ascendancy has led to some reversal of this assessment, but bias is no less apparent.

It was these kinds of reflections that led me to conclude that a well-grounded study of small-group activities that included a European country would be of considerable value in casting new light on work organization and the various options open to national actors. In fact, I was raising the N from two to three. While some would emphasize that this is still a very small number of cases, I sensed that with proper selection for maximum variation, one could get a lot of leverage from that additional case. Sweden has a strong reputation as an innovator in work organization and its selection as a third country to be compared seemed ideally suited to my aims.^[1]

[1] By limiting the analysis to the United States, Japan, and Sweden, I ensured linguistic competence in the three countries to be compared. This made possible field and archival research based on primary source materials, as well as interviews with key informants in their native language, and facilitated survey research on "early adopters" of quality circles in the United States and Japan.

In the case of Sweden, in addition to the voluminous scholarly and management literature on the subject in Swedish, the newspaper coverage was also extensive. This allowed a different vantage point on developments. These data sources were supplemented by interviews with key informants and recognized experts in the unions, companies, universities, the employers' confederation, and research institutes. Volvo and Saab, two major innovators, were particularly receptive to my initiatives.

For the United States a large scholarly and management literature on direct participation on the shop and office floor level exists. In addition, the author, through lecturing, consulting, and participating in professional associations, has been an active observer of, and sometimes participant in, the process of diffusing quality circles.
In the case of Japan, a significant scholarly and management literature exists on the subject in Japanese. Selected company visits and interviews supplemented this literature, as did interviews with employers' confederation officials and union leaders. The Japanese Union of Scientists and Engineers (JUSE), a key actor in the diffusion of quality circles in Japan, was particularly forthcoming in its cooperation and arranged for a variety of meetings with key informants.

Some Introductory Thoughts on the Role of Culture

In speaking of small-group activities, the question inevitably arises of whether one nation can learn from another's experiences. This in turn raises the issue of culture as a potential barrier to the learning process. Powerful cultural stereotypes in each country bias how one assesses the impact of culture on the cross-national borrowing process. These stereotypes, themselves part of powerful societal ideologies, constitute an important set of resources and constraints on the learning process. As ideologies, moreover, they also distort our understanding of actual social behavior and thereby conceal important resources that are available to the learning process.

Americans and Japanese have an image of Japan as a group-based culture and of the United States as one based on individualism. Many go on to conclude that a group-based activity such as quality circles will not work in the individualistic American culture. Such stereotypes clearly have a basis in fact; Americans do have a strong tendency to assign tasks to the individual and to hold the individual accountable, while the Japanese are more inclined to assign tasks to the group and make the group accountable for performance.

These differences do have their roots in the respective national histories of the two countries, and thus constitute different cultural constraints. In the case of Japan, for example, the legal system during the Tokugawa period (1603-1867), which was based on the principle of collective responsibility, allocated complete authority and responsibility for the performance and conduct of group members to group representatives such as the village headman (see Smith 1983:38). In practice, the system operated over time to assign a high degree of autonomy to village communities and urban neighborhoods alike.

Nevertheless, these stereotypes and generalizations can be terribly misleading when it comes to assessing the practices of individual companies, whether in Japan or the United States.^[2] First, there is ample evidence from social science that under certain conditions, group activities and decision making can be stultifying. Consider the observations of Karl Weick (1984:22), a noted student of organizations:

Take the common demonstration that groups, in the interest of coordination, often accommodate to the least accomplished member. When this happens, members set in motion a mechanism that winnows some intelligence out of the system. In fact, most group mechanisms that have been uncovered (e.g., groups promote polarized beliefs; groups are solution centered rather than problem centered) imply that groups and organizations seem to be exquisitely organized to vulgarize their own minds.

Seen in this light, the question is not whether groups are superior to individuals or vice versa, but under what conditions and in what environments groups perform effectively.

Moreover, in the extent and nature of group activities, there is tremendous variation between firms in Japan and in the United States. Nor are group activities foreign to many American firms. In some cases, the nature of the technology or the service to be provided almost dictates group organization. Yet the teamwork imposed by management policy or dictated by technology must be distinguished from the attitudes internalized by employees. In their recent large-scale survey of workers' attitudes in Japan and the United States, Arne Kalleberg and James Lincoln (1989) report that the Japanese employees are only slightly more inclined to "favor working in groups" than their American counterparts. While the data lend themselves to a variety of possible interpretations (perhaps the Japanese respond the way they do because they have "had it" with group activity), they do suggest caution in embracing popular stereotypes too closely.

To look at the matter differently, consider the practice of individual reward and recognition. One can make the strong argument that in U.S. manufacturing firms the individual blue-collar worker is typically treated in the mass (that is, as a member of a large group) in the sense that he or she receives little recognition in the wage sys-

[2] A biting critique of the groupism model as applied to Japan appears in Mouer and Sugimoto 1986.

tem or informally as to individual performance. The auto worker who told me during an interview, "Nobody gives a damn whether I do my job well or not!" captured this situation all too well.

Conversely, in Japan, there is tremendous effort made in personnel policy to individualize treatment of blue-collar workers and tailor training and wage increments to individual performance.^[3] Now in which economy is the individual "king" and in which does the group dominate? One of the major implications of this focus on the individual in the Japanese reward system for lower-ranking employees is that it leads to tremendous competition among individuals. This means that the highly publicized teamwork among Japanese workers gets played out in the context of a delicate balance between cooperation and individual competition. Such subtle relationships are notably absent from typical Western treatments of the subject, with their overwhelming focus on groupism. Yet they have been a major component in the engine driving continuous improvement activities in postwar Japanese firms.

Similarly, there is the Western image of the Japanese individual subjugating himself or herself completely to group-defined objectives. Such a view does not allow for the real life subtleties whereby individuals seek to control the definition of group objectives in ways that serve their individual interests or values. Nor does it allow us to grasp the changing situation reported in Japan whereby the individual increasingly appreciates the group to the extent that it responds to individual needs (Cummings 1980: 197).

Much of our image of the Japanese group model is an account of how human relationships ought to be. That is, it is ideology. The reality is that groups are often subject to competition and conflict, with even the appearance of harmony being quite deceptive. Moreover, Americans are misled by Japanese stereotypes of themselves as group-oriented and harmonious, just as we are misled by stereotypes of ourselves as exclusively individualistically oriented.

Americans have, in fact, conflicting cultural stereotypes. There exists a communitarian ethic in the United States that has powerful roots growing out of our traditional agricultural society. While we highlight the role of individualism, we also have a set of stereotypes, commonly emanating from the sports metaphor, that point to the powerful benefits to be achieved from teamwork. Only with

[3] These differences do not need to be large to be significant in the eyes of the employee.

the close coordination of individual assignments are the individual talents on a football team channeled to bring about maximum group performance. These analogies are in common use in our business firms, and motivational speakers from the sports world who invoke such comparisons, like Fran Tarkenton, are in great demand for speaking engagements.

Consider also the contradiction involved when observers applaud teamwork but criticize meetings as wasteful! Again, the issue is what kinds of teamwork or meetings, under what kinds of conditions, produce what kinds of outcomes? To argue that group-based activity such as quality-control circles will not work in American firms because they violate our individualistic ethic hardly reflects the reality of competing values and circumstances.

Some Japanese scholars go so far as to argue that Japanese firms succeeded with small-group problem-solving activities in spite of not because of culture. This turns the conventional wisdom of Westerners, not to speak of Japanese traditionalists, on its head. Fujita Yoshitaka argues that only through a strong commitment and strategic and long-term efforts were Japanese managers able to overcome cultural constraints (Fujita 1982:77-86). He notes that traditionally Japanese workers were accustomed to a "do as you are told" style of work in which they had little incentive to think creatively. He sees such practices as deriving from the permanent employment and seniority wage systems, practices often associated with Japanese culture.

Similarly, Shimada Haruo argues that the cooperative labor-management relations upon which small-group activity, information sharing, and other contemporary worker-manager relationships are based, were not bestowed upon Japanese corporations from the beginning. "Rather they were constructed deliberately and at considerable cost." By cost, he refers to the difficulty of coming to terms with a hostile and aggressive labor movement in the early postwar period (Shimada 1982:246). Support for this interpretation can be found in the detailed accounts of individuals who established postwar Japanese employment practices. The Japanese Institute of Labour published a set of interviews with key figures of the postwar period in the personnel departments of major companies such as Nippon Steel and Toyota Motors. What comes through strongly in these presentations is the sense of struggle and personal hardship (expressed in rather typical Japanese melodramatic fash-

ion) experienced by those trying to create new personnel practices during a period of tumultuous upheavals (see, for example, Tanaka 1982). These perspectives provide a healthy corrective to the conventional Western view that simply sees quality-control circles as coming naturally to the Japanese by virtue of their group-oriented culture—a gross distortion of what actually happened and the amount of "sweat" involved in making it happen.

That is not to say that culture does not bestow some advantages—to be read as providing resources—on Japanese managers seeking to introduce small problem-solving groups. It is apparent as one examines the training packages of the Japanese Union of Scientists and Engineers and the typical packages produced by American organizations that the Americans put far more stress on group dynamics. There is an implicit assumption that Americans need more help in navigating these problems. Our supervisors seem less well trained in such techniques as individual motivation through group activity and group problem solving than are the Japanese. Small-group organization is pervasive throughout Japanese society in a way that suggests it provides a basic "code" for how contemporary Japanese think about organizing to solve problems. For example, the group organization of normal learning activities and student responsibility for classroom maintenance tasks documented in Japanese schools stands in sharp contrast to practices in American schools (Cummings 1980). It is reasonable to believe that such experiences during one's formative years provide a strong foundation for subsequent group-oriented activities.

Yet even in Japan tremendous investments were made in training and organizing to get these groups operating effectively in a new problem-solving mode (see Lillrank 1983). After all, you can have close teamwork that is not particularly task-oriented or, worse yet from management's perspective, close teamwork that operates in a way that works against formal organizational objectives. There is, in fact, a large social science literature in the United States on the way in which informal organization at work functions in exactly that fashion.

Basic Arguments

The central questions raised in this book rest on the observation that, in the main period under investigation, 1960-85, American

firms moved more slowly and less effectively in adopting small-group activities than Swedish and, especially, Japanese ones. Why was this the case? Generally speaking, the answer is to be found in a confluence of factors, ranging from different labor-market circumstances, labor-management traditions, and the predispositions of major institutional actors (especially the degree of top management consensus) to the willingness of these institutional actors to take joint action to build national infrastructures for diffusing small-group activities.

Through various comparisons I elaborate below on the relevance of these various factors and the linkages among them. Moreover, I also show how the small-group-activities movement took different forms in each country, reflecting the labor-management balance of power, cultural traditions, and, above all, the degree to which national infrastructures developed and operated to diffuse innovations effectively. I focus particularly on the evolution of these national infrastructures, for they tell us much about the commitment of various national interest groups.

The reader will note in examining the above list of causal factors that I take a decidedly "macro" approach in trying to understand what happened, or did not happen, to the small-group-activities movement in each country. It is only in this fashion that one is able to distinguish the forest from the trees. Much contemporary theorizing and empirical research on small-group activities involves researchers, especially industrial or organizational psychologists, reporting factors within organizations that inhibit or enhance the probability of success (Dachler and Wilpert 1978:29-35). Success is typically measured in terms of survival of the innovation or contributions to improved quality and productivity, increased participation, job satisfaction, or reduced employee turnover and absenteeism. The critical factors typically identified as responsible for these outcomes include the mode of innovation, the amount of authority given to group sponsors and leaders, the makeup of the group, its leadership, the group's procedures and objectives, the nature of the organization's technology, how the small-group activities fit in with wider organizational practices, and so forth.

A pattern emerges (also seen in other domains of organizational research) of a variety of often contradictory studies showing how under different conditions one or another micro variable contrib-

utes marginally to affecting outcomes. Charles Perrow (1979:96-112) documents just such a pattern in his scathing critique of the accumulated human-relations literature, showing how it fails to demonstrate a clear link between leadership and morale on the one hand and productivity on the other. The application of increasingly complex research methodologies and causal models has only resulted in a loss of applicability and theoretical power. The variables become so numerous that one can hardly generalize to organizations or even to types of organizations.

By shifting from the micro analysis of the social phenomenon under examination to a macro analysis, the opportunity develops for grasping the broader environmental factors behind innovation. As Perrow (1979: 110) puts it, "It may be, hopefully, that any theory that has the power to explain a good deal of organizational behavior will have to deal with more general variables than leadership and small-group behavior." In emphasizing such macro explanations as I do, I further assume that the micro variables associated with small-group behavior are randomly distributed and thus have little effect when a large number of organizations are being studied.

In keeping with this vision, I show the importance of understanding the position of the national business leadership toward such innovations. Also to be considered is how these positions interact with national labor-market conditions as well as other variables. In summary, the overall perspective adopted in this book is that this macro approach will produce a far better and more generalizable understanding of the spread of small-group activities across and within firms than a detailed analysis of what makes small-group activities work or not work in particular organizational settings.

This is not to say that I ignore all the micro factors on which many of these outcomes rest. The decision by individual firms to invest in building a national infrastructure rests on the trade-offs these firms perceive, and they will be the subject of further discussion.

Chapter Two
What Is Small-Group Activity?

To say that one is going to study the spread of small-group activities in three countries requires that one identify the phenomenon being spread. The careful reader will have noted some ambiguity implied by the different terms small-group activities and participatory work structures . Clarification of the relationship between the two terms is central to any serious understanding of the subject.

There are many types of what Europeans, Americans, and Japanese alike usually call "worker participation" (Dachler and Wilpert 1978). It can take the form of representative democracy, with workers' representatives serving on boards of directors (for example, the German co-determination model). It can manifest itself in councils of workers having the dominant decision-making role in firms, as in Yugoslavia (that is, workers' control and self-management). Typically, these types are formal in the sense that the arrangements are prescribed in government rules or collective bargaining arrangements.

Participation can also, however, take the form of direct participation of employees in the everyday decision-making process on the shop or office floor. This may be either formally prescribed by management rules or collective bargaining arrangements or of a more informal nature (the voluntaristic model). In this book, I focus on direct participation, first, because this is an area in which the Swedes and the Japanese have been most innovative; second, because scholarly research suggests that it is the area that matters most to workers (see, for example, Kohn 1976:111-30); and, third, because direct shop and office floor participation is the primary focus of organized participatory work practices in the United States.

My examination of participation focuses on efforts to directly expand employee control in the everyday decision-making process on the shop and office floor. While the topic is direct participation, it is important to examine the process by which these "solutions" emerged in competition with potential alternatives, such as board representation. It is also important to ask why some forms of direct participation were selected over others and what the implications for the diffusion process were. Finally, it is important to keep in mind that these various approaches to participation are not mutually exclusive. Often we find mixtures of representative and direct participation, as well as formal and informal modes, within the same system of industrial relations.

To examine only part of the participatory system operative in a given country invites distortion. The benefits involved are, however, considerable. One is able to pursue an intensive study of a limited number of cases. Limiting the study to countries that have all emphasized small-group activities enhances the search for explanations of their similarities and differences.

Japan, Sweden, and the United States span a range of variation favorable to the task of comparison. The United States is an example of limited change relative to the other two, both of which have experienced a significant, although varying, degree of institutionalization of small-group activities. There is also variation in their social, economic, and political systems; Sweden may be characterized as social democratic, the United States as liberal capitalist, and Japan as liberal capitalist with a strong developmental thrust. These different political systems vary considerably in their dimensions: group interest, labor-market conditions, managerial problems, political organization, and institutional mobilization. Insofar as culture is a variable partially explaining small-group behavior, one might anticipate that Sweden and the United States, as heirs to the Western cultural tradition, would have more in common with each other than they would with Japan. This is a theme to which I shall return.

What's in a Name?

My analysis thus far is a bit too neat. What are we actually comparing when we say that we are trying to explain the spread of direct shop and office floor participation in decision making in the three

countries? It is hard to get a grip on the concept of participation in decision making.

This is because participation can occur in many work domains and in different parts of the decision-making process. There are small-group activities with greater and lesser degrees of employee participation. I selected small-group activity as a generic term that would not be weighed down by the specific connotations of direct participation . Such a neutral term makes it possible to examine the processes operating in the three countries with less likelihood of attributing values and intentions to actors that do not apply. While small-group activity is hardly a household term, at least one U.S. company, Hughes Aircraft Co., has created the title of manager of small-group activities.

In any case, my focus is on the introduction and spread of small-group activities at the workplace . Specifically, I examine quality-control circles in the United States and Japan and what were once known as semi-autonomous work groups in Sweden. As implied above, the small-group activities represented by these labels vary in structure and process across the three nations and indeed over time in a given nation.

What is involved in quality-control circles and self-managing teams? Since QC circles were introduced in Japan under the auspices of the Japanese Union of Scientists and Engineers (JUSE) it is appropriate to use its definition:

The QC circle is a small group within the same workshop that voluntarily carries out quality-control activity. The small group continuously conducts control and improvement of the workshop as one part in the chain of companywide quality-control activity. In this fashion, utilizing quality-control techniques, the small groups carry out self-development and mutual development.
(JUSE 1970:1; my translation)

The circles are relatively autonomous units composed of small groups of workers, usually of common status, in each workshop. A 1983 JUSE survey reports that typically circles had five to eight members. Some 60 percent of the groups choose their own leaders, and some 20 percent rotate the leadership; still others have staff specialists as leaders, or leaders are appointed by the supervisor (Lillrank 1984). The circles usually meet once a week for an hour or so; they are parallel organizations in that they are not directly a part

of the work process. The workers are taught fairly simple statistical techniques and modes of problem solving and are guided by leaders in the selection and solving of problems. The circles concentrate on solving job-related quality problems, broadly conceived as improving methods of production as part of companywide efforts. The circle solutions are presented to management for action, with the circle members having no authority to implement the solutions on their own. Quality-control circles are supposed to allow for the acquisition of skills by workers, the development of career potential, cooperative activity, and the like. If functioning properly, they should give workers a sense of control over their everyday activities on the shop floor.

As the concept of the circles spread to the United States, they became separated from their quality-control foundation in Japan. As a consequence, they also became less concerned with statistical methods and more focused on building process skills (for example, how to run an effective meeting). Thus, a case can be made that the circles themselves are not the same institution in Japan and the United States. Reflecting this difference, they are known in Japan as quality-control (or QC) circles, while in the United States they are generally known simply as quality circles.

The Swedish idea of self-managing teams or autonomous work groups^[1] is that workers make their own decisions regarding work allocation, recruitment, planning, budgeting, production, quality, maintenance, and purchasing. The group members are not merely carrying out a certain number of tasks. They are working together on a continuous basis to coordinate different tasks. They take responsibility for the organization of work and take the necessary measures to cope with the work of the entire unit. Job rotation among members at their initiative is seen as a normal part of work activity. A great deal of mutual help, joint responsibility for the total operation, and continuing opportunities for learning are expected.

This ideal has seldom been fully realized in practice, but it constitutes a set of objectives toward which many Swedish companies have made partial progress. We can see that even modest implementation of the concept of autonomous work groups would give workers greater control over the work process than is the case with

[1] I use the terms semi-autonomous work teams, autonomous work teams , and self-managing teams interchangeably.

quality circles. Workers in such groups can not only make recommendations based on their analysis, but can make decisions about their work and implement them in a broad framework established by management and unions. To be sure, such action tends in practice to have been focused on working conditions, task assignment, work methods, personal equipment, and establishment of daily routine, while broader management issues tend to be left to higher-level management and union officials. The former are the same work domains that occupy quality-control circles, but the self-managing teams, unlike circles, in principle involve themselves in personnel matters and budget decision making as well. In practice, too, self-managing teams often seem to have greater access to decision-making processes than do quality circles. Moreover, they are part of the work process, not a parallel institution.

In contrast to self-managing teams, there tends to be less of a gap between the ideal of quality circles and how they actually operate. This relatively close correspondence suggests that quality circles represent a tighter (better-defined requirements) and less demanding (more realizable in terms of fewer required changes from conventional practices) package than self-managing teams.

Some Differences among the Three Countries

Consider the following differences: in Japan the key term used to explain the innovation as it began to be applied was decentralization of responsibility . By decentralization the Japanese do not generally mean delegation of authority to offices down the hierarchical structure. Rather, for them decentralization means the taking of responsibility for objectives by large numbers of people (cf. Kalleberg and Lincoln 1989). Generally, the Japanese simply speak of small-group activities (shoshudan katsudo ). In Japan quality-control circles have allowed workers to play a greater role in the input phase of the decision-making process, but management retains control of the decision itself. The term decentralization of responsibility does not suggest a voluntaristic process in which workers organize to obtain greater participation for themselves or choose to democratize the workplace. Rather, we get a sense of small-group activities being imposed by management as part of its per-

sonnel practices. The extraordinarily high rates of participation in small-group activity in Japanese firms support this interpretation.

Small-group activities were commonly introduced as part of a corporate strategy to mobilize all the resources of a firm to overcome foreign and domestic competition. In this sense, management sought to make participation a responsibility, an obligation, of employees rather than to provide an opportunity for them to express their individual talents or "self-actualize," California style. Moreover, the term democratization was seldom heard. It was not until the late 1970s that one even began to hear of a stress on participation in management (keíeí sanka ) in a sense even approximating American usage. Rather, the focus initially was on engineers aiming to solve quality and cost problems in the workplace; the motivational aspects were an afterthought that only came to be explicitly discussed many years later.

One of the first scholars to talk about small-group activity in the context of participation was Ueda Toshio, whose 1975 book with Hirota Kimiyoshi (Hirota and Ueda 1975:25) discussed the newly raised expectation that small-group activities might lead to greater participation (sanka ). This was some thirteen years after the first quality-control circles were formed. The increasing tendency of interpreters of the movement to link it to participation occurred partly in response to positive American definitions of what they were doing. Prior to this time, when the Japanese used the term keiei sanka and spoke of democratization of management, they were not even thinking of small-group activities. This remains true to a great extent even today. Rather, the Japanese think primarily in terms of systems of indirect participation, such as the widely diffused labor-management consultation system. Internationally, the term keiei sanka conjures up for them images of the German co-determination system and other formal representational approaches in continental Europe.

There is another sense in which the term participation has been used by the Japanese, and it is quite revealing. This usage refers to the necessity of the act of participation in small-group activities. Shiba Shoji summarizes this perspective well when he bluntly states that the real meaning is that:

Every worker participates in the same workshop. Let us say there are seven workers in the workshop who work together on the same produc-

tion line in the workshop. All seven workers have to participate in the circle activities. Isolation of workers in the workshop is not allowed. Small-group activity with the participation of every worker is aimed at improving the work in which they are engaged. QCC activity is not an activity for amusement but an activity for the improvement of work.
(Shiba 1983a: 13)

We see clearly here just how Japanese use of the term participation contrasts sharply with how Americans have typically used it.

In the United States, managers often speak of participation in management, quality of work life, and employee involvement to explain what they are doing (Kanter 1983:44-46). It is precisely for this reason that we mistakenly equate participation in the American sense with what the Swedes and the Japanese have been doing in the area of small-group activities.

To a somewhat greater degree than in Japan, the terms democratization and humanization have been used—and often confused—in the United States. Nevertheless, this is a rather subdued theme relative to other terminology. One leading academic consultant explained to me in 1985: "I used to use the term democratization , but I don't anymore because I find it gets in the way of operational objectives." There appears to be considerable confusion about what is meant by the terminology used in the United States. A 1982 study by the New York Stock Exchange that focused on the concept of quality of working life (QWL) and investment in human resources (Freund and Epstein 1984:120) defines QWL as "the effort to encourage employees to participate in the key decisions that affect and determine day-to-day work patterns" (for more elaborate definitions of QWL, see Walton 1975:91-97). The concept of participation appears to be central to how most American managers have understood what they are doing. Management sees itself as developing tools to tap unused human resources through participation. The motivational consequences in terms of job satisfaction and worker morale are commonly highlighted.

In Sweden the early movement was strongly punctuated by expressions stressing joint influence and democratization of the workplace. In an influential book entitled Form and Content of Industrial Democracy , originally published in Scandinavia in 1964, Fred Emery and Einar Thorsrud argue that the objective of industrial democracy cannot be attained solely by worker representation on

boards of directors. Rather, this must be supplemented by an approach at a level where "a large proportion of employees are both able and willing to participate" (Emery and Thorsrud 1969:86). They have in mind, of course, direct shop and office floor participation. The Swedes have stressed changing power relationships between managers and employees at all levels. As one Swedish commentator notes, even when the North American advocates talk about workplace democracy as a goal of new work structures, they commonly have in mind "democratic leadership styles" rather than a transformation of structural relationships as envisioned by many Swedes (Leymann 1982:47).

A central element of the democratization theme in Sweden has been the focus on work group autonomy as an end in itself. The contrast here with the Japanese case is particularly stark. The Japanese version of participation emphasizes teaching workers conventional and nonconventional engineering and statistical skills to design and improve their own jobs under close managerial direction and control. Japanese managers have achieved a more thoroughgoing implementation of scientific management to an extent that its founder, Frederick Taylor, could not even have imagined. Whereas Taylor stressed the importance of managers' getting to know worker capabilities, in the relatively uneducated immigrant America of his time, managers saw these capabilities as quite limited. In contrast to that situation, Japanese managers have gotten increasingly well educated workers to participate in their own job design, thereby tapping powerful motivational principles associated with challenging workers' individual abilities and providing them a limited sense of control over the workplace (see Shimada 1988). To be sure, to accommodate worker interests, managers made some important modifications of traditional hierarchical control structures and scientific management practices. They moved toward decentralized operational decision making. This meant that workers together with production engineers could make many decisions with regard to organizing production as long as they met the demanding improvement goals set by management. It was in this context that circles emerged. Finally, in contrast to traditional scientific management, Japanese management aimed at building a broader rather than a narrower range of individual job skills for its labor force.

The popular symbols and choice of language used to characterize

a social movement tell us a great deal about the motivation of actors and the kinds of constraints they impose or have imposed on them. In the case of Japan, the focusing of the debate on decentralization of responsibility tells us that management was pretty much in charge and could impose its own categories and labels on developments. In the case of Sweden, the focus on industrial democracy tells us first that management did not have full control of the agenda. Moreover, the Swedes had a highly centralized labor-management decision-making system. This meant that advocates of semi-autonomous work groups in labor, management, and academic circles could argue forcefully and believably that there was something missing at the shop and office floor level in terms of democratic decision making. Based on that view, they further maintained that semi-autonomous work groups could fill that vacuum. In the United States, unlike in many Western European countries, there is a labor movement with an active shop floor presence, at least in the unionized sector. This has to a considerable extent preempted the industrial democracy theme. Instead, in the United States there is a rather more modest focus on participation and employee involvement , with the motivational consequences being given great emphasis.

The major implication of this extended discussion is that in examing small-group activities cross-nationally, there is inevitably an element of comparing apples and oranges. Small-group activities are not the same across nations. Rather than try to eliminate this difference in what is being compared—which I judged to be a futile exercise—my strategy has been to acknowledge those differences. The task then is to try to develop plausible explanations for the way these practices developed and the effects they might produce.

Three Snapshots

The United States

In the early 1970s extensive discussion of the need to humanize work began in the United States. We heard of job redesign, job enlargement, and job enrichment. Although the terms are often used loosely, the job humanization movement has focused primarily on the redesign of jobs, envisioned as occurring either horizontally

(job enlargement), creating more variety in the job, or vertically (job enrichment), expanding the domain of employee decision making. Although there are numerous similarities with the earlier human-relations movement (Berg, Freedman, and Freeman 1978), a distinctive characteristic of the new movement is its focus on changing the job itself.

Whatever the forms, the programs and proposals designed to humanize work have one common denominator: they all involve attempts to reduce alienation and increase job satisfaction by increasing employees' participation in workplace decisions and job variety, thus making more effective use of workers' potential. Ideally, these changes are supposed to improve organizational efficiency and productivity while enhancing the quality of work life. Improved quality of work life involves workers controlling those aspects of their work that directly affect their everyday lives.

When one examines the adoption and diffusion of small-group activities of this sort in the United States over the period 1960-85, one is hard put to say that many of the noble ideas outlined above have been institutionalized in large numbers of American corporate entities. In the 1970s self-managing teams won only very limited acceptance here. In companies like General Electric, where a number of plants adopted some version of self-managing teams, they had all died out by the 1980s. Generally, the number of plants adopting them was small, and the labor force of such plants was often atypical as a result of special recruitment procedures. The little that was accomplished in the 1970s was commonly done in nonunion plants. Generally speaking, more was done in new plants; far less was tried or succeeded in established plants. Even less was accomplished in office work. In addition, even in allegedly "successful" experiments, such as the General Foods plant in Topeka, Kansas, the innovation was slow to diffuse to other plants in the company, not to speak of its offices (Walton 1978).

Finally, there is more form than substance to many of the announced programs of U.S. companies. Given the ambiguity of goals and technology, it is not surprising that it is often difficult to judge what has actually happened, why it happened or did not happen, and whether what happened is good (see March and Olsen 1976). Organizations often develop myths about what they are doing in small-group activities, and, because of the lack of agreement over

what constitutes participation, they have a great deal of latitude to do just that. This can be used as a vehicle to encourage or discourage the diffusion process. Small-group activities can be adopted ceremonially, but in practice be buffered from actual work, so that little substantive change occurs (Meyer and Rowan 1977). This allows forms to be adopted without producing significant change. With their stress on "off-line" activity, quality circles lend themselves to this outcome.

I attended a number of company presentations in the late 1970s in which, once all the rhetoric was stripped away, the bulk of the participation program consisted of workers being able to have lockers near their work sites or the plant management having established a recreation program. In general, one sees many "programs" in the United States that have beginnings and ends, but relatively few instances in which small-group activities are highly institutionalized.

Notwithstanding these observations, there is a way in which, as Rosabeth Kanter reminds us, talking about innovation often serves as a precursor for it. She documents the transformation of topics and treatments in business periodicals between 1960 and 1980, noting the rise to prominence of "human resource management" and "participative management" (Kanter 1983:44-46). It is a sign of our times that a company such as Motorola used the slogan "Quality and productivity through employee participation in management" as part of its newspaper advertising in 1985.

The 1980s have indeed produced a burst of experimentation with small-group activities by U.S. firms. Inspired in part by the rapid internationalization of the economy, and in particular by the powerful competitive threat posed by the Japanese, both management and labor have been forced to reexamine the contributions of small-group activities. The New York Stock Exchange national survey of human-resource programs and QWL in 1982, based on a sample of 49,000 U.S. corporations (27 percent response rate), and including both manufacturing and nonmanufacturing firms, documents this explosive growth. Of the firms sampled with 500 or more employees, 15 percent had initiated quality circles. The figure rose to 22 percent in firms of over 5,000 employees. Overall, 19 percent of manufacturing firms reported adopting quality circles, compared to 9 percent of nonmanufacturing firms. In keeping with my observations about the suddenness of this growth, the study re-

ports that in corporations of over 500 employees, quality circles were the most rapidly growing human-resource activity. Indeed, the report states that quality circles had come to epitomize the QWL concept. Some 74 percent of the quality circles surveyed in firms of over 500 employees in the spring of 1982 had been established for less than two years, with 85 percent reporting less than five years' experience (Freund and Epstein 1984).

Yet caution is necessary in interpreting these data. For a firm to say it has initiated quality circles does not mean that a majority of its employees, or even a significant minority, are in fact participating in them. Indeed, one seldom finds large plants or offices, much less companies, in which more than a minority of the labor force are participating in quality circles. In my 1985 survey of 236 circle adopters, only 24 percent of respondents reported more than 25 percent of their workforce participating in quality circles, and only 11.9 percent reported more than 35 percent of their workforce doing so. Moreover, there are no data at the national level to indicate that quality circles have been institutionalized in American corporations. Circles also have a more limited impact and pose a more modest threat to management's hierarchical decision-making structure than self-managing teams.

Despite the initial excitement over quality circles in the United States in the early 1980s, by the mid 1980s the amount of attention being devoted to them in the business press and management journals had fallen off dramatically. Inside companies they were less and less a topic of conversation . They were now being publicly described as a fad, with discussion focusing on what would succeed them (Lawler and Mohrman 1985: 65-71). For many companies, self-managing teams became the new game in town.

Japan

The Japanese movement for small-group activities began in the early 1960s and accelerated thereafter. There was an increasing emphasis on decentralization of decision making; employees were to take responsibility for a variety of everyday decisions for which management representatives had hitherto been responsible (functions like maintenance, quality control, and safety). The vehicle for these efforts was "small groupism" (shoshudanshugi), as the Japa-

Fig. 1.
Annual New QC Circle Registrations in Japan, 1963-86. Source: Lillrank 1987.

nese called it. The idea was to make the small group the responsible unit in this decentralization effort.

A variety of surveys suggests that these practices are now widespread. The Japanese Union of Scientists and Engineers, which registers QC circles, reported some 250,000 circles, with over two million members, in early 1987. In addition, JUSE "conservatively" estimates unregistered QC circle members to total an additional five times the number of those in registered circles. I would suggest the more conservative figure of three times that number, which would nonetheless bring the total to about eight million employees, or roughly one out of every six Japanese employees.

Figure 1 depicts annual new quality-control circle registrations from 1963 to 1986. There was one relatively short burst of increases in the registration rate and one major sustained one. The first was from 1966 to 1969, when registrations rose from roughly 2,000 to almost 10,000 circles a year. The second occurred between 1977 and 1984, when new registrations rose from some 5,500 in 1977 to almost 30,000 circles a year in 1984. This latter point represents the peak of the diffusion rate; growth has since slowed, as befitting a mature movement.

The difficulty of estimating the amount and spread of QC circle activities arises in large part because many firms do not use the term quality-control circle , but rather engage in a variety of related small-group activities (typical categories of small-group activities in Japanese surveys in addition to circles include zero-defect groups, improvement groups, and Japanese versions of management by objective).

A survey of small-group activity by the Japanese Ministry of Labour in 1972 reveals that some 40 percent of firms with over 100 employees then reported such activity. For firms with over 1,000 employees, the figure was 70 percent, and in every size category firms with unions were more likely to have small-group activity than those without unions. About half the establishments reported that they established their small-group activities on existing organizational arrangements, with the rest stating that they had created new structures (Japan, Ministry of Labour 1973:36, 138). A 1977 follow-up Ministry of Labour survey reports strikingly similar findings, with 40 percent of the establishments with over 100 employees again reporting small-group activities (Japan, Ministry of

Labour 1978: 166-71). By 1984, however, 60 percent of establishments with over 100 employees reported such activities, with the figure rising to 84 percent for firms with over 5,000 employees (Japan, Ministry of Labour 1985: 14). These and other survey data suggest again that the growth of small-group activities in manufacturing has peaked. Manufacturing and utilities record the highest levels of small-group activity with services and the wholesale and retail trades recording significantly lower levels.

Although it does not show up in the aggregate data, JUSE reports that QC circles have spread rapidly in recent years to the nonmanufacturing operations of manufacturing firms (for example, sales and administration) and to the service sector, particularly banking, insurance, and hotels, and including even restaurants, nightclubs, and bars. Finally, in contrast to the United States, it is typical to find firms and plants and offices reporting over 90 percent participation in small-group activities. Just how typical can be seen from a JUSE survey that found that 74 percent of the companies with quality circles had a policy requiring all employees (zenin sanka ) to participate (Lillrank 1984:8). That many of these small-group activities have survived for twenty or more years suggests that reasonably high levels of institutionalization have been achieved. Relative to the United States, there is little basis for thinking of the quality-circle movement in Japan as a fad.

Consistent with earlier discussions, a major characteristic of QC circles is that they tend not to threaten the hierarchical structure of authority as much as some other forms of direct participation. First-line supervisors tend not to be threatened by the circles, and indeed in the initial stages of the movement, they often served as leaders of them. In my survey of the 267 early adopters of QC circles in Japan, I asked managers: "Is it the experience of your company that the scope and/or content of the first-line supervisor's authority must be changed significantly in order to have this form of small-group activity function properly?" Only 6.5 percent of the respondents answered yes, with the remainder answering no.

Sweden

Swedish efforts to develop small-group activities crystallized in the late 1960s with emphasis on autonomous or self-steering work groups (självstyrande grupper ) as the unit of production. These

ideas about decentralization of decision making spread rapidly. There was a long and intensive public debate centering on broad issues of democratization and social justice. It is more difficult to evaluate how well these activities are actually diffused in Sweden than in the case of Japan. The reasons are themselves revealing; they reflect the Japanese penchant for national surveys and systems of formal registration. On the Swedish side they reflect, in part, the strategy of diffusion adopted by the technical department of the Swedish Employers' Confederation (SAF).

As noted earlier, the goal of autonomous work groups is for workers to make their own decisions regarding work allocation, recruitment, planning, budgeting, production, quality, maintenance, and purchasing. It is an ideal seldom, if ever, realized in practice, and certainly far less commonly achieved than the English-language public relations efforts of SAF and of companies like Volvo would suggest (Gyllenhammar 1977; Swedish Employers' Confederation 1975). Yet considerably modified versions of this decentralized work system did come into effect in most large Swedish firms, and, unlike in Japan, they spread rapidly to the public sector as well. Gunzburg and Hammarström (1979:39-40) report estimates of well over one thousand firms with significant work reorganization efforts in the mid 1970s. Albert Cherns (1979:360-61) ranks Sweden first among European countries in the extent to which the "quality of working life movement" had been diffused. Volvo, the most prominent Swedish innovator, points out that since the opening of the revolutionary and much-discussed Kalmar plant, with its group work organization, in 1974, it has built nine new plants whose design has been strongly influenced by the Kalmar experience.

Yet it has not been easy to maintain the early momentum of shop and office floor innovation. During the 1970s some twenty new laws relating to labor and national collective agreements came into effect, mandating representation for employees on boards of directors, new labor welfare measures, changes in working conditions, and a co-determination law among other things. There was a shift toward working out the implications of this new legislation, particularly the co-determination law. Many shop and office floor participatory efforts dried up or failed to diffuse from experimental units to the wider organization. Such was the case with Atlas Copco, a large manufacturer that figured prominently in the early days of

experimentation. The uncertainty surrounding the loss of power by the Social Democratic government in the late 1970s also contributed to a wait-and-see attitude on the part of many employers. The struggle over the "wage earners' fund" proposed by the Social Democrats further diverted attention from small-group activities.

It was not until the 1982 agreement between the unions and the employers' federation, not long after the Social Democrats had again returned to power, that the framework for the application of the co-determination law was put in place and one began to see new life in the shop and office floor small-group-activities movement. But it is an approach chastened by past experiences. Managers no longer talk about autonomous or self-steering groups; instead their vocabulary stresses group work (grupparbete ), group organization (grupporganisation ), teamwork (lagarbete ), and "the little factory in the large factory," with a renewed stress on productivity and quality improvement and new concern for flexible manufacturing systems and proper use of databases.

Based on the above observations, it seems reasonable to conclude that in the three countries surveyed, small-group activities have spread most rapidly and broadly in Japan, followed by Sweden and then the United States. However, we must constantly remind ourselves of the partial character of such observations. We are recording the tip of the iceberg. Much small-group activity is of an informal nature that is not easily recognized even by participants, much less recorded by outsiders.

Moreover, these observations do not mean that Japan has the greatest amount of employee participation in the American sense of that term or the most democratized workplace in the Swedish sense of that term. Consider the findings of the 1983 cross-national study of workers' attitudes sponsored by the Aspen Institute for Humanistic Studies and the Public Agenda Foundation, which surveyed the amount of workplace discretion (freedom to make decisions about work) employees said they had. Across industries, and for both white-collar and blue-collar jobs, the researchers found that Swedish employees consistently scored higher on their scale of employee discretion than did Japanese employees (Yankelovich et al. 1983:44-45; Zetterberg et al. 1983:140-41). Clearly, the amount of discretion employees have at work is a major factor in how much control they have over their immediate work environment.

These findings underscore my earlier observation that the spread of small-group activities in Japan was tied to management decentralization of responsibility rather than to voluntary employee participation in decision making per se. One can assume more responsibility without necessarily having a sense of more discretion. We should not impute increased employee participation and control to the spread of small-group activities without sound empirical support. Moreover, the amount of discretion at work that employees report arises not only from small-group activities but from other sources as well. Not all cross-national differences in work relationships result from recent experiences with small-group activity.

Chapter Three
Decision Making and Research Strategies

Central to this study is the concept of innovation. I am interested in how managers, union officials, and employees came to conceive, introduce, and spread the novel approach of using small-group activities in industry—a social technology—to solve problems and sometimes to organize production. I define innovation as a technology new to a given organization and diffusion as the process by which this innovation spreads to different organizations and to different parts of the same organization (see Tornatzky et al. 1983:2). Although the rate of diffusion may be faster in one country than in another, they may all end up in the same place over time. This represents a convergence perspective based on the assumption that there are functional forms of diffusion models, an approach that would leave little room for the long-term influence of cultural factors.

I shall examine the incentives to adopt such new practices and the nature of the obstacles that stood in the way of such adoption. In examining the spread of new ideas from one location and set of individuals to others, I ask: What are the patterns regulating the flow of ideas? How is it that some social strata get identified as carriers of these new ideas? Are there special agents of diffusion? How does one set of ideas get established, if it does, as "best practice"^[1] and come to be seen as a general solution to a particular set of prob-

[1] Best practice is a term used in economics to suggest that there is one specific best way (best technology) to accomplish a particular objective.

lems? Many of these questions, though sometimes expressed in different terms, have been asked in typical studies of the diffusion process, and this book will be no exception (cf. Rogers and Shoemaker 1971).

Still, there are a number of problems with the traditional social science literature on diffusion that need to be addressed. In dealing with the political sphere, the classic literature on diffusion centered on the development of an appropriate infrastructure for diffusion and the receptivity of adopting units, rather than on how these factors are influenced by the interplay of interests in the broad political and economic arenas (see, for example, Rogers and Shoemaker 1971). It is for this reason that I shall weave the perspective of political economy into my analysis.

Most important, one needs a causal model of decision making to order and explain the process of social change. It has often been pointed out that studies of the diffusion process seem to lack theoretical content and therefore ultimately reach a dead end. What kind of theory of decision making will help us understand the diffusion of small-group activities?

Small-group activities take place, or do not take place, in specific local organizations, which may create organizations at the national level whose mission it is to spread these new practices. In the end, the latter organizations are the main focus of my analysis. We need a mid-range theory of decision making that provides an explanatory framework for these organizational and interorganizational activities.

The test for a useful theoretical framework is what provides the best fit with the data and forces the researcher to ask novel questions, thereby revealing fundamental characteristics of the phenomenon under investigation. Theories of "loose coupling" partially fulfilled these objectives for this research project. It is a decision-making model that allows us to flesh out the skeletal characteristics of the diffusion process through a set of contingent decisions made by the various actors.

The Loose-Coupling Model

Social scientists developed the "loose-coupling," or "garbage can," model in the 1970s, primarily in the context of their study of educa-

tional institutions. In this model the stress is on more or less serendipitous combinations of problems and solutions. That is, decisions occur when elements from streams of problems, solutions, participants, and opportunities come together. But they often come together under conditions of great uncertainty, such that there is a weak correspondence between specific problems and particular solutions.

This model may be juxtaposed to the traditional rational model of decision making in which problems are identified and a search is instituted for the most suitable solutions. In the rational model, means are tied to ends in a logical fashion, actions are tied to intentions, solutions are guided by imitation of one's neighbor, prioritizing of goals takes place, feedback and evaluation control subsequent decisions, and past experience constrains present activity.

In the case of small-group activities, the rational model would work as follows. Managers identify a set of problems they are having, such as poor productivity, low worker morale, or lack of control over the labor force, and after a search conclude that small-group activities will best address these problems. They learn what "best practice" is and they introduce small-group activities, constantly modifying their practices based on feedback so as to best achieve their objectives.

My research suggests that while such a rational model can be relevant to certain parts of the adoption and diffusion process, a fuller accounting requires a much more flexible theoretical model. For as managers and employees wrestled with the adoption process, what I often saw were ambiguity, uncertainty, ignorance, conflicting goals, and "solutions chasing problems." Managers don't know what they want until they see what they can get; in this sense, solutions are used to formulate problems.

How can this be? The early research on loose coupling and the garbage can model was conducted in universities, on school boards, and in local government institutions that were characterized by an "organized anarchy" in the sense that the objectives or participants often varied, best practices were hard to define, and responsibilities for decisions often unclear. We, however, are examining business firms in the private sector, where objectives can presumably be stated relatively clearly, where a technology associating alter-

natives with outcomes is reasonably well known, and where there is a stable division of labor by which specific individuals and groups specialize in certain decisions (Cohen, March, and Olsen 1976:24).

The resolution of this seeming paradox lies not in the characteristics of participation or small-group activities per se but in the fact that we are talking about innovation. Uncertainty and ambiguity are the central features of innovative activity, and they are resolved only over time through learning activity. Louis Tornatzky and his associates arrive independently at this same conclusion when they state that "a typical innovation process is likely to have many of the aspects of 'garbage can' decision sequences" (1983:18). Characterized by uncertainty and ambiguity, the innovation process is like a floating capsule within the organization, partially insulated from normal organizational rules and processes. Indeed, one of the areas of ambiguity is the extent to which it should be insulated from normal rules. Only with the passage of time and reutilization of the innovation does the protective wall dissolve and "normal functioning" resume.

We can see the uncertainty operative in the introduction of small-group activity through examining the organizational choices it represents for a firm. Even allowing for significant national variation, we find that decisions to adopt small-group activities involve organizational choices in which goals are problematic and technologies are ambiguous and only partially understood. Participation in the decision to introduce small-group activities is itself often fluid.^[2]

Consider first the issue of goals and objectives. For management, there are many possible diverse goals that can potentially be served by small-group activities. Among them are increasing productivity, increasing product quality, reducing the ratio of indirect to direct employees, improving worker morale and reducing alienation, democratization of the workplace, meeting public pressures, more effective utilization of the labor force, good public relations, reducing employee turnover and absenteeism, reducing worker grievances, weakening unions, and developing cooperative real-

[2] March and Olsen (1976:38-53) refer to this as "attention structures." Studying the process of participation in decisions about participation contains its own set of ironies.

tions with the unions. While some of these objectives are not additive and are achieved only at the expense of others, others complement one another. Moreover, a variety of measures other than small-group activity may potentially contribute to the solution of each of these "problems." In addition, action taken to achieve some of these objectives might have adverse consequences on the achievement of still other organizational objectives.

Management goals will vary depending on the level of management. Middle management often sees small-group activity as a threat to its authority and prestige, and top management may see small-group activities as an opportunity to cut out layers of deadwood in management. Other parties, such as unions and workers, will have their own objectives. In sum, the goals toward which small-group activity is directed are problematic, with management having difficulty choosing the objectives to be served.

A similar case can be made for the ambiguity of the technology, by which I mean the technique of conducting small-group activity. The dispute over technique has been considerable. Scholars have often introduced their ideas on small-group activities at a high level of abstraction, so that practitioners are at a loss as to how to proceed. Competing agents of diffusion will offer different solutions on which small-group activities to adopt and how to execute them, with the potential user again left in a confused state. If failure occurs, it can be attributed to poor technique or the inappropriateness of the overall approach. Moreover, various proprietary interests may limit and distort the flow of information between firms, The belief that every firm, plant, and office has a unique culture that requires a special adaptation of the technology makes it difficult to specify the same appropriate technology across firms.

The difficulty of measuring the outcomes resulting from use of the technology of small-group activities and relating these outcomes to specific inputs also contributes to its ambiguity. Although some of the specific models being advocated are presented as self-contained packages, many provide little in the way of detailed guidelines. This diffuseness makes unclear just how much of an impact adoption will have on the organization in the long run and how much and what else will have to change in the organization as a result of adopting small-group activities.

All this further reinforces the ambiguity of the technology.^[3] Managers prefer packaged solutions whose cost and outcomes are, if not guaranteed, at least defined and limited (Cherns 1979). I shall examine how the three nations handled this ambiguity in technology, whether there were variations in the degree of ambiguity of different packages, how these ambiguities were resolved over time, and whether the resolution involved joint organizational activity.

The issue of who can, should, and does participate in the decisions by the firm to innovate in the area of small-group activity is also commonly unclear, at least in the initial stage. James March and Johan Olsen (1976) argue that this follows logically from a situation in which goals and technology are unclear.

Consider the case of the United States. My survey of American firms that have adopted quality-control circles shows that sponsorship was widely dispersed among a broad range of departments, from quality assurance to personnel. The personnel department often has an adversary relation with labor and a vested interest in conflict. Consequently, it is often ill-equipped to undertake such initiatives. Other departments, such as human-resource development and training and education, are possible candidates for the role of sponsor, except that, like personnel, they are commonly low-status departments without the necessary clout in the organization to ensure the acceptance of proposed work reorganization. Quality assurance sometimes took responsibility, but it, too, tended to have relatively low status, at least in the late 1970s and early 1980s, when much of the initial experimentation took place. Moreover, at a number of companies I discovered that small-group activities emerged in the late 1980s as disputed turf between the newly revitalized quality department and the labor relations managers. Finally, the relative obligations and rights of staff versus line operations personnel to make and enforce decisions about small-group activities were often unclear in many firms. Staff were given the responsibility of undertaking small-group activities, but only the line people could "make it happen."

The ambiguity of goals, technology, and who is to participate in

[3] Sahal (1981:60) adopts a similar view on the ambiguity of technology transfer. He focuses on the "technical know-how" bottled up in the system of its origin and the need for painstaking adaptation of the technology to its new environment.

the decision to innovate all make the choice process surrounding small-group activity one in which loose coupling is the norm. Only with the passage of time and either the dropping of the innovation or its routinization do these conditions disappear.

Finally, in the context of the loose-coupling model, it is important to examine the process by which given models of small-group activities were selected over available alternatives, and by what criteria and with what incentives for individuals and interest groups. It is also important to investigate whether certain characteristics of specific organizational sponsors, such as their membership, constituency, prestige, or size of budget, aid significantly in explaining the success or failure of the diffusion of given models of small-group activity (see Hirsch 1972). Where possible, I specify how the agenda of opportunities for choice emerges among relevant interest groups and organizations. This is a major focus of my investigation, especially in part 2.

The Microeconomics of Decision Making about Small-Group Activities

Thus far, I have concentrated on the nature of the problems faced by individual firms considering the adoption of small-group activities. Problematic goals, ambiguous technology, and confusion about decision rights are characteristics of the innovation process and define the loose-coupling model. Now I want to address the question of whether there are ways in which these problems can be resolved more efficiently than through trial-and-error corporate activities.

The analysis in the second half of this book concerns the development of the national infrastructure for the diffusion of small-group activities. That is, it deals with the macropolitics of the spread of small-group activities. Embedded in this focus is a central premise of the book: that the building of a national infrastructure for diffusion has been central to the successful spreading of small-group activities. By "successful spreading," I mean that this infrastructure has the potential to contribute to smoothing the decision-making process, thereby resolving the various problems just described more quickly and efficiently.

Yet there are many innovations that spread through the mar-

ket as firms adopt them without requiring the creation of a national infrastructure. What is it that distinguishes these alternative approaches? Why would the spread of small-group activities be strengthened through the support of a national infrastructure? Why can't firms themselves simply build or acquire the necessary expertise to conduct small-group activities? Here the macropolitical approach requires an understanding of the microeconomics involved in firm-level decision making.

Transaction cost analysis, the framework elaborated by Oliver Williamson, provides a basis for understanding the underlying principles involved (Williamson 1985). I provide a very brief overview to set the stage for subsequent analysis. Transaction cost economics maintains that the choice of institutional arrangements (governance structures) results from the underlying differences in the attributes of transactions. This involves "an examination of the comparative costs of planning, adapting, and monitoring task completion under alternative governance structures." The expectation is that the firm will adopt a strategy of economizing on transaction costs under conditions of uncertainty.

A key concern is identifying under what conditions a firm will internalize a particular function, as opposed to contracting for its provision. Attributes of the contracting process include above all asset specificity. Here the question is: Are the assets that are developed "dedicated assets" or can they easily be redeployed for use in other transactions with other firms? In the area of small-group activities, we are concerned primarily with human asset specificity and, by extension, with what I would call "organizational routine" assets.

The other two attributes of importance are the behavioral assumptions of bounded rationality and opportunism. Bounded rationality refers to the fact that human agents intend to be rational but are only partially successful. Opportunism is defined as seeking one's own interest with guile. Creating and taking advantage of an asymmetry in information with one's bargaining partner with regard to one's true expertise or costs is an example of opportunism.

The principal ways in which transactions differ are in asset specificity, uncertainty, and frequency of occurrence, and it is the nature of transactions that will determine which governance struc-

tures (types of organizations) are appropriate. We can clarify this most easily by turning directly to elaborating the case of small-group activities.

In all three countries, businesses faced the issues of whether to adopt small-group activities and, if so, how to go about acquiring the necessary expertise. They could choose to develop such expertise themselves, building their own internal competence and relying on sketchy reports in the public domain; they could choose to acquire much of that expertise through the market, primarily through contractual arrangements with consultants; or they could choose to join forces with other firms to build a national organization (or take advantage of an existing one) to serve as a forum for a dialogue among companies and develop and diffuse the necessary assets—human capital and organizational skills—to successfully implement small-group activities.^[4] There were a variety of collaborative arrangements that could be considered, some of which I shall discuss in describing the various forms that developed in the three countries.^[5]

Which governance structures should be selected to provide the organizational expertise required for introducing and operating small-group activities? Let us consider the abovementioned options in order. Firms could simply opt to develop the required expertise among their existing employees. But this would be a laborious task given the vast difference between these new practices and conventional practices, uncertainty about what is required, and the rapid evolution of best practices among the many firms experimenting with them in each country. There are large start-up costs associated with instituting small-group activities. The cost of creating state-of-the art quality-circle training materials for managers, facilitators,

[4] This approach is broadly comparable to William Ouchi's (1984) application of Williamson's framework to his analysis of industrial associations. Ouchi, however, never works out the linkage to the specific elements of Williamson's theory that make such associations a logical approach to governance.

[5] Other nonmarket collaborative solutions include, for example, visiting firms that have greater experience in the area of interest. This is widely practiced (usually based on a norm of reciprocity) even among firms in the same industry. While useful, such visits are typically of limited duration, and there is the possibility that information viewed by the lender as proprietary will be held back, that the information may not be adapted to conditions in the borrowing company, and that it may lack a "hands-on" quality.

leaders, and group members in the United States in the early 1980s, for example, easily ran to a quarter of a million dollars and more. Nor does this cover a variety of other start-up and operating costs, such as staffing, and opportunity costs, including time lost from production by group members.

Even large firms have difficulty recovering the extensive costs associated by spreading their application over a wide range of clients; small-group application is limited primarily to one's own employees. While small-group activities among one's employees may be frequent, the size of each transaction is small. In a few cases, especially in the United States, some large companies have tried to market the expertise of their small-group-activities staffs to recover their investments, but success has been limited. Notwithstanding, larger firms stand a better chance of recovering their investment costs than do smaller firms and thus have somewhat less of a need to contract for outside help.

One could use the market to replace existing employees with new employees with the required small-group activity skills were such individuals available, but this would obviously disrupt organizational performance dramatically if done quickly. After all, small-group-activity skills are only one of the many qualities that one would want employees to have. On the other hand, if such a substitution of employees is done gradually, one risks losing competitive advantage to other firms that are moving more rapidly to gain the productivity and quality advantages associated with effective small-group activity.

Yet the asset specificity of the human and organizational skills required to operate small-group activities is high. To be sure, there are generalized leadership, content, and process skills involved. Notwithstanding, the effective integration of small-group activities with an adaptive authority structure and flexible shop or office floor practices requires the development of specialized norms and practices. Individual employee "ownership" of these new practices is required for them to be effective, and that requires in turn that the expected users participate in their design. Thus it is ultimately firms that have to undertake and manage such activity. It cannot be farmed out.

We see, however, from this analysis that while firms have to do it themselves, they need help to acquire the necessary expertise.

They need help to obtain the general purpose investment in small-group activities that they can themselves later adapt through continuous experimentation to produce their own specialized version. They need help in clarifying goals, reducing ambiguity about the technology, and sorting out individual decision rights. There are two primary sources of such help: consultants and the pooling of corporate resources.

The consultants range from one-person companies to the specialized staffs of large consultant firms. Especially in the United States, these firms are often staffed by individuals who have acquired "know-how" from leading small-group activity developments in the private sector and then quit to try their hands at selling such knowledge. As we shall see, there is sharp variation among the three countries examined in the extent and type of consultant activity.

What is the appeal of contracting with such consultants for firms seeking to develop small-group activities? Most consultants in the United States offer an implementation package based on materials they have developed, purchased, or appropriated (from other consultants or firms using small groups). Consultants have typically had access to the experience of many companies and can therefore provide a pool of knowledge. It is in their interest to develop a generalized package that they can then market to a large range of companies. As we have seen, it is in the adopting companies' interest as well to acquire general purpose knowledge. Consultant companies typically offer implementation packages with known and fixed costs. For those companies that see small-group activities as a simple package to be slotted into existing structures and practices, this is particularly appealing. Such companies were more common in the United States than in Japan or Sweden.

An internal survey of managers at one large U.S. manufacturing company found managers giving the following reasons for seeking an external consultant for small-group activities in the early 1980s:

Need for expertise in how to implement employee involvement; nobody knew what to do to start the process

Perception that a consultant could help overcome local inertia in getting the process started

Belief that a third party would be credible to all parties involved in implementation (especially important in plants with strained union-management relations)

Feeling that an outside consultant could be candid to all parties

Perceived need for a coach as the process unfolded

Perception that a consultant could resist and buffer demands by higher management for premature evaluation of the process

Perception that it was division policy to use outside consultants

There are, however, considerable drawbacks to consultants. Above all, U.S. managers often reported concern about consultants being opportunistic. What is the nature of such opportunism? Above all, consultants may misstate their expertise. In the late 1970s and early 1980s, there was a bandwagon effect as new consultants entered this new and rapidly evolving field in great numbers. There was also tremendous uncertainty among managers as to what was involved in engaging effectively in small-group activity. Under such conditions, the cost of verifying the expertise of consultants was considerable.

The extent of these costs can be seen in the experiences of the Ford Motor Company, whose Personnel and Organization Staff designed a handbook in 1984 to help units decide whether they needed outside consultants, how to screen them, and how to deal with them in connection with their employee-involvement (EI) activities. The handbook itself was the product of fifty hours of interviews with Ford managers and union officials and required a task force of twelve managers. Suggested procedures:

First evaluate the adequacy of internal resources

Review external consultants certified for EI work within the company and choose a number of them for further consideration

Contact locations that have used consultants being considered

Compile a list of several candidates for review

Invite candidates in for on-site interviews with the steering committee or operating committee

Review the services each consultant provides

Determine consultant's range of resources that can be used to support the growth of the local EI process

Identify a consultant who can relate to employees at all levels of the organization and gain their trust and confidence

This list does not exhaust the factors that the handbook recommends local managers examine, but it does make clear that there are considerable costs in time and manpower associated with trying to determine the qualifications of consultants. As implied in the second item, Ford adopted corporationwide certification of consultants; this represents a centralized screening procedure to reduce costs and the likelihood of error. At Ford that certification process in turn necessitated a corporationwide panel to review the credentials of consultants, which conducted a ninety-minute group interview with prospective consultants.^[6]

Beyond such general sorting activities, the costs to local units of finer screening to meet their specific needs are substantial. Smaller companies with more limited resources find it more difficult to evaluate, and therefore more expensive to screen, external consultants. However extensive the screening and however large the company, the danger that consultants may exaggerate their contributions and the company may not be able to see through self-serving claims remains significant.

Furthermore, managers may fear that consultant information and materials are slanted in ways that encourage continued dependence on consultants and their training materials. To rely on consultants to keep abreast of changes in best practice entails continued expense.^[7] In practice, most companies reduce their use of external consultants as they gain experience with small-group activities.

Apart from opportunism, consultants have little specific knowledge of their client firms. There are thus considerable limits to

[6] This corporationwide activity was eliminated in the late 1980s as reliance on external consultants diminished and the central staff developed greater confidence in the sophistication of local units in such matters.

[7] It should be noted that the relationship is fraught with opportunism from the consultant's point of view as well. I witnessed a number of cases in which firms passed a rough rewrite of consultants' training materials off as their own, ignoring the (obviously weak) copyright protection. This enabled them to avoid royalty payments for each set of training materials purchased from a consultant. Other companies bought a few copies and then let their copy machine_s take over, again avoiding royalty payments.

their potential contribution to integrating small-group activity with shop or office floor practices. While some consultants are willing to acquire such knowledge with a view to building long-term relationships with a firm, this can be an expensive proposition from the firm's point of view.

For firms or unions that see small-group activity as part of a national movement (whether to gain competitive advantage or advance industrial democracy), such piecemeal efforts to acquire general purpose knowledge about small-group activities are inadequate. An alternative is to pool corporate and/or union investment (formally or informally) in industry- or nationwide infrastructure-building activities.^[8] By absorbing what was going on in the field, refining it, and feeding back best practices to individual firms, the resultant associations could develop the general purpose knowledge required. They could conduct research on promising new directions and help forge a consensus on what these were, thereby smoothing the decision-making process within each firm. Such associations are instruments that allow competitors to develop a non-adversarial relationship for mutual advantage.

As agents of their constituencies, such associations are not very likely to engage in opportunistic actions.^[9] Firms drawing upon their services can still develop competitive advantage through more effective tailoring of the associations outputs with their own norms and practices. Moreover, to the extent that a national associations general purpose knowledge is superior to the general purpose knowledge being generated by international competitors, cooperative action among domestic firms can still yield international competitive advantage for all. Yamazaki and Miyamoto (1987: 313) make the same point in their comparative study of trade associations.

Such associations, with their wider network, in principle offer greater opportunities for keeping up with the developing standard of best practices than do individual consultants. For one thing, they are less likely to conceal failure and more likely to share learning that results from failure. Finally, such nationwide associations offer

[8] These collaborative efforts may be seen as institutions midway between markets and hierarchy, to use Williamson's categories.

[9] To be sure, there is a literature on the problem of accountability of agents to owners (see Jensen 1983), but relative to independent consultants it seems clear that the incentive for opportunism is likely to be limited.

the possibility of creating a national social movement for small-group activities. This can be organized in conjunction with unions under the banner of industrial democracy, but it can proceed without union involvement as well. These efforts can provide a motivation for employees to engage in small-group activity that neither individual firms nor consultants can offer.

This is not to say that such national associations do not have drawbacks. They may become caught up in bureaucratic procedures that strangle innovation, or they may fall victim to internal political struggle among member firms. Should the association stress developing new applications of small-group activities to better serve its more sophisticated members, or should it concentrate on innovative ways to get firms started on small-group activities?^[10] Notwithstanding such dilemmas, associations on balance have tremendous potential for advancing the small-group-activities movement.

These are all issues addressed in the second half of this book. Central to consideration is the question of whether firms in any given country took greater advantage of the collaboration and pooling of resources in the associational framework to better spread small-group activities and capture their benefits. In so doing, there should have been less of a garbage can quality to their decision-making process. My purpose now is simply to demonstrate that the evolution of a national infrastructure can be crucial to the spread of small-group activities. This evolving infrastructure economizes on bounded rationality and minimizes opportunism.

Areas of Comparison

To provide a map for the subsequent chapter-by-chapter analysis, I take as a starting point the standard rational model of the diffusion process. It generally involves the following stages:

Such stage models are useful intellectual tools for simplifying a complex process. The complexity of the actual decision-making

[10] Ouchi (1984: 181-84) documents comparable conflicts in associational activity.

process, occurring as it does at many levels and involving many individuals, is such that it is, in fact, extremely difficult to see the innovation process as one in which decisions feed each other in a linear or logical sequence. One important piece of evidence in this regard is that decision makers themselves are often unaware of just what stage they are going through (Tornatzky et al. 1983: 19). Indeed, this is the point of my using the loose-coupling model as a guide.

Notwithstanding, I shall follow the above model for the subse-quent chapter divisions, always, however, keeping in mind the way the ambiguity of the problems and solutions suggests different processes. This latter consideration requires exploring the possibility that "latter" phases precede "prior" phases and in particular the conditions under which discovery and transmission precede the search stage. Exogenous developments also display an ability to transform the process. The political dimension operates as a crosscutting variable that has an impact on all stages. To begin at the beginning requires an exploration of the motivation for small-group activities. This is the subject of the next chapter.

Once this exercise is complete, I shall turn my attention in chapters 8-13 to the nature of the national infrastructure for diffusing small-group activities in the three nations. This infrastructure has the potential to influence decision making in all phases of the diffusion process.

Chapter Four
The Motivation to Innovate

The greatest disservice that fashion does is carelessly to turn life's most precious and fragile assets into marketable products of transient worth.
Kennedy Fraser, The Fashionable Mind

It might seem a simple matter to read the explanations of key actors as to the central problems they saw themselves facing and follow the problem-solving activity that led to the identification of small-group activities as a solution to problems of work organization. The social reality is a good deal more complicated. Many possible incentives and precipitating factors are capable of producing the set of decisions that lead to the introduction of small-group activities. Moreover, not all those factors are equally salient for the many social actors involved in the process, and their importance varies according to time and place.

Consider first the matter of the explanations given by key actors. If we allow for the possibility that experience creates values, rather than the conventional emphasis on value preferences determining choices, the problem of reconstruction becomes far more difficult. The diffusion of quality-control circles in Japan provides a textbook example of how interpretations and explanations involving the value of participation per se were called forth long after the initial innovations (March and Olsen 1976). As we saw in chapter 2, participation emerged in Japan as an ex post facto rationalization of the small-group-activities movement. While Japanese managers were

indeed responding to a specific set of problems as we would expect based on a rational model of decision making, they were not necessarily the problems they now cite. Such rewriting of history is hardly unique to the Japanese.

Notions of democratizing the workplace, reducing worker alienation, and giving dignity to the worker are often given as motives for adopting small-group activities in Western countries. Sidney Harman, one of the early management pioneers of the new movement, spoke in 1972 of the need to see "work satisfaction—which is to say the attainment of a sense of purposefulness in his or her work, the achievement of a sense of personal worth and dignity— ... as a fundamental right of employees and therefore a fundamental obligation of employers" (U.S. Department of Health, Education and Welfare, Special Task Force 1973: 25).

Typical statesmanlike pronouncements are made on the subject by academics and high-level company and government spokesmen. Notwithstanding, one must establish whether there is an incentive structure in the organization that would encourage managers to identify alienation and lack of democratization, work satisfaction, and dignity as problems and small-group activities as a solution to these problems. By and large these incentives do not seem to have been as significant as the many public statements would suggest, except insofar as national political pressures were brought to bear. Even then because our cases are market economies, a focus on managerial incentives seems very much to the point.

None of this is meant to suggest that managers are any more disingenuous about the motives for their social behavior than any other group. Sociologists have often noted the need of individuals and organizations to rewrite their personal and organizational histories to fit current circumstances (Berger 1963). Rosabeth Kanter goes so far as to suggest this is necessary for organizational change to occur (1983:278-306).

An Analysis of "Structural Causes"

Apart from the difficulty of relying on the explanations given by key actors for their motivations, one may also question the ease with which we can identify the structural conditions underlying the decision to adopt small-group activities. Thomas Sandberg takes up

this issue with an interesting analysis of how it came to be that small-group activities in the manufacturing sector diffused more widely in Sweden than in Norway in the 1970s. According to Sand-berg (1982:93), the likelihood of adoption of small-group activities increases with:

the share of people employed in industry;

the importance of industry for the national economy;

the share of the engineering industry—including the machinery and transportation equipment sectors—in total industry;

the size of workplaces and firms;

the degree of work rationalization—efficiency brought about by introduction of productivity improvement machinery and work reorganization;

the level of welfare;

the level of employment;

the share of the population in the labor force; and

the share of women, immigrants, and other new groups in the labor force.

Sandberg recognizes that these structural explanations alone are insufficient for understanding the emergence of small-group activities, but nevertheless makes a plausible case for showing how these conditions facilitated developments in Sweden and inhibited them in Norway (Sandberg 1982:96-128).

What happens when we place the United States and Japan in the picture alongside Sweden? The utility of these structural explanations disappears; they cannot explain the speed of the Japanese in widely diffusing small-group activities relative to Sweden, much less the United States. By some measures, such as size of workplace, industrial concentration, the importance of industry, and degree of work rationalization, one might well have predicted that the United States would be the earliest and most successful adopter of small-group activities, with Japan coming last.

Consider the impact of labor shortage, which seems to underlie a number of Sandberg's measures, though he has relatively little to say directly on the subject. It might seem a simple matter to operationalize this structural condition in terms of employee-turnover

data, the assumption being that high rates of voluntary employee separations are associated with labor shortage as employees exploit their superior bargaining power with employers. This, in turn, leads employers to search for new approaches to committing their labor forces to the firm. If we examine such data for the 1970s, we find that turnover rates, even allowing for differences in data collection and presentation, were high in Sweden and the United States and low in Japan. Koshiro< (1983: 67) reports that monthly separation rates for workers in manufacturing in 1970 stood at 4.1 percent for the United States and 2.3 percent for Japan; they declined steadily thereafter in Japan, falling to 1.4 percent by the late 1970s, but held constant in the United States. In the case of Sweden, the monthly separation rates for private sector industry stood at 3.2 percent in 1970, but fell to 1.9 percent by 1976 (Statistiska Central-byrån 1971; 1978). This puts Sweden squarely between the United States and Japan in 1970 and closer to Japan by the late 1970s.

These data do not support the belief that labor shortage provides a guide to explaining the emergence of small-group activity. If anything, the above figures suggest that U.S. managers had the greatest incentive and Japan the least incentive to innovate. This does not appear to have been the case, as reflected in the amount of diffusion that took place in the two countries.

We can also examine another measure said to be loosely associated with labor shortage, absenteeism. The assumption here is that under a labor shortage, workers can be rather cavalier about being absent from work, as they can always find another reasonable job. Yet here as well, we do not find much guidance if we compare absenteeism across the three nations (excluding annual paid holidays and national holidays). The absenteeism rates for employees in the nonagricultural sector are highest for Sweden, 13.8 percent (1978), and lowest for Japan, 2 percent (1977), with the U.S. figures standing at 3.5 percent (1978). Thus, even making ample allowances for differences in data sources, the data still suggest that Sweden would have had the greatest incentive to develop a new approach to employee commitment and Japan the least incentive. Again, this does not conform with the national experiences.

An approach focusing on the structural conditions leading to the adoption of small-group activities also requires that we understand the link that ties these structural conditions to the solutions arrived

at by individual decision makers. This scenario presumes decision makers develop a set of articulated interests and have an ability to identify problems and formulate solutions and the political will and power to implement solutions.

Still, it is not my intent to suggest that structural conditions are not basic precipitating factors in the emergence of small-group activities. They operate according to my model as changing environmental conditions, thereby leading managers to question the efficacy of existing organizational arrangements. Keeping in mind the Sandberg list, and looking for common underlying factors that would distinguish the early adopters (Japan and Sweden) from the late adopter (the United States), we do find some suggestive patterns.

A fine-grained analysis does suggest indeed that the perception of severe labor shortage did create the conditions for the emergence of small-group activity in Japan and Sweden. How, the reader may ask, can this be, since such an outcome is not suggested by the comparison of objective measures discussed above? Before turning to that issue, we should note a further complication. Decision makers tend to see themselves as responding to specific manifestations of labor shortage, such as lack of control over the labor force (rising turnover and absenteeism and loss of competitive strength reflected in poor quality and productivity), rather than a labor shortage per se. As we shall see, "motivation" is a peculiar mixture of the driving forces leading managers to reexamine existing practices (thereby creating an occasion for making decisions) and their strategic responses to a problem based on finding and choosing possible courses of actions. They formulate their objectives in the course of carrying out these activities, and may reformulate them over time as they come to reevaluate accomplishments. Viewed in this light, we can see that motivation is not an event but a string of decisions over time.^[1]

It is difficult to mobilize managers and employees in response to abstractions such as the labor shortage, so broader ideologies, involving the virtues of participation as leading to self-actualization, democratization, dignity, improved quality of work life, and so on,

[1] This perspective builds on the early observations of Herbert Simon (1965: 53-54).

are often brought into play. A good example of the themes just discussed is a statement in the Volvo company report of the draft 1984 OECD study, which starts out in straightforward fashion listing as aims of work organization development in Volvo plants over the past ten to fifteen years the following objectives:

Increasing production

Improving quality

Reducing absenteeism

Reducing the turnover of personnel

Facilitating recruiting

The report then goes on to conclude that all these objectives can be summarized as increasing efficiency and improving the quality of working life. Of particular interest is the way in which the labor-shortage issue gets converted into a quality-of-working-life issue. To pursue this matter, we need a more detailed understanding of the identification of labor shortage as a problem and how small-group activities came to be seen as one of the solutions. For this, let us turn to an examination of each of the three national experiences.

The Japanese Case

In the late 1960s and early 1970s Japanese employers found themselves faced with an increasingly tight labor market as the economy grew rapidly and the agricultural sector could no longer serve its historic function of providing surplus labor (Minami 1973). It became increasingly difficult for the major manufacturing firms to recruit those select employees they desired, and management came to believe that it was increasingly difficult to retain such recruits. (The evidence for the latter proposition is somewhat problematic.) Even though, as reported above, turnover remained low by Western standards, it occurred in the context of a severe shortage of new male school graduates, especially middle school graduates, at a time of rapid economic growth. Replacement was both difficult and costly because of the absence of a pool of workers willing to take the most disagreeable jobs in the manufacturing sector. This situation threatened future corporate growth prospects and market share by restricting the ability of firms to meet market demand.

These observations make clear the limitations of relying on simple aggregate measures of labor shortage to capture the precipitating factors leading to the adoption of small-group activities. A meaningful model of decision making must build not only on structural conditions but on expectations based on past experience and perceived future prospects.

Between 1970 and 1973 momentum began to build to reverse Japan's long-standing policy of relying exclusively on domestic labor and to admit foreign migrant labor. However, the oil shock of 1973 and the deflation of Japanese economic growth, along with the decision to increase offshore equity investments, resulted in the issue being shelved. Rising educational levels led meanwhile to an increasing proportion of workers who were reluctant to accept the least demanding jobs. The educational system was producing more and more high school graduates who had been led to expect white-collar jobs commensurate with their educational achievements. Instead, an increasing number were being assigned blue-collar jobs. Management was greatly concerned about these trends. Many predicted an increasingly militant labor force unwilling to be satisfied with menial jobs. Surveys reported that workers wanted jobs that would allow them to develop their abilities and talents, whereas in the past workers had given priority to job security. It may be that management overestimated the problem, egged on as it was by exaggerated media reports and predictions that unrest on campuses would spread to the shop floor.

By the late 1960s evidence that supported management fears began to surface. The labor shortage was intensified for just those firms in the manufacturing sector that had the most standardized and routinized jobs. Industries still characterized by hard physical work under trying conditions and those requiring routinized job performance had a good deal of difficulty recruiting and retaining their labor forces. The auto and steel industries experienced great difficulties and had trouble meeting their expanded production schedules. A survey of 1,579 establishments in the machinery and metal-manufacturing industry reports that of the new employees in spring 1969, 50 percent of both middle and high school graduates recruited had quit within a three-year period (Japan, Minis. try of Labour 1974: 72).

These circumstances constituted a major motivational force for Japanese management to search for solutions to deal with what had become identified as a significant problem. Although the QC circle movement had come into being in the early 1960s, circles became immensely more attractive to firms outside the original early adopters in the materials industry (steel and chemicals in particular). QC circles became a key managerial strategy for making firms more attractive to highly educated potential recruits and reducing the likelihood of turnover and labor unrest. Those firms and industries that had the greatest recruitment problems, such as the auto, machinery, and other assembly industries, took the lead in introducing small-group activities during this time frame. They were also, of course, the industries that were growing most rapidly during this period and absorbing more and more labor. The growth in QC circle registration at JUSE took its first sharp spurt upward during this period, 1966-69 (see fig. 1 in chapter 2).

In the interwar period, Japanese firms had carried on discussion and study practices among work teams (see Cole 1979). They thus had experience with small-group activities in their behavioral repertoire (tradition), albeit on a more authoritarian model. This experience made the selection of small-group activities a reasonable solution to their problems. However, it was hardly the only solution. At the same time that managers were introducing small-group activities, the pace of technological innovation accelerated. That is, companies adopted the standard management response to labor shortage—capital investment. Koshiro< Kazutoshi develops a more complex causal framework for explaining the emergence of small-group activity, stressing that labor shortage and pollution problems led to large-scale capital investment (Koshiro< 1983:70-86). The costs associated with this large-scale capital investment to mechanize the production process and modernize work organization were very high; they could only be recovered by the introduction of small-group activity that would maximize the efficiency of the new arrangements.

It is important to note that Japanese managers generally do not see capital investment as a solution that permits them to avoid dealing with issues of work organization. Rather, their experience has taught them that to get the full benefits of capital investments

in equipment, they must work to optimize the relationship between the new technology and human factors. They have not, however, embraced any formal way of representing these relationships such as that embodied in the sociotechnical framework adopted by many Swedish managers.

Still another approach to dealing with labor shortages developed in the late 1960s and 1970s; offshore equity investments in production facilities grew at a rapid pace as Japanese firms sought to break the bottleneck imposed by the shortage of labor, take advantage of lower costs, and meet the domestic content requirements of Southeast Asian nations. Small-group activities were thus only one of several measures designed to deal with the effects of the labor shortage, and some of the alternatives had multiple objectives.

Another strategy adopted in some industries, especially by auto manufacturers, was the recruitment of large numbers of agricultural workers as seasonal labor. A related and somewhat later strategy used in many industries involved the increasing use of part-time workers, especially women. Initially, this strategy developed to meet labor shortages, but after the mid 1970s it was increasingly used to provide greater flexibility to employers trying to limit their regular employment in the face of slower economic growth; part-time female employment more than doubled, from 700,000 in 1975 to 1,800,000 in 1982, rising from 2.9 percent of the labor force to 6 percent. Yet for the critical period in question, 1965-75, female participation in the labor force actually fell from 48.8 percent in 1965 to 44.8 percent in 1975. This primarily reflected the reduction of the labor force in the agricultural sector, where many female workers had traditionally been concentrated.

The slowdown imposed by the oil shock of 1973 and the subsequent recession in 1975 provided an external solution to the labor shortage. Small-group activities were thus only one of the answers to the labor shortage. Moreover, they were only partially successful in addressing it. A substantial case study literature focusing on the contributions of small-group activity to quality and productivity improvement developed, however, and many managers accepted that small-group activities were meeting their objectives even though they did relatively little to moderate the labor shortage. In a 1977 nationwide survey, in 61 percent of all establishments surveyed, management respondents reported that small-group ac-

tivity produced "good results." The larger the firm, the stronger the positive evaluation—83 percent of all establishments with over 5,000 employees reported good results (Japan, Ministry of Labour 1978: 168).

There were other problems relevant to the selection of small-group activities as a solution. Small-group activities were part of a corporate strategy designed to mobilize all resources to deal with heightened competitiveness in domestic and foreign markets. The growing internationalization of the Japanese economy, and the tremendous threat this posed for the survival of Japanese firms, was one of management's major themes in the 1960s. Better utilization of its human assets could make significant contributions to improved productivity and quality. With labor to a large extent a fixed cost under conditions of low turnover, it made sense to fully employ the abilities of that labor force. Small-group activities were a logical follow-up to the growing interest in improving the training and education of workers.

Finally, a quite important point, the Japanese had by the early 1960s identified decentralization as an important strategy for reinvigorating their large organizations, thereby improving quality and productivity. Movements to decentralize safety and maintenance responsibilities were followed by efforts to decentralize quality responsibilities, leading in the early 1960s to the birth of the quality circles. Kobayashi Shigeru's influential book Sony Revitalizes Its Employees stressed his efforts at Sony's Atsugi plant to increase employee responsibility at the level of the small work group (Kobayashi 1966). Efforts to decentralize and to build small-group activity through increasing employee responsibility were part of a common effort, with the same actors serving as sponsors. The attempt to improve competitive performance through decentralization and small-group activity was both a response to the labor shortage and an independent outcome of the decentralization movement.

Management devised a variety of responses to environmental changes (labor shortage, rising educational levels, public concern about pollution, and internationalization of the economy). The objective was to raise productivity and quality while bringing about better communication, teamwork, and motivation in the workplace. Much of the literature of the QC circle movement stresses the importance of not aiming directly for productivity and quality

Fig. 2.
Managerial Motivations for Introducing Small-Group Activities in Japan, 1960-80.

improvement, but rather letting them arise indirectly from the improvement in communication and teamwork among workgroups (Shiba 1983a: 17). Japanese managers, being no more saintly than other managers, have tended in practice, however, to stress the productivity and quality-improvement side at least as much as, if not more strongly than, teamwork, communication, and motivation.

The preceding discussion of management motivation in Japan is summarized schematically in figure 2, which depicts the driving forces, the various strategic responses, and the objectives of Japanese management. Some of the strategic responses complemented one another; others represented clear alternatives, at least with regard to particular production outputs.

The Swedish Case

The labor-supply situation was even more serious in Sweden. By the 1960s a severe labor shortage developed. Swedish workers were increasingly unwilling to take jobs characterized by routinized tasks and poor working conditions. A number of studies in the early 1960s showed that many of the large industrial enterprises and public enterprises on Sweden's west coast were losing half of their newly employed Swedish labor within some seven and a half months of hiring. Overall, it is estimated that annual labor turnover at the beginning of the 1960s was running about 35-50 percent of the total number of employed workers (Berglind and Rundblad 1978: 102-3).

Absenteeism also grew rapidly. Sick days per person were estimated to be running at 15.5 in the mid 1960s and rose to 19 days per person by the early 1970s (Jönsson 1981: 10). This aggravated a growing labor shortage. Among new entrants to the labor force between 1970 and 1976, it is estimated, almost every second person was simply replacing working hours lost owing to absenteeism (Jönsson 1979: 3).

At the same time, educational levels rose rapidly. In 1962 the government initiated the nine-year compulsory comprehensive school (ages seven to sixteen), raising the school attendance requirement two years over the old system. By 1968, moreover, 80 percent of sixteen-year-olds were continuing their education be-

yond comprehensive school. The growth in educational attainment led to new entrants to the labor force whose profile was dramatically different from that of the existing labor pool. Nowhere was the gap greater than in the manufacturing sector. Whereas 69 percent of manufacturing employees were graduates of primary schools in 1977, only 41 percent of the national pool of 20-24-year-olds fell into this category (Jönsson 1981:9). There developed a strong sense among employers that these new workers were not satisfied with existing working conditions and that heightened labor turnover and absenteeism were the result. Swedish employers sought to deal with these conditions through various approaches.

First, they welcomed more women into the labor market, a policy reinforced by a variety of government measures, including the establishment of a national network of day-care centers and labor-market training courses. The entire growth in the number of people in the labor market from 1965 to 1974, from 3.7 to 4 million, consisted of women, primarily married women. As a percentage of the labor force, women rose from 36 percent in 1968 to 42.5 percent by 1975 (Statistiska Centralbyrån 1978:56). Female participation in the labor force in Sweden rose from 46.6 percent in 1965 to 55.2 percent in 1975. These levels stand well above those recorded by the EEC nations, Japan, and the United States (U.S. Department of Labor 1983: 422).

Interestingly, these increases did not especially accrue to the manufacturing sector, in large part because the traditional industries employing female labor, such as textiles, were rapidly losing employment. Women accounted for 24-4 percent of all manufacturing employees in 1965, increasing only fractionally to 24.7 percent by 1975. However, in those growing sectors hardest hit by the labor shortage, such as auto manufacturing and steel, the percentage of women in the labor force grew significantly. It rose in auto manufacturing from 13.7 percent in 1965 to 18.4 percent by 1975, and in iron, steel, and metal work from 12 percent in 1965 to 16 percent by 1975 (Statistiska Centralbyrån 1978:71, 73).

A second strategy has involved an increase in the employment of aliens. The number of aliens taking jobs in Sweden increased steadily from 121,747 in 1962 to 221,925 in 1973 (Statistiska Central-byrån 1977b:257). The Scandinavian component of this migration has been approximately 65 percent, with the majority of foreign

workers coming from Finland. Nevertheless, the Finns have presented substantial adjustment problems; for those who have not grown up in Swedish-speaking areas of Finland, the language hurdle is considerable. Yugoslav immigrants have come to constitute 10 percent of the non-Scandinavian inflow.

In 1977 approximately half these migrants (61 percent in 1971) were concentrated in the mining, quarrying, and manufacturing industries, as opposed to only 25 percent of the Swedish labor force; migrant labor constitutes 10 percent of the total employment in this sector (Statistiska Centralbyrån 1978:195). Moreover, within manufacturing, the migrant workers were concentrated in just those industries characterized by the most routinized jobs and poorest working conditions (Statistiska Centralbyrån 1977a:86). In the early 1970s only half the workers in these manufacturing sectors were Swedish and one-fifth were women.

By the late 1960s Swedish manufacturing employers became increasingly concerned that reliance on female and foreign labor was not solving their problem. They still had difficulty finding enough workers to do the least desirable jobs and found themselves relying on the lowest-quality labor. Doubts began to increase about the wisdom of relying on an increasingly larger number of foreign workers. In addition, absenteeism and employee turnover continued at high levels, with the rates being even higher for foreign and female labor than for male Swedish workers. Annual turnover by 1970 was running at about SO percent in a number of large plants in metropolitan areas (Jönsson 1979:3). High unemployment and sick pay benefits made turnover and absenteeism relatively costless to employees.

As we saw with the Japanese, one strategy under these circumstances is to increase investment abroad, and there is indeed evidence of an increase of such investment in the early 1970s. Swedish assets in overseas manufacturing companies were characterized by high levels of employment, reflecting the concentration in transportation equipment and machinery. The pattern was for large Swedish manufacturing companies such as Volvo, Ericsson, and ASEA to establish overseas subsidiaries for parts manufacturing and assembly operations (Jones 1976:172-73).

Yet Swedish firms were hardly ready to abandon production in Sweden. Consequently, attention turned to strategies for restruc-

turing work so that Swedish workers could be brought back into the factories and turnover and absenteeism reduced to manageable proportions. This restructuring of work had a number of dimensions, including the introduction of new technology and changes in the architecture of work organization. Small-group activities were an integral part of this response.

As in Japan, major innovations in small-group activities took place in just those industries suffering the severest recruitment, turnover, and absenteeism problems. Within Sweden, activities at Saab-Scania's engine plant and Atlas Copco's machinery assembly plant received great attention. Volvo experienced some of the severest problems and achieved fame internationally as one of the more adventurous firms in designing small-group activities (for a friendly assessment of its experience, see Jönsson 1979). Volvo's leadership role in Sweden needs to be fully recognized; its dominant position resulted from Volvo's status as the largest private firm in the Swedish economy. GM was the largest private firm in the United States until 1974, but it did not have the commanding role in the U.S. economy that Volvo did in Sweden. One estimate puts the impact of Volvo's operations on the Swedish economy at about twice that of GM's impact on the U.S. economy (Jönsson 1981:2).^[2]

The Volvo company report for the 1984 OECD study summarizes its strategic responses from about 1975 as follows:

the introduction of new techniques;

modification of production systems/structures;

changed "architecture" of working conditions;

reformulation of tasks for supervision and production techniques;

modified arrangement of workers' tasks. (OECD 1984b: 13)

With regard to the last item, the report identifies a "package" of activities, including delegation of responsibilities, integration of jobs where other skills are required, job rotation, participation in decision making about the job, and changes in physical working conditions. Volvo characterizes these responses as addressing the

[2] This role has continued to grow. In 1985 the amount of capital associated with Volvo and its allied firms accounted for 20 percent of the capital of all firms listed on the Swedish stock market. See Per Afrell, "Fonder viskning mot Volvo-makt" (The [wage-earner] fund is a whisper compared to Volvo power), Dagens Nyheter , 10 February 1985, p. 10.

problem of integrating "hardware" with "software." That is, it asserts recognition of the need to properly prepare the "social" side of the organization for the introduction of technology. This represents a modified version of the original sociotechnical position, with its stress on changing both social and technical systems to better coordinate them. Small-group activity, while implicit in the above description, is also given a smaller place in the arsenal of company responses as compared to what it would have reported ten years earlier.

We see here a situation in which successive choices (recruiting women, recruiting aliens) were unsuccessfully applied to the labor-shortage problem until a new, "more attractive" choice, small-group activities, came along (cf. Cohen, March, and Olsen 1976). Rather than abandoning previous choices, however, employers simply added this new tactic to them in an effort to resolve the problem. As in the case of Japan, there were additional factors responsible for the identification of small-group activities as a solution. These factors were primarily political in nature and involved other interest groups. Such alleged motivating factors as the "postindustrial ethos" to which Cherns (1979) refers seem, however, to have operated more as legitimizing ideology than as causal factors in the adoption process.

A notable difference with Japan lies in the relation of the small-group activity movement to the decentralization movement in Sweden. As in Japan, a decentralization movement developed in Swedish industry in the early 1960s. It gradually gathered momentum over the next two decades, but was slow to merge with the small-group-activities movement. The focus for much of the early period was on the decentralization of functions such as marketing, and the advocates of decentralization, generally top executives, tended not to link their advocacy with the small-group-activities movement. Similarly, the sponsors of small-group activities, especially line managers, did not particularly connect their efforts with decentralization. Perhaps the gap was sharpest for union leaders (who assumed membership on boards of directors in 1973); they tended to oppose decentralization insofar as it might lead to independent subsidiaries that would limit employees' job transfer rights. At the same time, many firm-level-union leaders tended to be supportive of the small-group-activities movement.

It was not until the appearance in the late 1970s of the Swedish

Fig. 3.
Managerial Motivations for Introducing Small-Group Activities in Sweden, 1965-75.

Employers' Confederations new version of decentralization (based on what it saw as starting to happen in a number of leading firms) that a firm link could be established between the two movements (Agurén and Edgren 1979). With the publication of the results of its new factories (nya fabriker ) project, SAF endorsed the "little factory in the large factory" concept and stressed a vision of relatively autonomous "product shops." This decentralization solution provided a better fit with the small-group-activities movement, but arrived on the scene at a time when that movement had already lost much of its steam and national politics was focusing on other solutions.

One can only speculate whether the consequences would have been different in Sweden if events had followed the same course as in Japan, where the two movements—small-group activities and decentralization—developed hand in hand, with the result that they strongly reinforced each other. In Sweden, however, middle-level staff and line management (quality, maintenance, production managers) found it easier to resist the movement for small-group activities when it was not reinforced by a top management push toward decentralization. These middle-level personnel had ample motivation to do so, since the small-group-activities movement was seen as threatening many of their prerogatives.

The failure of these two solutions—small-group activity and de-centralization—to land on management's agenda at the same time in Sweden arose perhaps from the tumultuous political events of the 1960s. This reminds us of the serendipitous quality to problem-solving activity and the different effects resulting from different combinations of problems and solutions.

Managerial motivations in Sweden for adopting small-group activities are summarized in figure 3. There are striking similarities to the Japanese managerial motivations depicted in figure 2, as well as some significant differences because of political pressures in Sweden and the role of internationalization in Japan (Sweden already had a highly internationalized economy in the period under question). The turn to foreign labor and the relatively independent role of the decentralization movement in Sweden are also notable. While these differences had real impacts on the speed and shape of the new movement, one is still struck by the overall similarities. But such schemata tend to freeze motivation in time, when in fact it is constantly evolving. To deal with this, let us turn to an evaluation of efforts over time.

Evaluation and Goal Succession

As in the case of Japan, it is appropriate to ask what success the Swedes had in using small-group activities to address the various problems of turnover, recruitment, and absenteeism, as well as quality and productivity. The Swedish case is more interesting in this regard because of the changes that did and did not take place. Employee turnover has declined dramatically over the past decade. From the range of 35-50 percent in the 1960s, national annual turnover figures for production workers show a decline to the low 20s by the late 1970s (Statistiska Centralbyrån 1981: 74). At the Volvo group (in which car manufacturing accounted for only 20,000 jobs out of a total of 76,000 in 1983), annual employee turnover fell from 20 percent in 1973 to 9.6 percent in 1981. Much of this decline can be attributed to a sharp weakening of the economy and the resultant tightening of the labor market. Thus, again, as in the Japanese case, small-group activities were at best only modestly effective in reducing employee turnover. Yet because the decline in turnover paralleled the introduction of small-group activities, and because it is impossible to separate the relative causal impacts of the different factors, it is possible for managers to proceed with the unchallenged assumption that some of the decline was attributable to the new work structures. This provides continued motivation and legitimation for pursuing small-group activities.

Some modest success has also been achieved in increasing the proportion of Swedish workers in the manufacturing sector. While the mining, quarrying, and manufacturing sectors accounted for 24 percent of the employment of all Swedish citizens (including naturalized citizens) in 1981 (down slightly from 25 percent in 1977), the proportion of foreign residents accounted for by these sectors fell from almost 50 percent in 1977 to 40 percent in 1981 (Statistiska Centralbyrån 1981: 189). In the case of the Volvo group, the percentage of Swedish workers increased from 69.7 percent in 1978 to 73 percent in 1981 (of the 27 percent consisting of foreign residents, Finns accounted for 17.2 percent and Yugoslavs for

4.7 percent). At Torslanda, Volvo's largest plant, which employed some 10,000 workers in 1984, the percentage of non-Swedes fell from 50 percent in 1974 to 35 percent in 1984 (OECD 1984b: 18).

The weakened economy was primarily responsible for a dra-

matic reduction in immigration to Sweden during the latter part of the 1970s and the early 1980s. The weakened economy and rising unemployment, the latter still low by American standards, led some Swedes to take jobs that they might have scorned a decade earlier. Again we see that small-group activities had, at most, a modest influence on developments. Again, even if the trends could not be used as dramatic evidence of the success of the new work initiatives, at least they could not be used to criticize them. It should also be noted that foreign labor is much less of a problem in Sweden in the 1980s for still another reason; most of the older immigrants have now learned the language and acclimated themselves to the Swedish work environment.

We turn now to an assessment of the impact of small-group activities on absenteeism. Unlike the changes in turnover data or the data on the proportion of Swedes in the manufacturing sector, the absenteeism problem has grown dramatically, rather than abated, over the past decade. Experts generally attribute much of this to the further improvement of sick leave benefits. Moreover, the reduced opportunities for voluntary job changing may be channeled into higher absenteeism. Others see the growth in absenteeism more positively as a new welfare society benefit, contributing further to employee flexibility.

Temporary absenteeism for a full measured week, including illnesses, vacations, military service, service leave (for example, for pregnancy or educational purposes), and strike activity, as a percentage of the number employed, rose from 9.6 percent in 1965 to 11.6 in 1970, 14 percent in 1975 and 17.8 percent by 1980 (Statistiska Centralbyrån 1981:44). It has since moderated somewhat. A 1985 report states that on the average 10 percent of the work force are away from work every day on grounds of sickness.^[3]

Small-group activities were apparently powerless to arrest the growth of absenteeism because of improved sick leave benefits. In-

[3] "Nyttigt att vara sjuk ibland" (Useful to be sick sometimes), Dagens Nyheter , 24 February 1985, p. 7. It was with some amusement that I read a telex message from the Central Bureau of Statistics apologizing to staff at the Swedish Embassy in Washington for a month-long delay in sending data I had requested on absenteeism and labor turnover. The responsible official explained that owing to absenteeism resulting from sickness and the press of other work, the bureau had been unable to get the request processed earlier.

stead of being used as an argument against small-group activities, however, the growth in absenteeism has been used from the beginning to justify the use of group work methods! Group work has been seen as a strategy to avoid the disruption introduced by absenteeism under traditional work methods. The reasoning is that through teamwork based on broad skills and job rotation, one can avoid the disruption caused by the loss of any one worker. The more absenteeism you have, the more useful group work thus is, not in reducing absenteeism but in cushioning its impact. By contrast, Japanese managers have typically used employee teamwork in conjunction with a tight (no slack) production system to reduce absenteeism by mobilizing peer pressures against workers "who create hardship" for their workmates by not coming to work. Here in microcosm we see a display of relative managerial and worker power in the two societies.

Moreover, using group work to cope with absenteeism in Sweden has its limitations. At its showcase Kalmar plant, Volvo was forced in the late 1970s to add still another reserve pool of workers to its existing absentee pool "to resolve the problems arising from absenteeism, which affects the different work teams quite unevenly" (Efficiency and Participation Development Council 1984: 44).

What can we conclude from this assessment of the factors that provided much of the initial impetus to the small-group-activities movement? As we examine what has happened to turnover, recruitment, and absenteeism over time, we see that performance has been influenced by a variety of factors and allows for a variety of interpretations. While this does not allow proponents of small-group activities to decisively rout doubters and nonbelievers, it does, given continued top management support, allow them to pursue their efforts without being subject to severe criticism.

With such tenuous claims, however, new objectives for small-group activities are needed to sustain top management support. We can see this transition at Volvo. When Volvo initiated its Kalmar plant in 1974, Pehr Gyllenhammar, the corporations president, stressed that it would be a factory that, without any sacrifice of efficiency or the company's financial objectives, would give employees the opportunity to work in groups. By 1984, with heightened competition in the worldwide automobile market, the stress was no longer on small-group activities not leading to any disad-

vantage but rather that they should confer competitive advantages in quality and productivity.

When viewed over the perspective of a decade and a half, there is considerable ambiguity about how to interpret the results of introducing small-group activities. This returns us to one of the early themes in chapter 3. Ambiguity is pervasive in this process, and the simple notion that it is reduced through learning over time hardly does justice to the complexity of the matter. What rather happens for many managers and workers is that new practices may gradually become institutionalized as normal organizational responses without much thought being given to what the original results were supposed to be. In the process managers assign new objectives to small-group activities. In short, a process of goal succession has set in whereby the original objectives are supplanted by new ones.

The U.S. Case

To contrast the situation in Japan and Sweden with that in the United States is revealing. A large reserve pool of unemployed labor is available in the United States to fill the most disagreeable jobs. Although labor turnover is remarkably high by world standards, an ever-ready source of replacements is thus available. This constituted a major barrier to the introduction of small-group activities in the United States, particularly in a context where most domestic firms were insulated from foreign competition (see Wool 1973).

The differences are most obvious in a comparison of the automobile industries. Gyllenhammar recounts how he came to the United States in the early 1970s and found that some American automobile plants had turnover figures similar to those of Swedish car factories—that is, "at the worst, half the employees left every year." At General Motors, it is reported that between 1965 and 1969, absenteeism rose roughly 50 percent, turnover rates were up more than 70 percent, grievances rose 38 percent, and disciplinary layoffs rose more than 40 percent (Kanter 1983:315).

Gyllenhammar was astounded that American managers typically did not perceive this as a serious problem. Rather, the American managers were accustomed to greater mobility of workers and rea-

soned that they could train people and put them on simple jobs so quickly that the turnover figures did not matter, even if they did make extra planning necessary (Gyllenhammar 1977). In short, because of this reserve pool of labor, management tolerated high turnover and absenteeism rates and had only modest incentives to engage in searches for new ways of organizing work. Indeed, the assumption that they had large quantities of unskilled labor being processed through the firm at regular intervals to a significant extent determined the simplifying job design specifications produced by engineers. It led them to expect less, rather than more, from their blue-collar employees, the opposite direction to that being taken by the Japanese and Swedes.

A similar argument can be made about rising educational levels. Although a number of scholars pointed to the rising level of education on the part of the American labor force, most of the change related to the rapid rise in those going on to higher education. The concern was whether there would be enough jobs for them. This was not seen by managers as affecting the supply of labor for lower-level shop and office floor jobs, and few American managers therefore saw a need to restructure their work organizations.

In the important book Work in America , published in 1973 as the report of a special task force to the Secretary of Health, Education and Welfare, the authors endorse participative management as one of the solutions to American work ills. Interestingly, the primary treatment of education in the report is as a means (through retraining) of improving work in America rather than as a force requiring management to reorganize work. The authors also point out that raising the educational level (credentials) is not terribly helpful, since there is little correlation between educational achievement and job performance (U.S. Department of Health, Education and Welfare, Special Task Force 1973:121-52).

Notwithstanding the stabilization or decline in college enrollment rates for the young in the early 1970s, the overall educational attainment of the labor force has continued to grow simply as a function of the retirement of older employees and their replacement by younger, better-educated workers. This gradual process is not, however, of the kind that leads management to dramatic policy changes. A rapid shift in the numbers and quality of high school graduates could potentially have had such impact, but the percent-

age of seventeen-year-olds graduating from high school was virtually unchanged from 1965 to 1975. It stood at 76 percent in 1963-64 and actually fell below 75 percent in 1974-75 (Grant and Lind 1977:65). If anything, concern was being registered about a decline in the quality of American high school graduates. In sum, upward shifts in educational quality were not as apparent in the United States as in Japan and Sweden, nor did such changes as occurred combine with a shifting labor supply situation to produce a major management concern.

Female labor-force participation did rise significantly in the period between 1965 and 1975, from 39.3 to 46.3 percent. This was not, however, a response to labor shortages, and certainly not a well-articulated employer and government strategy, as was the case in Sweden.

For all practical purposes, small-group activities did not represent a choice for American managers during the 1960s and 1970s. The problems they recognized, the available solutions, and the focus of key policy makers combined to make small-group activities a non-issue. There were occasional managers and companies who for one reason or another took the initiative in establishing participative management practices, sometimes involving small-group activities. Companies like Donnelly Mirrors, Lincoln Electric, and Harman International Industries come to mind. These tended to be isolated examples, however, without much interest being shown in them by other corporations or the media. They were seen as mavericks rather than as trendsetters.

To be sure, some American managers faced problems of labor turnover, absenteeism, and rising educational levels similar to those in Sweden and Japan. In the case of automobiles, Gyllen-hammar of Volvo was apparently unaware that there were key U.S. managers concerned about the impact of turnover and absenteeism on quality. In the 1973 letter of understanding between the UAW and GM establishing joint QWL efforts, GM specifically gave reduced employee absenteeism and turnover as the benefits to be achieved by the corporation. In turn, the consumer was to benefit by improved quality.

This letter of understanding established the first national joint labor-management committee on QWL in a major corporation. While the initiative came very much from Irving Bluestone, the

then vice president of the UAW, the corporation had already been exploring "organizational development." Kanter (1983:331-47) describes the internal momentum building for participative management within GM in the late 1960s. What is striking in comparison to Japan and Sweden, however, is how long such discussions were confined to the corporate level. Despite a long developmental period, the degree of diffusion to GM plants, notwithstanding such showcase factories as Tarrytown, was extremely slow during the 1970s.

The 1970s were a decade in which internal and especially external pressures for change in the way American firms organized and managed themselves were building. Relatively few firms responded to these pressures; most continued patterns of behavior built up in the post-World War II environment of economic growth and expansion of world markets. Among business leaders generally, there was little sense of urgency, no conclusion that dramatic changes might be called for in how employees were organized and managed. Those who were concerned often had little support in the power centers of their corporations and had to move slowly to build a constituency. The environment that made such lackadaisical responses possible was changing rapidly, however, as one American industry after another became subject to intense international competition.

The pressure finally broke through and forced American firms into a decade of turmoil, experimentation, and testing of new approaches. Ernest Savoie, director of the Labor Relations Planning and Employment Office at Ford, reflecting on the experiences of Ford's eighty-six large facilities, made a statement that might have been echoed at other major corporations such as GM, General Electric, and Westinghouse. At a 1985 University of Michigan workshop, Savoie said that more had happened with respect to small-group activities and other forms of participation at Ford in the early 1980s than at all U.S. companies combined in the 1970s!

As American markets were being internationalized, the role of the United States as a leader in manufacturing was being eroded, with market share declining in industry after industry. With the growth in the strength of the dollar in the early 1980s, incentives increased for American producers to move toward offshore facilities and suppliers. To use the words of Jim Baker, executive vice presi-

dent of General Electric, in a widely quoted speech of May 1981, the choice was seen increasingly as one of "automate, emigrate, or evaporate."

With automation came increasing recognition of the need to use the new technology in the most effective fashion, and many companies concluded that this involved greater reliance on an intelligent labor force. Small-group activities were part of this effort. At the same time, firms instituted strong cost-cutting measures, including especially measures to trim domestic labor costs. Notwithstanding Baker's colorful depiction of the hard choices facing management, these choices were not necessarily mutually exclusive. Companies like Ford that were investing heavily in new technology and small-group activities were the very same firms that were actively exploring offshore facilities and suppliers. This is one reason why workers often mistrusted management's commitment to employee involvement. Even General Electric, so prominent in selling its automation services to other U.S. manufacturing firms, was actively encouraging its purchasing department to explore offshore suppliers.

Ironically, it was only with the remarkable success of the Japanese in competing in American export and domestic markets that a more general reevaluation of small-group activities began. Suddenly in the early 1980s managers and the media were looking for the key to Japanese success, and participatory work practices based on small-group activities were identified as part of the package. Consider the following statement, which appeared in a friend of the court brief filed with the National Labor Relations Board by the International Association of Quality Circles (IAQC), the major organizational exponent of quality circles in the United States:

Faced with . . . challenge from Japan, American businessmen began to examine the situation and study the phenomenon of Japan's rapid economic expansion. They quickly became aware of the fact that Japan's success was largely attributable to its national concern for, and an almost single-minded dedication toward, achieving superior quality and high levels of productivity in the workplace. Moreover, American businessmen learned that the approach most extensively utilized by the Japanese to achieve these goals had its roots in the United States, to wit, "Quality Circles"—a specific and unique type of formally structured system involving employee participation.

Fig. 4.
Managerial Motivations for Introducing Small-Group Activities in the United States, 1975-85.

Ignoring the question of the correctness of these statements on the origin of circles and the specific causes of Japanese success, such views came to be widely disseminated in the popular management literature during the early 1980s. Management's objective in introducing participatory work practices was to provide for greater commitment of workers to their jobs in ways that would improve quality and productivity. The belief that the Japanese were making better use of their human resources through such practices as small-group activities was widely shared. Japanese managers were also seen as having developed more constructive relationships with their unions, and this also served to spur various participatory practices in the United States.

Decentralization as an approach to increased productivity and quality arrived in full force on the American scene in the early 1980s (see Kanter 1983). The rallying call was for fewer layers of management and more line responsibility at lower levels. This was rather later than in Sweden and almost two decades later than in Japan. Moreover, the American approach to decentralization had a peculiarly American quality, with stress being placed on "uncorking entrepreneurial talents" within the firm, or "intrapreneurial behavior," as it came to be called in management jargon. While some scholars such as Rosabeth Kanter tied this focus to small-group activities, for others the linkage was weak or nonexistent.

Figure 4 summarizes my analysis of the forces driving U.S. small-group initiatives. There is a certain "compactness" about the process in the United States that is lacking in Japan and Sweden, where the driving forces and strategic responses were more varied. Labor shortage in all its various manifestations was not a driving force in the United States, and the strategic responses to reduced competitiveness in an internationalized economy were thus more focused. Also notable are the differences in emphasis on "people-building." Whereas Swedish managers were likely to talk about changing the social and physical environment of the workshop in keeping with the sociotechnical system, and the Japanese about the need to build teamwork and communication among work group members, Americans were more likely simply to stress worker satisfaction. Moreover, the link between "people-building" objectives and increased productivity and quality improvement tends to be less well

developed in the United States than in Japan and Sweden (Cole and Byosiere 1986:25-30). In conclusion, despite expectations based on cultural differences, my analysis of managerial objectives shows a good deal more similarity in this respect between Japan and Sweden than between Sweden and the United States.

Chapter Five
Bandwagons versus Dominoes

Although all fashion looks mobile and rebellious at times, its roots are surprisingly constant: to think or act for reasons of fashion in any given field is to support that field's established centers of power.
Kennedy Fraser, The Fashionable Mind

Thus far I have talked about a diffusion of small-group activities that is not wildly at variance with the rational model of decision making. That is, notwithstanding public rationalizations and reliance on the current visible and available solutions, the process is one in which managers identify problems of high turnover and absenteeism and resultant problems of quality and productivity. This, in turn, gives rise to a search for solutions and adoption of small-group activities. Moreover, the search for solutions is greatest in industries suffering the greatest problems.

This process may be characteristic of the early adopters in each country, but does it apply to all adopters? Early adopters account for only a small number of total adoptions. In any innovation process, some organizations or parts of organizations take the lead and others lag behind. Social scientists have documented the characteristic patterns associated with these developments (Griliches 1957).

A bandwagon effect typically develops as the process of diffusing new technology acquires momentum. This is well documented in terms of decisions within industries in which prominent firms adopt QC circles and are quickly followed by major competitors who fear

a loss of competitive advantage. In Japan, for example, quality-control circles achieved rapid growth in the mid 1980s in the brewery and construction industries. In each case, a pioneering firm (Suntory and Takenaka Komuten respectively) adopted QC circles and the other firms in the industry responded in rapid succession. Both are highly competitive oligopolistic industries in which any management innovation that promises competitive advantage is likely to be picked up quickly by peer firms.^[1] If the enthusiasm for the new ideas cannot be sustained, they will be labeled a fad. Those hostile to the innovation may anticipate such an outcome, possibly generating a self-fulfilling prophecy.

In Sweden a 1969 "hallelujah" conference featuring the noted advocate of semi-autonomous work groups Einar Thorsrud kicked off the bandwagon phase. In Japan it occurred somewhat earlier, and in the United States we seem to have entered this phase in the early 1980s, as testified by the treatment of the subject in business journals.^[2] The participative programs of well-known large manufacturing firms like General Motors, Westinghouse, and General Electric were written up widely in the business press, and seminars and workshops on small-group activities were conducted by and for many major firms. The emotional character of such publicity, the promotion effort, and often the sheer power of the promoters tend to push careful evaluation of results by potential adopters into the background. At this point, companies begin adopting participatory work practices because it is the "thing to do." The public message is that your ability to compete successfully will be impaired unless you adopt this magic new formula. Prestige accrues to the early adopters (cf. Meyer and Rowan 1977). They are held up in the media as models to emulate and large-scale visits from other companies take place. The more skillful, like Volvo, Saab, the Buick Motor Division of General Motors, Hewlett-Packard, Sony, and Toyota, will convert this into a public relations success, increasing their legitimacy both internally and externally. Edward Lawler (1986:230-31) comes to similar conclusions, pointing out that such publicity often precedes any firm evidence of success.

[1] The particular examples used here are documented in Lillrank 1988: 197.

[2] See, e.g., "The New Industrial Relations," Business Week , 11 May 1981, pp. 84-98.

NASA: Getting In on the Action

To more fully understand the extent to which the choice of small-group activities is separated from systematic search and evaluation procedures as the movement gains momentum, let us consider the experience of NASA. This is of interest because it allows us to investigate whether NASA's rationale for adopting participatory activity and the mode of diffusion differ from those of the mass-production blue-collar industries. Some 85 percent of the NASA budget goes to contractors (mostly in the private sector), and the impact of NASA policies on them is instructive.

As a government agency with a mission that is very much in the public eye, NASA has a strong interest in maintaining a good reputation for quality. Every failure or delay of a widely promoted space mission owing to quality defects is a source of national embarrassment, a demoralizing experience for NASA employees and suppliers, a threat to budget appropriations, and an invitation to political meddling, as well as a source of fear that the agency may lose launch customers (satellite owners) to NASA's European competitor, Arianespace. Last, but not least, failures or delays are a threat to individual careers, jobs, and contracts.^[3] The 1986 shuttle disaster brought many of these very factors into play.

It is thus not surprising that quality circles came to the attention of NASA, which is continually looking for ways to raise quality and reliability levels. The agency recognizes that its quality problems have many causes, and that there is therefore a good chance any particular quality-related solution will address at least one of them. Consequently, NASA scans the environment for solutions rather than problems and seeks to match any quality-related solution it finds with some relevant quality problem.^[4]

At the same time, NASA regards itself as an elite government agency performing a "sacred" national mission. At least until the Challenger accident, it also saw itself as practicing state-of-the-art management. There is a certain hubris associated with this world

[3] See Arlen Large, "Shuttle to Try Satellite Rescue in Display Aimed at Customers," Wall Street Journal , 6 November 1984, p. 2.

[4] James March (1981:569) arrives at these same conclusions from a theoretical perspective.

view that leads to a resistance to innovation. The mentality is, "If it's good, we must already be doing it."^[5]

Around 1978 some key NASA officials who had been exposed to quality-circle ideas encouraged Martin Marietta Aerospace, one of NASA's prime contractors for the manned shuttle program, to initiate circle activities. Concerned about morale problems in their own operations and not oblivious to the political benefits to be derived from cooperating with new initiatives from a major customer, officials at Martin Marietta agreed, and activities were begun at Martin Marietta's Michoud Division in 1979. There was considerable trial-and-error experimentation over the next four years. Michoud personnel developed a set of measures that indicated that the circles were making significant quality and productivity contributions, and it proved a clever political stroke to create "integrated teams" of on-site NASA officials and Martin Marietta middle managers. The participating NASA officials sent back positive reports on their experiences.

NASA hung back despite repeated urging by Martin Marietta that nothing would come of the effort unless the agency took the lead. You have to "practice what you preach" was the argument used. Independent of NASA headquarters, the Huntsville NASA facility began experimenting in this area as well. At the same time, officials from NASA headquarters began visiting private sector firms such as Westinghouse to examine their productivity and quality-improvement efforts. Circles were identified as a key element in these efforts. By 1983 NASA began advocating employee participation through quality circles.

First, NASA pushed its own operations to adopt circles, known as NETS (NASA Employee Teams). Headquarter briefings were held to get the message across. The NETS were to be part of the NASA Productivity Improvement and Quality Enhancement Program (PIQE) adopted in 1983. The agency funded Martin Marietta's Michoud Division to provide the training for personnel at its various installations. In April 1983 Michoud held a facilitator training course for personnel at NASA installations. A 1983 summary report was written up and distributed widely within NASA and

[5] Kanter (1983:347-51) identified a similar world view at General Motors, with an added dose of belief in organizational immortality.

among suppliers detailing the achievements of the employee-participation efforts at its headquarter operations and at each of its installations (NASA 1984a). Such reports are "show and tell operations" designed to put pressure on each installation to have some solid material to report that will make it look good next time around. The manager of a unit with a weak report would be on the spot. To give the reader a feel for such statements, here is the published report given by NASA headquarters Code M:

Code M established a NET in the Resources and Institutions Office, with participants ranging from GS 11-15, both supervisory and non-supervisory personnel. Two NET members were trained by Martin Marietta's Michoud Division's System Refinement Team (SRT). They, in turn, have trained the other team members in NET techniques. SRT training resource materials have been made available. A sizable list of problems have been generated. From the list, the consensus is that the top two problems are POP/budget techniques and staff communications. A second "unofficial" NET has been established and is comprised of the OSF secretarial and clerical work force. A secretarial retreat has been scheduled for early January that should encourage more team spirit, establish better communications, and provide an opportunity for the secretaries to complete several current projects. These projects should result in more consistency within each OSF organization in handling correspondence and personnel-related items.

In the External Tank Program system, refinement teams [quality circles] at the contractor continue to produce cost savings, reduce attrition, and enhance productivity.

(NASA 1984b)

We see here a situation in which, contrary to the rational model of decision making, a scanning of the environment for quality-related solutions led NASA to identify circles as one such solution and to experiment with applying it to its problems. The heads of each of its installations were not all equally convinced of the merits of this new approach, and we have seen one way in which pressures were brought to bear on these installations for adoption.

The rational model of decision making, which links the prior identification of problems with the subsequent search for solutions, breaks down even further when we examine NASA's activities vis-à-vis its suppliers. As NASA officials became convinced of the utility of participative management through the Michoud experience,

they began to encourage adoption of small problem-solving groups in meetings with their other prime shuttle contractors (for example, Rockwell International and Morton-Thiokol). Soon the word was out among suppliers that it was advisable to begin employee-participation activities through the adoption of quality circles.

The Michoud Division of Martin Marietta again became a major training center for NASA. This time an informal arrangement through a "technical directive" from NASA at the Marshall Space Flight Center at Huntsville provided the funds to teach suppliers how to organize quality circles. The technical directive from Huntsville was a strategy to avoid the inevitable delay that would have resulted if approval had been sought from the bureaucracy at the national level. Suppliers learned that funds could be inserted in their budgets to cover the costs of such training without reducing their profits . In the world of bureaucratic organizations, that is a very powerful message! Moreover, for those contractors working on a cost-plus-award-fee basis, a strong favorable evaluation of their circle efforts by NASA could lead to points being added to the overall evaluation that went into their award fee (that is, their profit). The message was very clear: NASA wanted its suppliers to adopt employee-participation activities. Moreover, with a large number of agency officials working on-site at the supplier firms themselves, NASA could make its views known quite clearly.

In September 1984 a two-day conference was held for all major NASA contractors on productivity and quality improvement; executives from 110 corporations attended. Employee-participation workshops were a prominent feature of the two-day activities, again a clear message as to what NASA expected of its suppliers. When one of the participants asked the NASA administrator at the public forum if quality circles were going to be mandated in the future, he said no. My conversations with a number of the participants afterwards made clear that such a disclaimer was to be taken with a grain of salt. While it would be "bad form" to publicly impose circles on contractors, there was little doubt among most participants as to the direction NASA expected them to take. The signal was there for all who could read it. Customers are continually sending out signals to their suppliers in all realms about appropriate business behavior. The more adept among the suppliers are skillful in getting an early reading and turning it into a competitive advantage.

The Auto Industry: The Word Is Out

The kind of process described above is neither unique to NASA nor confined to the public sector. We see the same two-step process in the case of the U.S. automotive industry. The first step involved leaders of the auto firms issuing a strong message to their various installations in the late 1970s that employee-participation activities were strategies that had been identified as effective in dealing with various problems relating to productivity, quality, and worker satisfaction. At Ford this came in the form of a direct letter from the president of the company on 5 November 1979 announcing the new policy of employee involvement. The policy letter was accompanied by a directive to all general managers and above informing them in unequivocal terms that they were expected to develop employee-involvement practices (see appendixes A and B). Management was informed that it would be expected to operate within a set of guidelines, one of which stated: "Methods of managing should encourage employe participation in identifying and solving work-related problems."^[6] In short, though it was not entirely clear to plant managers what form their employee-involvement program should take, it was clear that they had better "get cracking." Clearly, "circle-like activities" were high on the list of appropriate responses. The only option managers had was in choosing the particular form of employee involvement, and even that was subject to approval by the personnel and organization and labor relations departments.

General Motors made the same decision in 1979, with President Elliott Estes and James McDonald, then executive vice president for North American Operations, providing the sponsorship (Kanter 1983:340). GM took NASA's more gentlemanly approach. Every GM plant was "strongly encouraged" to launch a QWL program.

[6] As if to illustrate the power of directives at Ford, the company still insists on spelling employee with one final e , a practice begun by a directive from John Bugas in the mid 1940s! Bugas, an ex-FBI man, was brought into the company by Harry Bennett in 1944 and became a confidant of Henry Ford II's. As vice president of industrial relations, and the man in charge of plant newspapers and news releases, he is said to have insisted that the spelling of employee conform to current newspaper style. Still another story that Ford managers have told me is that an efficiency expert sold the idea to Bugas, claiming that it would save a lot of money in typing and paper costs. I am indebted to David Lewis of the University of Michigan School of Business Administration for much of this background.

The message, however, was no less clear than at Ford. QWL programs, including small-group activities, were now mandatory.

The second step, taken in the early 1980s, was the strong initiative by the automotive companies to begin working with their suppliers to improve quality and productivity. Suppliers were made to understand that quality improvement was a precondition for maintaining their customer's business and some version of small-group problem-solving activity was clearly part of the quality-improvement package. With the widely anticipated reduction in U.S. automotive suppliers gradually taking shape in the mid 1980s, it behooved every supplier to take such messages seriously. Those who could move rapidly saw themselves as developing a competitive edge.

It should not be thought that the processes described above are unique to the United States. In the mid 1960s Toyota, with its well-developed communication links with suppliers, became one of the first large Japanese companies to make the decision to introduce its quality-circle program to its suppliers. We can be quite sure that these highly dependent suppliers responded with great alacrity, alert to the political implications rather than out of any profound understanding of the intrinsic benefits to be derived from circle activity. As one student of Japanese QC circles puts it, referring to the Japanese manufacturing sector generally: "If the parent company executives believe in QC circles, the subcontractors believe too" (Lillrank 1988: 127).

Summary

We can see the bandwagon effect in both the diffusion of small-group activities to specific units within large firms and their spread to "captive" suppliers and contractors. But the popular image of the bandwagon effect as a kind of spontaneous combustion does not capture the full range of the process. Such spontaneity can apply among "peer" organizations, but is far less relevant when inter-organizational relationships are hierarchically structured. The popular imagery of the bandwagon has its counterpart in the social sciences, where some scholars see the process of diffusion as one of contagion. Borrowing from the epidemiological model, the new be-

havior spreads in response to contact among organizations and the attractiveness of the behavior or beliefs being imitated.^[7] In contrast, the spread of small-group activities from customers to suppliers involves the exercise of power on the part of customers. Power creates the opportunities for streams of problems, solutions, and actors to come together at a given time and place. For Williamson's transaction cost economizing to have any meaning in this situation, the role of power in influencing future income streams (for suppliers) must be given center stage in any trade-off analysis.

We see here less a "bandwagon effect" than a "domino effect." Each successive adoption adds pressure for subsequent adoptions. Recalcitrant parties or "slow learners" are brought into line by sanctions if necessary. Internal diagnosis of work-related problems and a search for solutions is replaced by the adoption of "the solution" for quite unrelated, though nevertheless powerful, reasons. The approach outlined here is quite consistent with what has been labeled the "emergent norm" approach to the spread of fads (Turner and Killian 1972:21-25). Adapting this perspective to the organizational level, I would say that organizations are recruited to fads through preexisting networks of interaction, structured by power relationships. These networks account for the selective clustering of adopters. Fads are accompanied by strong informal pressures (emergent norms) to participate and penalties for those reluctant to do so (see Miller 1985: 152). Whether these fads evolve into incipient or lasting social movements is a matter still to be addressed.

One consequence of this kind of process is that adopting units may have relatively little commitment to small-group activities per se. Many managers tend to see employee participation as just one more productivity program along the road—little more than a new set of acronyms. They respond to the political pressure to adopt quality circles with inflated reports of their achievements and hope that by the time it is obvious that not much has happened, their customer will be onto a new program. While this is not necessarily the case, the tendency to such attitudes is strong in situations where the innovation in question implies a delegation of power, thereby

[7] See James March's discussion of this tradition (March 1981).

threatening existing authority relationships, and when the parent company or agency has a history of short-term programs that produce no lasting effects.

Over time there is a decoupling of motivation and search from solutions. Instead of problems chasing solutions, as posited in the rational model of decision making, solutions begin to chase problems. The solution is publicly dangled before companies with the promise that it will address a wide range of problems. Or, as we saw in the preceding examples, it is forcefully presented to specific operations and suppliers, who recognize that it is politically expedient for them to adopt the new practice if they expect to have successful careers in the company or to maintain their contracts. The presumption, whether true or not, is that the proponents have searched and found the most suitable alternative. Contrary to the standard model of diffusion, we see that in the "bandwagon" or "domino effect" phases, discovery and transmission clearly precede the search.

Chapter Six
Search, Discovery, Transmission

To learn, one must be humble.
James Joyce, Ulysses

In this chapter I examine the sources of the ideas that formed the basis for small-group activities in Japan, Sweden, and the United States and discuss some of the transformations that took place in the course of their transmission. When I began investigating small-group activities in the three countries, I was soon struck by the importance of the flow of ideas across national borders. How this operated and its significance became the subject of further investigation. The chapter also examines the nature of the search process at different phases of the diffusion of small-group activities, which sheds further light on the various decision-making models in use.

Swedish Developments: The Paradox of Success

The ideas adopted by Swedish employers during the early 1970s can be traced to a small degree to the contributions of American behavioral science and particularly the work of Abraham Maslow, Douglas McGregor, and Rensis Likert (Swedish Employers' Confederation 1971). The central line of influence, however, derives from developments at the Tavistock Institute in England (Emery and Trist 1969). The Tavistock researchers were strongly influ-

enced by theories of group dynamics and particularly by the work of Kurt Lewin, a German Jew who came to the United States in 1932. Lewin did pioneering research into the determinants of group decision making, the productivity of groups, and the influence on individual behavior of participation in group discussions and decisions (Homans 1968: 259).

The first social science experiments directly ancestral to the early 1970 Swedish efforts to adopt self-steering groups were conducted by Tavistock researchers in the early 1950s. These efforts began in a study of English coal mines and involved systematic comparisons of different kinds of work organization and wage systems and their impact on work performance. They led the researchers to conclude that "composite" forms of work organization with a holistic approach that stressed group work and responsibility were superior to conventional approaches. The new work methods involved integration of tasks rather than the differentiation of tasks characteristic of conventional mining methods.

These Tavistock ideas emphasized the development of the organization as an open sociotechnical system, with work organization based on the interaction of social and technical factors (that is, behavioral and physical events). Work was to be seen as a sociotechnical whole. At the group level, this means that the whole work group and the entire activity cycle must be taken into consideration in the design of work.^[1] In coordinating the social and technical factors, the Tavistock researchers made the assumption that there are several degrees of freedom on the technical side, so that the form of technical organization that fits best with the social factors should be selected. This is in contrast to the traditional view that the technical side always determines the parameters for the social organization of work. The aim is to develop small work groups that maintain a high level of independence and autonomy. This involves highly group-oriented work processes and group decision making. It is expected that as a consequence jobs and learning possibilities will be enriched and individual responsibility increased. All this is to be achieved without any loss in productivity compared to conventional work organization.

[1] See Sandberg 1982 for a fuller discussion of the development of ideas on work organization at the Tavistock Institute from a Swedish perspective.

English employers were slow to adopt these ideas, but they were carried to Scandinavia—especially to Norway—by the charismatic and visionary Norwegian scholar Einar Thorsrud (see Cherns 1979). The ideas came to Sweden indirectly through Norway and directly through the work of scholars like Eric Rhenman (1964), a researcher associated with the Swedish Employers' Confederation, and Reine Hansson, who conducted studies of how work motivation is influenced by organization and wage systems in 1967-69 (see Sandberg 1982: 146-50).

Three of the key events that served as a transmission belt from Norway to Sweden were a visit to Norway in 1966 by a Swedish union and management team led by Reine Hansson, who was close to Thorsrud's research; the translation of Thorsrud's research into Swedish in 1969 under the auspices of a joint labor-management publication company (Thorsrud and Emery 1969); and the "hallelujah conference" held at the Museum of Modern Art in Stockholm in 1969 with Thorsrud as the featured speaker. The conference was sponsored by the technical department of the Swedish Employers' Confederation (SAF) and attended by officials of major companies, as well as by selected union officials. The original plan did not include inviting union representatives, but Thorsrud protested, and at the last minute some were invited.

As discussed earlier, Thorsrud's ideas were not as well accepted in Norway as they were in Sweden. This is commonly explained by the lack of relevant mass-production industries in Norway, middle-management opposition, too heavy a reliance on academic consultants, and the hostility of organized labor. The unions saw many of Thorsrud's ideas on direct shop floor participation as a threat to the centralized decision-making power of the labor movement (cf. Jenkins 1974). Swedish scholars often say that the ideas developed in England, were tried out in Norway, and achieved their widest diffusion in Sweden.

The sociotechnical approach is a package with far-reaching implications for a firm's organization, but has a diffuse quality, deriving from ambiguous operational goals and technology, that undoubtedly slowed the diffusion process. It was not a calibrated solution with known costs and outcomes of the kind management would have preferred. As Pehr Gyllenhammar, the president of Volvo, remarked in a 1984 speech at a Volvo workshop on produc-

tion technology and quality of working life: "Management wants balanced change where you know what will happen—where you can manage change, and therefore it is good."^[2]

As noted earlier, in addition to its diffuseness, the sociotechnical package has strong democratizing overtones. The books of Thorsrud and his collaborators typically focus on the themes of industrial democracy and democratizing the workplace. So important were these themes to the authors that the terms often appear in the titles of their works, which strongly challenged the traditional hierarchical control structure. The notion of autonomous work structures was explicitly linked to the freedom of workers from oppressive, arbitrary management. In the early days of the movement, Thorsrud was known in Sweden as "the foreman killer." We can see this challenge in the "psychosocial job design criteria" laid down by Thorsrud and his collaborators in the early 1960s (which assume that acceptable levels of income and job security have already been achieved). These criteria were expressed somewhat differently from publication to publication, but they typically included the following:

1. Freedom on the part of workers to make decisions about how to do their work

2. A meaningful set of tasks, offering some variety and some free space to develop the job over time

3. Opportunity for learning on the job and to continue learning on the basis of feedback of results and future needs

4. Freedom to give and receive help on the job and to establish mutual respect between people at work

5. Recognition and social respect outside the workplace for doing a useful job

6. Some form of desirable future at work, not only in the form of promotion

While a number of these criteria, seen in isolation from the others, could be supplied by management action, the first—freedom on the part of workers to make decisions about how to do their work—sets the tone for an approach to work design diametrically opposed

[2] From author's notes made at the time.

to the control system of traditional bureaucracy. These criteria were widely discussed in Sweden during the late 1960s and early 1970s (Swedish Employers' Confederation 1971; Törner 1976).

This initial discussion in Sweden was, however, followed by a gradual diminution of the democratic theme as the concept was applied by management. That is, there was a sanitizing of the originally radical concepts as management tailored the implementation process to its own needs. In the late 1960s and early 1970s the term democratization of the workplace was much in vogue, particularly among trade unionists and scholars and in the mass media. But as management took over control of the implementation process, the theme of democratization gradually diminished in importance, at least as applied to small-group activity on the shop and office floor. Nowhere is this seen more clearly than in the shift in the language describing what managers thought they were doing. The term autonomous work groups (literally "self-steering groups" in Swedish) fell rapidly into disuse among management in the early 1970s. Increasingly the term production groups came to be used. By 1974 the technical department of the Swedish Employers' Confederation could write in a summary report: "Autonomous groups became for a time somewhat of a fad—today this term is less common and one now speaks most often of production groups, which better describes what it deals with." By the late 1970s and early 1980s managers even replaced the term production groups and began to speak of "group work" or "group organization" as one part of the creation of "new factories" characterized by small independent production units, freeing employees from machine-pacing, jobs with more personal involvement, and fast and reliable production systems (Agurén and Edgren 1979).^[3]

Consistent with my observations about goal succession in chapter 4, concerns about democratization of the workplace had been replaced by a strong focus on improving productivity and quality.

[3] Interestingly, many Swedes continue to use the term autonomous work groups when writing or speaking to American audiences; if they used it in Sweden, however, they would be criticized by managers for being old-fashioned and out of step with the times. It is as if they find it too hard to explain the complicated events of the past decade to Americans and therefore look for an accepted shorthand that they know Americans will understand. This illustrates well the difficulties of cross-national communication, even with the best of intentions and language skills.

In 1984 a workshop was held for foreign and local experts at Volvo's Skövde engine plant to observe and discuss the current status of group work and new directions in work organization at the plant. At that workshop Rolf Lindholm, a major formulator of the Swedish Employers' Confederation's policy in the 1970s, commented to the plant staff: "When I talked with managers in the early 1970s, it was a concern to meet the needs of human beings that drove change forward, but today it has the smell of business."^[4]

Much of the discussion in the late 1970s centered around the kind of understanding that labor and management would reach regarding the implementation of the new co-determination law. But by 1984 the Dagens Nyheter , the leading Swedish newspaper, reported that unionists and management were no longer talking about co-determination but rather about "development," which they defined as "local solutions for local problems." Furthermore, "the new slogans, 'change' and 'development' belong together with service, efficiency, and all that is now called 'management.'"^[5]

What we see here is the conversion of a blossoming mass movement stressing democratic principles into a much narrower business tool. In the course of this transformation a good deal of the enthusiasm and motivation behind the original changes evaporated. The initial conversion of radical scholarly theories to practical management knowledge may well have been necessary to ensure managerial acceptance. But this narrowing seems, in dialectical fashion, to have limited the potential for more widespread diffusion and acceptance by employees and unions. When I met Thorsrud and Hansson in late 1984, not long before their deaths, they both expressed disappointment at the direction developments had taken. "Founders" often have a difficult time seeing their ideas evolve.

Japanese Developments: From Elitism to Mass Movement

In the case of Japan, there was initially a similar cross-national diffusion process, but the source of the ideas was primarily the United States. American behavioral scientists' research is well known in

[4] From author's notes made at the time.

[5] Anders Mellbourn, "MBL inte så dålig some sitt rykte" (MBL isn't as bad as its reputation), Dagens Nyheter , 29 April 1984, p. 16.

Japan as a result of the broad "management boom" of the postwar period in which, from the 1950s on, the ideas of American management in all fields, especially personnel administration, attained an exalted position in the eyes of Japanese management. It is not surprising that the Japanese were willing and eager to study American management techniques in the early postwar period, when the United States was unquestionably the world's most advanced industrial nation, in addition to being Japan's conqueror and occupying power. Generally, the Japanese were willing, independent of any objective confirmation, to make the assumption that American management techniques were the most advanced. Over time, however, they found that they had to adapt them to fit Japanese conditions.

In the area of small-group activities, as in Sweden, the research of McGregor, Maslow, and Likert, as well as of Chris Argyris and Frederick Herzberg, to name a few key individuals, is particularly well known in Japan. This is a function of the almost instantaneous translation of books and articles, the steady stream of Japanese students to the United States (many of them company employees doing graduate work), and invited lecture tours of American experts to Japan (cf. Kobayashi and Burke 1976). It would be rare to find the personnel head of a major Japanese firm who is not well versed in the various ideas of leading American scholars in this area. These ideas, combined with and adapted to indigenous values and practices, formed part of the foundation of Japanese small-group activities.

We can see these processes in the work and influence of Juji Misumi, a noted Japanese social psychologist who studied at the University of Michigan in the mid 1950s and came under the influence of a number of the students of Kurt Lewin (for example, Ronald Lippitt, Alvin Zander, and Dorwin Cartwright).^[6] Based on their understandings of group decision-making processes, he conducted research in Japan upon his return that led him and his coworkers to stress the importance of small-group discussions conducted prior to individual decision as a basis for group decisions. In 1963 he began work with a large transportation company, stressing

[6] Kurt Lewin may come about as close as anyone to being the intellectual godfather of industrial applications of small-group activities.

a series of small-group meetings to devise strategies for accident reduction.

These activities were quite successful and the company expanded them throughout the entire bus department in 1969. At the same time, it merged Misumi-derived programs with the zero-defect (ZD) program in vogue in Japan at that time (Misumi 1980). ZD was also a U.S.-originated program, begun as a motivational campaign for workers aiming at zero production defects by quality-control personnel at the Martin Company in connection with its contract for the U.S. Army's Pershing missile system in the early 1960s (Halpin 1966). The program was modified by the Japanese to stress small-group processes and in turn has subsequently merged with the quality-circle concept in the eyes of most Japanese managers. We see in microcosm here the process by which Japanese students of American social science and organizations incorporated their new ideas and modified them in the course of applying them to Japanese work organizations. We also see how they merged with the different streams of development arising from the technical quality-control discipline, as well as with traditional approaches to group activity in industry, as mentioned in chapter 4.

We can pursue the contribution of the quality-control discipline in Japan to the small-group-activities movement through an analysis of the evolution of quality circles. This will also give us a sense of the capacity of the Japanese to borrow and adapt Western organizational technology to their own needs. Quality-control circles are the essence of the Japanese approach to small-group activities, an example of the creative process of borrowing and adaptation in the personnel policies of large Japanese companies.

The Japanese Union of Scientists and Engineers (JUSE) was established in 1949 and became the focal point of efforts to introduce modern quality-control practices and, later, quality-control circles in Japan. An early postwar effort to improve quality was organized by U.S. occupation officials seeking to restore basic services such as telecommunications. They helped arrange for American statisticians to go to Japan and teach American wartime industrial standards to Japanese engineers and statisticians. Prominent in this early effort was a series of postwar lectures beginning in 1950 undertaken by William Edwards Deming to teach the Japanese sta-

tistical quality-control practices (for example, control charts and sampling inspection).

In 1954 the noted quality-control expert Joseph Juran arrived in Japan to give a series of lectures. He emphasized a newer orientation to quality control, stating that it must be an integral part of the management function and practiced throughout the firm. Armand Feigenbaum's book on quality control in 1960, with its emphasis on total quality control (TQC), was also influential (see Feigenbaum 1961).

These ideas spread rapidly in Japan from the mid 1950s through the early 1960s, but there were some critical innovations (adaptations) on the part of Japanese firms as they began to adopt them in their daily practice. In the Japanese reinterpretation, each and every person in the organizational hierarchy, from top management to rank-and-file employees, was to study quality-control concepts and take responsibility for their implementation. As Japanese managers began to teach foremen the concepts of quality control, the idea developed of creating workshop study groups composed of the foreman and his subordinates as a means of getting workers to take more responsibility for quality (JUSE 1971:9). In particular, these study groups were seen as a solution to the problem of how to make workers read the QC materials that were being prepared. Group activity would encourage those who were otherwise not so inclined to read.^[7]

Such developments in turn evolved into the quality-control circles of today. An examination of the first issue of The Workshop and QC , published in April 1962 by JUSE, reveals a conception of workshop QC activity still at something of a remove from the actual operation of quality circles today. The focus was on training fore-men how to work with employees to get the latter to accept and maintain work standards so that quality objectives would be met. Getting workers more involved in taking responsibility for setting and revising work standards was the strategy advocated (JUSE 1962). The circles developed as a rather spontaneous adaptive process when management began to extend the new ideas of quality control. Gradually, these ideas growing out of the quality move-

[7] Interview with Ishikawa Kaoru in 1977.

ment merged with ideas of decentralization of authority and group decision making.

Not only were all employees from top to bottom to be involved in taking responsibility for quality, but all branches of the organization, from marketing and design and purchasing down through sales, were also to do so in the new Japanese reinterpretation. Unlike Feigenbaum's narrow concept of TQC, which still left responsibility with the quality-control department, the broadened Japanese concept of TQC stressed that all departments had to take responsibility for quality control through coordinated action.

What happened in Japan was exactly the opposite of what took place in Sweden, where the democratic ideas of the Tavistock researchers were sanitized to make them acceptable. Instead, in Japan, the elitist American approach evolved into a management-led mass social movement. This is not to say that the Japanese adopted a Tavistock or American perspective on democratization. For the Japanese, democratization in this context meant mass participation more than it did expanded autonomy for individuals and groups of employees at the expense of the hierarchical control system.

Many of the Japanese innovations arose in incremental and spontaneous fashion as managers and workers took Western ideas to their logical conclusions in ways totally unanticipated by either the American experts or, often, even the Japanese leaders. Such was the case with quality circles.

Taking Western management strategies developed for use with management personnel and applying them to blue-collar workers has been a characteristic practice in postwar Japanese corporations. The evolution of quality circles thus fitted in with a more general borrowing pattern, reflecting egalitarian trends in the postwar period that broke down the barriers between white- and blue-collar workers (Shirai 1983: 117-43). The extension of monthly pay systems and bonuses to blue-collar workers under pressure from militant unionists in the early postwar period was one manifestation of this egalitarianism. Still another example is the transfer of American ideas on career development to Japanese firms. These ideas on the importance of developing careers for employees and associated policies were evolved in America and designed for application to

managerial employees. Characteristically, the Japanese borrowed many of these ideas and extended their application to blue-collar employees (Cole 1979:41, 172-73; Koike 1983:29-61).

The Role of Scholars

It is of particular significance that scholarly theories and ideas provide much of the foundation for the small-group activity movement that developed in the U.S. and especially in Sweden. In the three countries, the industrial relations scholars who most consistently ignored small-group activities were the Japanese (with some exceptions), at least until Japanese efforts acquired a worldwide reputation by the mid-to-late 1970s. Prior to that period, in characteristic fashion, the Japanese managers found that they could directly consult Western scholarly output as well as draw upon their own historical legacy. Swedish scholars were most directly involved as advisors to the government, management, and the trade unions in the formulation of policy. This seems to fit with the oft-noted penchant the Swedes have for relying on technical experts and seeing social problems as ones to be resolved primarily through the exercise of technical expertise. To paraphrase the noted Swedish social scientist Hans Zetterberg, the debate over how to proceed with participation at the workplace in Sweden tended to take on the character of a debate among rational experts, each side with its scholarly advocates, and the solutions proposed had the appearance of applied social science (Zetterberg 1984:85). Moreover, the academic fraternity in Scandinavia has generally been more closely associated with national political and economic processes than has been the case in other advanced industrial nations (Mead 1984: 12-13).

The United States provides a middle case, with scholars filtering some Western European ideas (semi-autonomous work groups become self-managing teams in the American context) and developing their own approaches to participation; the work of Rensis Likert and Edward Lawler comes immediately to mind as an example of the latter. At a 1985 workshop on participative work practices at the University of Michigan, I asked the some fifty members of the audience (mostly auto industry managers involved with participa-

tory efforts) who among them had heard of, or read, the work of Emery, Trist, or Thorsrud on the sociotechnical approach. Not one member of the audience raised his hand! Yet, undoubtedly, they were familiar with many of the ideas of these writers through the filtering process described above (see, for example, Lawler 1986). By and large, however, U.S. scholars have played only a modest role in mediating the flow of information about Japanese quality circles. American managers found that they could obtain access to information from U.S. consultants and later from Japanese firms through "transplants" and joint ventures.

Limiting Factors in the United States

To appreciate the different trends in the United States at the very time when the Japanese and Swedes were moving rapidly toward adopting small-group activities, let us turn to a consideration of the contributions of Charles Kepner and Benjamin Tregoe. The comparisons take on added meaning in the light of the aforementioned propensity of the Japanese to extend Western ideas developed for management personnel to blue-collar workers.

Kepner and Tregoe's approach to problem solving and decision making for individual managers was quite influential in American management circles in the 1960s and 1970s. At the time that the Japanese were developing quality circles, these researchers were crystallizing their technique for teaching individual managers how to solve problems systematically. In The Rational Manager , they claimed to have put 15,000 American managers through their courses (Kepner and Tregoe 1965).

An examination of their text leads to a number of interesting observations. First, the techniques espoused for identifying, prioritizing, and solving problems, though on an individual level, were not fundamentally so different from those being taught to blue-collar workers in Japan. In some respects, the Japanese techniques came to be more sophisticated! For example, the Japanese took another American development, value engineering, developed for engineers and used for assuring that the essential functions are provided at minimum overall cost, and introduced it for use among QC circle members. Many American managers would be ill-prepared to handle this technique. It is also a minor irony that in the late

1970s and early 1980s a number of major U.S. corporations explicitly absorbed some of the features of Kepner-Tregoe training into their quality-circle training materials.

In discussing the research of Rensis Likert, Kepner and Tregoe criticized his advocacy of group action and participative management for lacking problem-solving and decision-making techniques. In retrospect, we can see that what was required for the equivalent of small problem-solving groups at the shop and office floor to have emerged at this time was, first, for American managers to have combined Likert's ideas on group action with Kepner and Tregoe's ideas and, second, for them to have had the imagination to recognize that these ideas could apply to employees at all levels of the firm. The proprietary nature of the Kepner-Tregoe materials limited even other individuals from exploring those opportunities. More important, these ideas and solutions simply were not in the repertoire of practitioners. It was not until 1987 that Kepner and Tregoe announced in a promotional brochure a "totally new approach to Statistical Process Control which permits this powerful tool to be used not only by managers but other key shop floor workers and specifically develops trouble-shooting skills enabling them to participate in quality improvement opportunities." Even then, it was being limited to "key" shop floor workers.

We can see that the emphasis on the individual and on the decision-making authority of the manager blinded American managers and researchers at this time to the advantages of group action at the shop and office floor level. This is ironic; one of the typical criticisms of the Kepner-Tregoe training voiced by managers has been that once it is completed and you go back to your firm, there are no structured incentives for using the new approaches. This, of course, is exactly what the group framework of circles would have provided.

In the history of technology transfer—and we may think of these ideas on small-group activities as a kind of organizational technology—it is common that the inventors of a technology are not necessarily the commercializers. The jet engine, for example, was invented in England but successfully commercialized in the United States (Miller and Sawers 1970). The causes of this disjuncture form the basis for an interesting intellectual inquiry.

Why has American management been so slow in adopting the

ideas of American scholars in contrast to the interest shown in their work by foreign scholars and managers? Some explanations are suggested by the preceding analysis of the Kepner and Tregoe contributions and why they took the form they did.

Beyond that, Japan and Sweden are oriented to the outside world to an extent that is hard for Americans to understand. They perceive themselves as relying heavily on export industries to sustain their standards of living. Some one-third of Sweden's GNP is accounted for by exports. The Japanese widely believe that they are able to secure their national survival only by adding value to imported raw materials and then exporting the product. Both Japanese and Swedes thus believe that their success—indeed, their very survival as nations—depends on their ability to search out and absorb ideas from abroad rapidly and efficiently. In both countries, if a solution to a problem is not immediately at hand, it is second nature to management to look for solutions outside their national borders. To succeed in foreign markets, they have learned to be open to different cultures. In Japan this stems in large part from the "catchup" mentality that has dominated the thinking of industry and government officials over the past one hundred years. To catch up you had to be prepared to adopt the better ideas developed in the more advanced Western nations. Moreover, defeat by the United States in World War II and subsequent occupation by a foreign power for the first time in Japan's long, proud history had a profound humbling effect that is hard for foreigners to understand. In the case of Sweden, the humility necessary for strong receptivity to foreign ideas seems to derive more from being a relatively small nation. As one industry official explained to me: "Swedes believe that there is always something better somewhere else. It is inconceivable for them with only eight million people and such limited resources to think that they have arrived at the best solution to a given problem. Therefore, they are always looking for ideas from abroad." The net result is that for diverse reasons mapping their foreign environment in a systematic fashion is a well-institutionalized practice in both countries.

One must be careful not to overgeneralize these observations. For example, the Swedes have been decidedly slower than Americans in grasping the competitive implications of Japanese management practices as related to small-group activities. I would specu-

late that this is both because the Swedes have been more buffered from the strong winds of Japanese competition and because their ideas on democratization have prevented them from understanding what the Japanese have to offer.

Nothwithstanding, the situation of the United States is very different from the situations of Japan and Sweden. Until recently, Americans appeared confident of their own managerial abilities and technology and not very attuned to learning from abroad. Even in the area of hard technology, not to speak of organizational technology, American companies maintained few listening posts in Japan relative to the size of the Japanese effort in the United States. Indeed, even with all that has happened between 1970 and 1985, it can be said without the slightest fear of contradiction that U.S. monitoring of global developments in technology has been, and continues to be, woefully inadequate. Yet increasingly our firms operate in an environment where new developments in management and technology are occurring outside U.S. borders (cf. Row-and 1985:15). To take the example of organizational technology, we have seen how few American managers are directly aware of the contributions of Emery, Trist, and Thorsrud, and they are even less familiar with the Japanese gurus of small-group activity. With a vast domestic market, the United States cannot be said to be export-oriented to the extent of Japan or Sweden. American managers were accustomed to being on top, and until recently there was simply not the same incentive to learn from others.

Moreover, the ideas put forward by American behavioral scientists seemed to fit much better with existing organizational practices in Japan and with the prevailing Japanese managerial philosophy than they did in the United States. The Japanese scholar Takezawa Shin'ichi caught the flavor of this in the following remarks:

The behavioral science model of management, however, is not perceived as an antithesis of the organizational reality as it might be in the United States. Instead, Japanese managers tend to accept the model as an idealized goal which essentially lies in the same direction as their own behavioral orientation. Often, they are puzzled to find out that American management in practice fits the scientific management model far better than that of the behavioral sciences.

(Takezawa 1976:31)

This difference provides an explanation for why these ideas were so eagerly received in Japan and why, once combined with small-group activities, they led to a choice opportunity. Similar judgments could be made in comparing Swedish and American management responses.

In the United States these same ideas did not lead to a choice opportunity; rather, they were seen as threatening by many managers and union leaders. The prevailing adversary relationships between managers and workers and managers and unions constitute a formidable obstacle to the adoption of new ideas about organizing work in a cooperative fashion. The potential improvements in productivity obtainable through cultivation of worker's loyalty and cooperation have tended either to be seen as trivial or dismissed as unachievable. In short, the gap between existing practices in American industry and the new managerial philosophies has been so great as to make search for and adoption of innovative solutions problematic. In the unionized manufacturing sector, even when top management has pushed in new directions, it has often been unable to have its ideas implemented. The everyday routines of individual workers, middle managers, and union officials seem grounded in struggles for power and control in a way that often defeats promising initiatives.

The upshot of all this is that small-group activities were not a serious part of American management's agenda of solutions in the 1960s and 1970s. Nor was the issue the province of any particular managerial level or department. It is no wonder then that Japanese and Swedish managers acted more quickly than American management in searching out, adopting, and transmitting ideas about small-group activities.

Chapter Seven
Borrowing and Culture

The crow
imitating the cormorant
drowns in the water.
Japanese proverb

In the preceding chapter, I dealt with the centrality of foreign ideas in inspiring the introduction of small-group activities. I have yet to consider explicitly the role of culture in the transmission of ideas cross-nationally. It has often been argued that cultural differences inhibit the adoption of foreign modes of work organization, especially in the United States. In particular, some argue that the small-group activities developed in Japan are so closely wedded to their cultural origins as to be inapplicable to Western firms. We need to get a clear understanding of the ways in which culture is and is not a barrier to cross-national borrowing. My analysis leads me to the conclusion that culture is a barrier, but not necessarily in ways that are commonly recognized. The cross-national transmission of ideas and practices such as small-group activities is, in fact, a process of technology transfer, not fundamentally very different from the transmission of, say, steel-making technology.

In his pathbreaking study of industrialization Modern Economic Growth (1966), Simon Kuznets maintains that the increase in the stock of useful knowledge and the application of this knowledge are the essence of modern economic growth.^[1] This increase, he argues,

[1] This section draws on Cole 1980.

in turn rests on some combination of the growing application of science to problems of economic production and changes in individual attitudes and institutional arrangements that allow for the release of these technological innovations. As industrialization spread through the world, technological and social innovations cropped up in various centers of development. These innovations were the outcome of a cumulative testing process by which some forms emerged superior to others; each historical period gave rise to new methods and solutions. The economic growth of any given nation came to depend upon adoption of these innovations. Kuznets concludes by stressing the importance of the "worldwide validity and transmissibility of modern additions to knowledge, the transnational character of this stock of knowledge, and the dependence on it of any single nation in the course of its modern economic growth." Raymond Vernon has gone on to argue that the costs to nations of assimilating information from foreign countries have been greatly reduced. The propensity of technology to cross national boundaries is growing very fast, based on word of mouth, technical journals, patent applications, scientific meetings, licensing agreements, and communications among affiliates in multinational networks (Vernon 1986:100).

This book deals with the borrowing and adaptation of small-group activities, a social innovation. Although Kuznets speaks of both technological and social knowledge, his reasoning would appear at first glance to apply most forcefully to the realm of technological choice. It is here, one might think, that the selection of the most progressive technique will be made unambiguously in terms of cost-benefit analysis. For example, the blast furnace using a hot blast and a mineral fuel adopted in nineteenth-century America was clearly superior in terms of reducing costs and increasing productivity to its predecessor based on charcoal technology.

Yet even with "hard technology," uncertainty as to the benefits of new technology is typically a prominent feature of the innovation process. To pursue a more recent steel-making example, it took some time before the feasibility of the basic oxygen furnace, and therefore its relative advantages over blast furnace technology, became apparent (Lynn 1982). One can make a similar point with regard to the need to adapt a given technology to specific envi-

ronmental conditions. Thus, the basic oxygen furnace, which was developed in Austria, depended in part for its success on the availability of special heat-resistant bricks used to line the converters. The raw materials for bricks with these qualities were not found outside Austria. It was not until comparable heat-resistant bricks were developed outside of Europe that the basic oxygen furnace became economically feasible in North America and Japan. I distill two points from these brief observations. The success of a technological transfer process depends on the adopter, first, having the appropriate "know-how" and, second, having the requisite resources.

Application to Social Technology

With social knowledge and institutional arrangements, the situation is similar. Again, under the right conditions certain institutional arrangements are fairly rapidly grasped as being essential to economic progress. Consider the spread of the joint stock company, double-entry bookkeeping, and multidivisional decentralized management structures.

Many other institutional innovations, however, are not easily compared and evaluated vis-à-vis existing arrangements. Social innovations often interact with other processes in a way that obscures their respective contributions to economic growth. This was apparent in the ambiguous results of assessing the Swedish experience with small-group activities relative to the original managerial objectives for instituting them, Furthermore, the output of social innovations is often not as easily quantified as it is in the case of hard technology. Social innovations like small-group activities may have multiple goals, of which contribution to economic success is only one. Some advocates of small-group activities, for example, focus on their contributions to industrial democracy and worker satisfaction as ends in themselves.

The complexity and the lack of clarity in these relationships and the abundance of unwarranted inferences both lead to an element of fad in the adoption of social innovations. As a result, arguments are often grounded more in ideology and power relationships than in tested generalizations. It should be pointed out that such situations are by no means absent from the circumstances surrounding

the adoption of hard technology. In the course of evolving the concept of "satisficing," Herbert Simon and his colleagues studied the reasons companies bought computers in the mid 1950s. "The biggest argument for computers was because your competitor had one. . .. It was just pretty hard to show up at lunch without one."^[2] The inference that Simon drew from these observations was that the forces that result in the acceptance of innovations are not necessarily economic ones. Rather, economic decisions based on rational evaluation of carefully sorted and weighted criteria were no match for the emotion and excitement triggered by a technical or business innovation. We can see similar processes at work today as vendors "hype" the advantages to be achieved by adopting the latest technology. In the General Electric advertisement in appendix C, there is the not so subtle suggestion that those not jumping on the bandwagon will be left on the scrap heap of failed enterprises.

Viewed in this light, it should hardly be surprising that rapid diffusion of a particular social innovation may reflect such social and political considerations more than the economic superiority of the innovation in question (as measured in expected profit or transaction costs). We saw good evidence of these same phenomena in our discussion of the domino effect among supplier firms adopting small-group activities. Sometimes the claims to superiority of one social arrangement over another are cloaked in the language of objective social science. The task, for example, of evaluating the applicability of Japanese management practices in the United States and judging what the needed adaptations are is a herculean one. Many claims are made, often by those with vested interests. How is one to separate the wheat from the chaff? Can one identify "best practice," and, if so, how can one most effectively diffuse it? I shall deal with these matters in subsequent sections of this chapter.

One final introductory note is in order. Despite the seemingly greater difficulty in evaluating the objective costs and benefits of a social innovation, there is every reason to believe that once a decision is made to borrow, know-how and resources are just as critical to successful transfer of social technology as with hard technology.

[2] J. Williams, "A Life Spent on One Problem," New York Times , 26 November 1978, p. F5.

Culture as Obstacle

Kuznets's model of the transnational stock of technological knowledge provides a basis for recasting the way we think about the role of culture in cross-national borrowing. If his model holds for social innovations with an impact on economic growth as well—and I argue that it does—this suggests that culture is less a barrier than it has sometimes been portrayed. Rather, dealing with culture is part of the normal process of adapting an innovation that grew out of and was shaped by a particular environment to the conditions required for survival in a new environment. Sometimes this effort is relatively easy; at others it is complex, with a high risk of failure. Sometimes the know-how and resources are there to do the job; at others they are harder to obtain.

Notwithstanding, the view that cultural differences create impassable barriers, especially in the case of the borrowing of organizational practices, has a powerful hold. My initial task is not to suggest that this view is wrong in principle, but to understand why such views hold sway particularly in the United States. We can pursue the matter in the context of the discussions in the United States on the suitability of adopting small-group activities developed in other nations. We shall also have the opportunity to compare this response to the responses of the Swedes and the Japanese to similar ideas developed in other nations.

In the early 1970s there was modest interest on the part of American managers in the Swedish innovation of autonomous work groups. The general conclusion, however, was that they were not suitable for the United States. Economic factors, in particular the small scale of production in Sweden, were typically cited as factors making the transfer of such techniques difficult, if not impossible. A broader argument was that the cultural environment of Swedish managers was so different as to inhibit useful learning. Swedish managers were thought to be "soft" when it came to dealing with workers and unions—a common view was held by conservative American managers that in "socialistic Sweden they had given away the store." There was the further perception that profit and efficiency considerations were not as highly ranked as in the United States.

On a different level, in a Ford Foundation report, six UAW workers sent to Sweden as part of a work experiment reported that the workplace culture was not to their liking. They found the cultural environment inhibited "self-expression" and that it was harder to communicate with union representatives and superiors. (One cannot help but wonder how much language problems entered into these evaluations.) The reporter for the trip concluded: "Work reform in America can proceed only in the context of American life and culture. The study of work reform in other countries can broaden knowledge, but it does not supply a blueprint for replication" (Goldmann 1976:42). The overall tone of the report was quite negative.

With the benefit of hindsight, these observations are amusing. Not ten years later, the president of General Motors' Saturn project, William Hoglund, lauded Volvo, saying that it was the "industry innovator in humanistic production methods and worker relations" (Hamilton 1985:2). He might have added that in its attempt to leapfrog Japanese leadership in the industry, GM was using many approaches pioneered in Sweden, such as modular construction, automatic carriers, and self-managing teams. In Spring 1988 Volvo and Saab held a high level presentation for GM on its work practices. Apparently, as suggested by the loose-coupling model, "Timing is everything."

Similarly, there was an extensive debate in the United States in the early 1980s about the suitability of adopting Japanese work practices in the United States. Many have voiced the view that the unique group culture and strong work ethic of the Japanese make the adoption of such practices unlikely to succeed in our more individualistic culture. To take one expression of this view, Prakash Sethi, Nobuaki Namiki, and Carl Swanson conclude in their book The False Promise of the Japanese Miracle that "a group orientation based on coercion and perceived against individual self-interest, either as a social norm or as a corporate goal, has little prospect of succeeding in the United States" (1984:244). Strong words indeed.

A somewhat similar debate was taking place in Sweden in the mid 1980s. As the Japanese "wave" hit Europe after having rolled over the United States—and therefore having been partially re-interpreted through American experience—the question of the suitability of people-oriented management strategies such as quality circles to the Swedish environment was much debated. Oppo-

nents argued that Japan's authoritarian work culture was not consistent with Sweden's democratic institutions. The unions sometimes labeled new management ideas on work practices "Japanese-inspired" in order to discredit them.

Culture as "Non-Issue"

Working against this perspective of culture as an obstacle have been the reported experiences of Japanese firms operating in the United States with American workers. The U.S. mass media have seemingly delighted in showing how the Japanese managers have been able to transport their "more humane management methods" to the United States. Japanese and American managers of these plants have been quick to exploit and facilitate this interpretation. Marvin Runyan, the former head of Nissan in Smyrna, Tennessee, has been perhaps the most adept in this regard. Yet, the evidence of social science on the subject is quite sparse (most such companies are not eager to allow serious social science assessment of their practices and their effectiveness) and mixed at best. In a 1984 survey conducted by Columbia Business School and Coopers and Lybrand, a major consulting and accounting firm, only 10 percent of the 150 responding firms (a 20 percent response rate) reported that they were using quality circles (American Productivity Center 1984:6).

In these and other areas, the Japanese have, in fact, been quite cautious in applying their practices in the United States. This indicates that despite their own eclectic approach to borrowing from the United States, they were quite uncertain as to the applicability of their methods to the American situation. Japanese managers in Japan were generally puzzled by the seeming enthusiasm of their American counterparts for Japanese management practices; their own sense of the "uniqueness" of the Japanese experience made them doubt the transportability of their methods. Notwithstanding, it became typical in the late 1970s and early 1980s in the United States to rebut claims that we could not borrow from the Japanese because of cultural differences by citing the experiences of the Japanese in the United States. By the mid and late 1980s the Japanese had become more confident of their ability to export their techniques for use with an American labor force, and the idea that

we could not borrow from the Japanese because of cultural differences weakened among both Japanese and American managers.

The Reaction Sets In

The initial enthusiasm for things Japanese in the late 1970s and early 1980s was based on the view that the Japanese had found a better way. Their methods were presented as a strategy of human-resource management for restoring productivity increases and quality, thereby reinvigorating our economy. William Ouchi's Theory Z and Richard Pascale and Anthony Athos's The Art of Japanese Management capitalized on and contributed to these views (Ouchi 1981; Pascale and Athos 1981). The 1979 NBC documentary If Japan Can, Why Can't We ? received enormous circulation, with many companies renting the film to show it to their employees.

As these ideas moved down to the office and shop floor, however, it was found that American workers and managers had little tolerance for being told that the Japanese did everything better and that we ought to learn from them. During the height of the deep recession of the early 1980s, such talk tended to induce a sense of helplessness on the part of many managers ("We just can't compete with those guys") and anger on the part of many workers. In this sense, the issue of cultural differences as an inhibiting factor reappeared.

As a result, advocates of small-group activities quickly moved to purge these ideas of any connection with Japan. We were told that Americans, not the Japanese, had really invented quality circles (just as the Russians have had a chauvinistic need to claim that most important inventions emanated from Russia and not the West). One common version is that William Edwards Deming taught the Japanese quality circles. The facts of the matter are quite to the contrary (Cole 1987). To Deming's credit, he has never claimed otherwise.

In another claim, quality circles are equated with the Scanlon Plan, developed by Joe Scanlon in the 1930s (Cartin 1981:550-53). The fact that Scanlon Plans typically do not have a problem-solving mechanism such as lies at the heart of the QC circle process is conveniently ignored. Still others stress the checkered history of labor-management cooperation in the United States as evidence that "quality circles are a contemporary form of . . . recurring

labor-management cooperation" (Mitchell 1987:56-66). Even if there were similar organizational inventions at some U.S. companies, and surely there have been such cases, the compulsion to stress these similarities is in itself revealing. One could just as well reflect on why these isolated examples did not crystallize into a broad pattern of institutionalized behavior.

Yet one seldom sees such analyses. Instead, often-belligerent declarations on the U.S. origins of Japanese organizational techniques testify to the strength of the view that foreign ideas rooted in different cultures are unsuited for use in the United States. One should see statements claiming a U.S. origin for QC circles in terms of the myth-building process required for American managers to be able to borrow the QC circle idea from Japan. It was easier for American managers to think in terms of "borrowing back" than it was for them directly to swallow borrowing from a foreign competitor. This was especially the case when the borrowing reflected so negatively on their past management style and when pride had been hurt by a former student who was suddenly "beating their pants off" in the marketplace.

Some leading U.S. quality-circle consultants have aggressively stated in public forums that we have nothing further to learn from the Japanese. The basic training manual of the International Association of Quality Circles (IAQC), the leading promoter of quality circles in the United States, does not mention Japan once. This was the result of a deliberate decision to eliminate any mention of the Japanese origins of circles (the materials were originally developed by General Electric for use in its facilities). Similarly, the standard information package used in 1984 by the IAQC does not once mention Japan. One of the brochures in the package does, however, include a list of companies in the Western world that have adopted circles.

Pragmatism and Borrowing

From a cultural point of view, it is important to note that while Japanese managers adapted foreign ideas to their own needs as they saw them, they felt little need to naturalize them for the sake of avoiding "foreign pollution." As we have seen, American management was held in high esteem, and any adaptation could thus

proceed on relatively pragmatic terms. The Japanese have, of course, a long history of borrowing ideas from foreigners, most notably from the Chinese in the premodern era and from the West over the past one hundred years as they sought to "catch up" with the advanced industrial nations. Based on these historical experiences, they have institutionalized a highly eclectic approach to learning.

My point is not the earlier one, that the Japanese (and Swedes) have a greater tendency to search out foreign ideas, but rather that, given equal response to a foreign idea, the Japanese and Swedes seem to have less need to naturalize it for reasons of national pride than do Americans.^[3] In the case of their quality-control efforts, the Japanese have been only too happy to give credit to their foreign mentors, William Edwards Deming and J. M. Juran, and to deliberately downplay their own accomplishments. The use of the well-known Deming Prize in quality to further spur company improvement efforts provides a case in point.

In contrast to the materials of the IAQC, the widely distributed handbook of the Japanese Union of Scientists and Engineers, QC Sakuru Koryo (General principles of the QC circle), makes immediate mention of the origin of quality circles in Western ideas of statistical quality control. It discusses how these ideas were then modified and adapted to the Japanese climate by the development of the concepts of "companywide quality control" and "total participation." The Japanese continue to send teams of quality-circle leaders and professionals to the United States to study American progress in this area. Similarly, the Swedes continue to study and send teams to research American approaches to sociotechnical design and participation, though an independent assessment would conclude that both the Japanese and the Swedes were well ahead of American efforts in these areas.

This pragmatic approach gives the Swedes and the Japanese an advantage over those whose foreign borrowing concentrates on pu-

[3] I hypothesize that this distinction applies more to ideas affecting existing social and power relationships than to more economic and technical ideas. The latter are often seen, at least initially, as less threatening. Thus, there was less resistance to borrowing "just in time" delivery (a "pull" system of production and inventory control developed by the Toyota Motor Co.) from the Japanese than quality circles.

rifying foreign influences. It would be hard to imagine the Americans adopting an "Ishikawa Prize" named after Ishikawa Kaoru, considered the father of quality circles in Japan, to reward the U.S. company with the most effective QC circle activities. Indeed, the National Quality Award signed into law by President Reagan in August 1987 is dedicated to Malcolm Baldridge, a former U.S. secretary of commerce. Beyond the initial visits by American executives to see how circles worked in Japan, there has been relatively little serious sustained investigation by American firms and consultants of how circles operate in Japan, how they fit into broader organizational practices, and how the Japanese have coped with motivational problems and ritualism. The exceptions to this statement are to be found mostly when viable partnerships have existed between American and Japanese companies, such as between Ford and Mazda and Fuji and Xerox.

In summary, the Japanese experience has not been used very effectively as a learning laboratory. For example, one would be unlikely to conclude from the popular U.S. literature on participation in Japan that a well-institutionalized labor-management consultation system is central to its operation. Even companies that have borrowed directly from this approach, such as Ford, feel a need to go to some lengths to conceal its Japanese origins: Ford calls its labor-management consultation program "mutual growth forums." It is not without interest to speculate why it is that Americans have generally showed so much interest in quality circles and so little interest in formal labor-management consultation systems. Certainly, a major factor is that U.S. labor unions have generally been unwilling to take responsibility for business decisions and be accountable to the membership for their decisions, while management is not willing to give up control. Circles, by contrast, are seen to constitute nice, safe, clearly defined packages. In a similar vein, relatively few American companies have explored the link in Japan between participation in small-group activities and job security.

I do not need to belabor the point further. Clearly Americans feel a strong need to indigenize quality circles. In so doing, managers minimize, if not eliminate, the Japanese origins of the concept. The danger is that this reduces their opportunities of learning from the "available stock of transnational resources." This, in

turn, becomes disadvantageous for competitive economic performance. Thus, contrary to the popular wisdom, a major cultural barrier to borrowing Japanese small-group-activities approaches is not their rootedness in group-oriented Japanese historical experience. Rather, it is our own reluctance to seriously examine foreign ideas and practices in the area of worker-manager relationships that inhibits the learning process.

The Importance of Local Invention

I have documented the way in which resistance to foreign ideas and practices has been a significant cultural barrier in the United States to the adoption of small-group activities based on the Swedish and Japanese models. This does not exhaust the cultural obstacles. Another major cultural barrier is the failure to recognize the critical need for participants to have a sense of "local invention" at the national, firm, plant or office, and workshop level. Here I deal primarily with the national level. This is a cultural issue because it rests on the degree of discretion those in authority are willing to give to subordinates to initiate policy. Ultimately such willingness rests on the degree of trust and confidence that those in authority have in their subordinates.

Why is local invention so important, and how are we to understand this term? For the responsible parties to be effectively motivated to adopt the innovation in question, they must have a sense of having contributed to its invention. This is especially the case where participation per se is the desired social arrangement. Pure copying of someone else's solutions does not contribute to a sense of ownership, which produces commitment to the new activity and is thus critical to the success of the borrowing. Moreover, local adaptation (that is, invention) is also central to success because each adopting unit is unique and has unique problems.

We would do well at this point to recall the earlier discussion of technology transfer. My stress on the importance of know-how and resources in securing technology transfer accords well with the way many social scientists have come to treat the concept of culture and tradition (cf. Eisenstadt 1973). That is, we may interpret the concepts of know-how and resources broadly to include symbols, ideology, and the generalized societal codes that provide a basis for

social action. They would, for example, include the kinds of authority relations and the kinds of interpersonal skills necessary for small-group activities to succeed. Culture provides a map of understandings that serves as a guide to social behavior. This map shows us how to proceed by outlining the options and how to value alternatives. Viewed in this light, culture becomes an obstacle to the extent that the adopter lacks the necessary know-how and resources to make the social technology work in its new environment. Know-how and resources constitute "local knowledge." This is the heart of the cultural barrier to borrowing.

The significance of this point of view is that contrary to the simplistic use of the term by many economists, there is, in principle, no such thing as diffusion of best practice. At best, there is only diffusion of best practices, practices that evolve in the course of their diffusion. Contrary to popular wisdom, there are times when it pays to reinvent the wheel! My stress on the importance of local invention grounded in local knowledge may appear less strange if we keep in mind that organizations themselves are social inventions designed to capitalize on the advantages of cooperation. Not surprisingly, changing them through "borrowing" thus involves further social invention. In many companies, top managers of highly centralized corporate structures have prescribed the form of small-group activity in a way that inhibits local invention. Each unit is instructed to do things in the same way, thereby reducing local spontaneous adaptation.

Devendra Sahal (1981:4), a sharp observer of the diffusion process, shows how significant innovation through adaptation to local needs occurs in the course of diffusion. He stresses that one of the problems with the contemporary theorizing of economists on the diffusion of technology is that it assumes that characteristics of the initial "innovation" do not change in the course of adoption. Many students of organizational behavior have, on the other hand, been quite sensitive to the way in which innovations are transformed as they move from invention to adoption to implementation (March 1981:569). As discussed later in this chapter, changes take place in the characteristics of the innovation as it diffuses, thereby affecting the rate of adoption. These observations allow me to append an earlier conclusion. While cultural barriers in terms of the historical rootedness of social ideas and practices may be less of a

problem in cross-national borrowing than commonly thought, the process of technology transfer is itself rather more problematic than commonly recognized.

Moreover, the additional innovation brought about by diffusion often creates new uses for the innovation. This, in turn, affects the extent of the adoption as well. In that context, I have mentioned both the development of the physical technology of heat-resistant bricks associated with the use of the basic oxygen blast furnace and the case of the social innovations involved in transforming Western ideas on quality control into the concept and practice of quality circles. By evolving into circles, formerly elitist quality-control ideas could be adopted by a whole sector of the labor force that had hitherto been excluded from involvement. By way of contrast, I have suggested how in Sweden managers saw the democratic ideas of the Tavistock researchers as threatening, and thus sought to purge them of their radical implications. While this facilitated initial acceptance by managers, thereby ensuring some significant diffusion, it also drained the movement of long-term, spontaneous support on the part of workers and the unions.

Some Examples of Adaptation in the United States

It is clear that adaptation is involved in fitting an innovation into the new environment in which it is expected to operate. This is true both of physical technology, where different factor endowments (different costs of land, capital, labor, etc.) press adopters to use the technology differently, and of social innovations. Perhaps the clearest example of this process in the material under consideration lies in the evolution of the name of this small-group activity in U.S. companies. In Japan QC sakuru (quality-control circle) is the standard term (though even in Japan companies may have their own terminology),^[4] reflecting above all the strong linkage of Japanese circles to the total quality control (TQC) movement, which stresses all-employee and all-department involvement in quality im-

[4] That they use Western terms for the activity is in itself quite extraordinary (though quite characteristic as well).

provement activities. In the United States many companies found that the name neither sat well with workers and unions nor conveyed sufficient breadth and importance. With regard to the former, the term control tended to have coercive connotations in the American context that were not what the promoters wanted to convey. As for the latter, quality-control departments often conveyed an image of narrow departmental and professional concerns in U.S. companies. Moreover, as noted earlier, they commonly had low status in the management pecking order. Quality circles thus came to replace quality-control circles as the preferred term. Moreover, as we have seen, companies adopted a wide variety of other names as well. In part, this was to avoid branding the technique as Japanese, but broadly speaking, it was part of the local invention process described above.

On a more substantive level, a variety of adaptations by U.S. firms can be noted. Whereas for the Japanese QC circles evolved in the course of quality-improvement efforts, in America they were introduced as part of the initial attempt to reinvigorate quality standards in many U.S. companies. QC circles also arrived in the context of an indigenous quality-of-work-life (QWL) movement in the United States, and the participation theme was thus quite naturally given high emphasis in many industries. American firms typically created the post of facilitator to apply human dynamics techniques. We have seen that the statistical component of the problem-solving skills has not been as prominent in U.S. efforts to date. Still another feature of U.S. methods relative to Japanese ones is the former's heavy reliance on consultants. For the moment, let us focus on yet another difference, the degree of voluntarism, to illustrate the tensions between different adaptive national responses and universal requirements.^[5]

U.S. firms acted initially to adapt the "all employee participation in circles approach" of the Japanese to fit a more voluntaristic model. There was much greater opportunity for employees to choose whether or not they wanted to participate in circle activity

[5] Wayne Rieker, the initiator of quality circles at Lockheed Air Missile and subsequently a leading consultant on quality circles, arrives at a rather similar list of adaptations (Rieker 1984).

(though not necessarily for managers) than was the case in Japan. The idea was that one would be working with truly motivated participants if participation was voluntary.^[6]

This stands in significant contrast to Japanese practices, where as we have seen typically over 90 percent of employees participate in circle activities. The Japanese do, in fact, stress the importance of voluntarism in making circles work in their literature, but they simultaneously stress all-employee participation. JUSE's handling of the matter in its handbook makes clear that it saw the reconciling of the two principles to be a matter of some difficulty. Consider the following tortured language:

Voluntary Participation by Everyone

Is it not a violation of the rule of "initiative," if some people, who [have] indicated that they do not like to participate, are still urged to participate? If it is clear that those people are reluctant to participate just because they are lazy, egotistical or conservative, they should be rightly encouraged to participate in QC Circle activities.

There are some people, however, who are individualistic by nature, and like to do their job without interference from others. . .. [But] the trend in the manufacturing workshop is such that no single person, however well he can perform his job, can fully satisfy the requirements now imposed on the workshop. It is no longer enough to "mind one's own business." The gathering of skill, technology and knowledge is a prerequisite of doing a job successfully in the workshop. Individual voluntary actions can be displayed on top of that.

Those who prefer working in solitude can be attracted by QC Circle activities, and participate on [sic ] their own will, if the circles can successfully create a magnetic atmosphere, and if those people are properly guided and persuaded to participate.

(JUSE 1970:55-56)

The different approaches taken by Japanese and American managers reflect differing norms and expectations, in part growing out of cultural differences. For the Japanese, voluntarism refers less to the individual employee's decision to participate in QC circle activity than to the relatively autonomous nature of peer decision making in circle activity, independent of normal line authority

[6] One of the few U.S. companies in the late 1970s that made circle participation obligatory in some of its units was Honeywell Corporation.

relationships (Lillrank 1985:5). At still another level, American unions have strongly stressed the voluntary principle in determining whether to support circle activity. Without such guarantees, they would have been far more hostile. Japanese unions, on the other hand, have shown relatively little interest in such concerns, except for protecting workers against compulsory overtime.

While Japanese managers use the rhetoric of voluntarism, it is apparent that in many companies strong management pressures, and peer pressures manipulated by management, make it difficult for workers to refuse to join in QC circle activity. The proper guidance and persuasion called for in the last sentence of the quotation above often takes the form of quite heavy-handed management pressures on workers to participate. Moreover, the autonomous nature of circle activities is often sacrificed through management manipulation of circle theme selection.

A 1975 Japan Employers' Association survey reported, for example, that of those enterprises involved in various types of small-group activities such as quality circles, 43 percent used performance in these activities as a factor in personnel assessment (cited in Tokunaga 1983:324). What employee wanting to get ahead in the corporation and impress management would refuse to participate under such circumstances? Here we see a good example of my earlier point about the subtle interaction between competition among individual workers and group activities.

Statements about differences in pressure applied by different management systems are, of course, relative. A leading promoter of the Japanese movement responded to my observations on this by commenting: "If you want to see pressure on workers to join circles, you should see the Koreans!" Over time, as the process has become institutionalized, the Japanese have moved in some respects toward a more voluntaristic model, especially in the selection of themes, quota requirements for suggestions, and actual participation in circle activities. The process is far from complete, however, with other pressures of institutionalization pushing toward ritualism and enforced conformity.

Indeed, in recent years new adopters of circles have taken a much more heavy-handed approach to getting employees to join circles. One Japanese observer refers to it as the "carpet bombing approach," alluding to the tactic in which planes are lined up in

tight formation and go back and forth over the target leaving no area uncovered. In a similar fashion, circles are targeted by top management and no employee is left with a choice. Typical is the case of the Sanwa Bank, which set off a wave of quality-circle activity in the banking industry in the early 1980s. The president, Akashi Toshio, was persuaded to attend a week-long JUSE seminar on QC circles in 1979. Ishikawa Kaoru was invited to give a series of lectures, and he stressed the importance of top management support. Akashi was so impressed that he put all his executive vice presidents through circle training. This process was repeated down the organization until no one was left unexposed. Within a year, there were 2,400 circles in the bank, with some 13,000 members. In my interview with bank officials at the time, they indicated that this constituted about 99.4 percent of eligible employees. All this stands in sharp contrast to the early days of the movement, which had much more of a bottom-up quality to it, with a strong emphasis on persuading workers to join of their own free will. In summary, we see a tension between the benefits resulting from ensuring that all employees in a workshop participate and the benefits resulting from spontaneous contributions.

In the light of this discussion, it is not surprising that by the mid 1980s many U.S. firms were beginning to reassess their voluntaristic approach.^[7] As one manager reported to me: "I have 15 percent of all my hourly employees in circles, but we seem to have hit the ceiling for volunteers and now I am worried about a loss of momentum." If a major objective is to build teamwork and communication in the workshop, then the typical American approach of having mixed workshops, with some in circles and some not in circles, seems unlikely to produce the desired outcome. By the mid 1980s, partly as a consequence of such concerns, firms were increasingly pushing mandatory involvement. As experience with circles and other forms of team building grows in the United States, we can expect more companies to take the position that teams are a part of normal job design, and that workers are therefore required to participate.

It is unlikely that the U.S. and Japanese approaches will con-

[7] See the "point-counterpoint" debate between Wayne Rieker and Harold Kay on this (Rieker and Kay 1985:6-7).

verge. Nevertheless, the pressures for differentiated responses to divergent organizational environments are balanced to some extent by pressures of a more universal nature. In both Japan and the United States, for example, we find managers struggling to balance the need to ensure exposure and teamwork against the need to develop and maintain spontaneity. Such universal imperatives will keep the two countries' efforts from diverging too sharply over the long run.

In conclusion, adaptation to local conditions, using local know-how and resources, is essential for successful technology transfer. By adapting to local conditions, one ensures the commitment and motivation necessary to drive the innovation forward. In the example of voluntarism, different cultural standards in Japan and the United States lead to different choices and opportunities as managers work with the resources available to them. At the same time, we see that attention must also be paid to certain universal factors, lest the borrowing process so bastardize the original innovation that its intrinsic benefits are lost.

The imperatives that derive from the universal aspects of a process have other implications for borrowing. This becomes clear when we examine the actual process of borrowing as it relates to quality circles in the United States and Japan. The experiences of Nissan in Smyrna, Tennessee, are a case in point. Prior to the initiation of QC circle activity in late 1983, the manager of organization development, Steve Neuman, was sent to Japan for training in the circle procedures and practices of the parent company. He came to realize that the activities in the parent company reflected a mature movement with some 4,000 circles and 99 percent participation in circle activities. Some circles had existed for as long as ten years, and it was not considered extraordinary for a circle to take a whole year to work out a solution to a complicated problem.

In one project the American manager observed, the circle was seeking to prevent wrong gaskets from being put on an engine. It turned out that this happened only three times a year (though with disastrous consequences for the operation of those three engines). He concluded that the problems the Japanese were working on were simply of a different nature than those that would be experienced by the Smyrna workers and managers for the first few years of QC circles. "They were working with a different part of the

curve." In contrast, he was wrestling with problems that the Japanese had dealt with over ten years before. How do you elicit middle management support? What is the best way to teach employees quality-circle procedures? How do you get employees to apply their training skills?

We can see from this concrete example both the difficulty of direct copying and the fact that the difficulty lies not only with cultural differences but with the differing imperatives arising from different stages of experience. Most American companies adopting QC circles did not have this direct pipeline to a Japanese company. The typical short-term visits of American executives hardly gave them time to figure out the right questions to ask, much less to get the right answers. What they generally see when they take their quick tour of Japanese plants is the result of some ten to twenty years' experience. Witnessing current practices does not provide much guidance as to how the Japanese got there.

Conclusions

We may return now to some of the original questions raised in this chapter on the role of culture and borrowing. Social scientists tend to operate with a holistic model; the suggestion that borrowing of social practices can take place on a selective basis tends to do violence to their perspective. The following quote from the 1984 report of MIT's International Automobile Program captures this view:

The impressive success of the Japanese has led many in the West to argue that their own industrial relations systems should be reshaped to imitate Japanese practices. The lifetime-employment system and the use of quality circles have been singled out for much praise and some emulation.

The difficulty with adopting individual Japanese practices is that the various features of an industrial-relations system interact with one another and may not be separable from other features of the total system. For instance, the method of setting pay interacts with the union representation structure (enterprise-based bonuses are supported by the system of enterprise-based unions in Japan) and also the form of worker participation. Even if they chose to, it may not be possible for Western producers and unions to adopt Japanese industrial-relations practices in a piecemeal fashion.

(Altshuler et al. 1984:215)

Underlying this argument is the view that if you borrow a discrete practice, you are pulling it out of a complicated and highly interdependent social nexus. In so doing, the further assumption is that the practice itself will not survive transplanting to the new environment. Interestingly, this viewpoint does not get applied to physical technology, which is assumed to be different from the transfer of social practices. Yet physical technology is ultimately made up of "know-how," and what is more social than that?

The thrust of the argument in this chapter is that the ability to borrow discrete bits of hard technology applies to social innovations as well. In both cases, the proper application of know-how and resources (local knowledge) can be brought to bear to adapt the practice or equipment in question to its new environment. The pattern of Japanese borrowing of American ideas of statistical quality control, leading eventually to circle activity, well demonstrates the possibility of borrowing highly discrete practices. It also reveals a process in which the original ideas or practices are literally ripped from their moorings to create something almost entirely different in the new environment. As paradoxical as it may seem to people accustomed to equating borrowing with imitative copying, borrowing is, above all, an act of social invention .

As if to demonstrate this process in the extreme, one Japanese scholar has spoken of "creative misunderstanding." What he has in mind is situations where because of language difficulties and cultural differences, the Japanese misunderstood what they saw, but found some way to borrow what they thought they were seeing and to make sense of the practice in their own environment. Lest this sound a bit too cute, we would do well to return to the insightful analysis of Thorstein Veblen, an astute observer of the process of German borrowing from England and Japanese borrowing from the West. Writing in 1915, this son of Norwegian immigrants stated:

The interposition of a linguistic frontier between the borrower and creditor communities would still farther lessen the chance of immaterial elements of culture being carried over in the transmission of technological knowledge. The borrowed elements of industrial efficiency would be stripped of their fringe of conventional inhibitions and waste, and the borrowing community would be in a position to use them with a freer hand and with a better chance of utilizing them to their full capacity, and also

with a better chance of improving on their use, turning them to new uses, and carrying the principles (habits of thought) involved in the borrowed items out, with unhampered insight into farther ramifications of technological proficiency. The borrowers are in a position of advantage, intellectually, in that the new expedient comes into their hands more nearly in the shape of a theoretical principle applicable under given physical conditions; rather than in the shape of traditional use, personal, magical, conventional.

(Veblen 1915:38)

While the language is somewhat old-fashioned, the thoughts are modern and applicable to my discussion of contemporary processes. The key to this whole process is adaptation, and it applies, as we have seen, to physical technology as well as to social innovations. That is, for effective borrowing to take place, one must make the kinds of adaptations that meet local needs. As we saw in the case of the United States, this may involve some purging of the foreign origins of the innovation, and it commonly involves creative interventions that change the characteristics and effectiveness of the practices in a dramatic fashion. This can be a very complicated process, as was the case with the Japanese reconceptualization of American ideas on quality control.

Commenting on the interdependency of borrowed elements and the dangers of "cherry picking," my colleague Michael Flynn observes that American managers "have been trying to model trees and haven't stepped back to look at the whole forest" (Holden 1986: 276). Flynn refers, of course, to the need to think systemically, and at first glance, his statement seems at variance with my analysis; his metaphor of the "whole forest" suggests the holistic model referred to earlier. But the question is, whose forest should we be looking at? It obviously can add important information to know how the original use of an organizational practice relates to other elements in the system, though Veblen cautions us that the borrower should not be so thoroughly imbued with such understanding as to lose the capacity to develop creative uses of the innovation. What is absolutely critical, however, is for the borrower to know how the intended use of the new practice in its new environment will relate to other elements in that environment. In other words, we need to be modeling "our" forest not "their" forest.

It may be that the term borrowing carries too much baggage. Especially in cases of extensive adaptation, where the original ideas are barely noticeable in the new practices, though they may have inspired them, or where foreign practices have served as a "negative role model," we might do better to speak simply of learning . It is clear that borrowing in the sense of mere replication simply does not exist for any complex social technology. Rather, there is adaptation and the invention of new social practices.