NOTICE: Please respect the preliminary nature of this draft: do not quote without permission of the author. ***** To Control and Inspire US Management in the Age of Computer Information Systems and Global Production* by Philip Kraft Department of Sociology and School of Management Binghamton University Binghamton, NY 13902-6000 (pkraft@bingsuns.cc.binghamton.edu) December 1995 *Forthcoming in Mark Wardell, Peter Meiksins, and Thomas Steiger (eds.), _Labor and Monopoly Capital in the Late Twentieth Century: The Braverman Legacy and Beyond_. SUNY Press (Albany, NY). I am indebted to the members of my graduate seminar on the politics of production. The comments of the editors have been especially helpful. As usual, I owe a great deal to my two unlikely but indispensable colleagues, Heinz K. Klein and Duane P. Truex. Please respect the preliminary nature of this draft: do not quote without permission of the author. wardel12. To Control and Inspire US Management in the Age of Computer Information Systems and Global Production I wonder if TQM is a strategy which is restricted to the internal functioning of a single company, or if it sees single companies as simply part of an economy which could be managed with total quality as its goal? I read about the futility of the us vs. them mentality within the functioning of a single company...would this also apply to the functioning of an economy? ================= TQM applied to a firm is capitalism. TQM applied to an industry is communism. See the difference? (Exchange on an Internet discussion group on TQM in Manufacturing and Services Industries)1 The worst thing an organization can do when it empowers people is to tell them to go out and make decisions. (Lynda Applegate, Harvard Business School, from a lecture on "Managing in an Information Age" at the Transforming Organizations with Information Technology conference, University of Michigan, Ann Arbor, Michigan, August 12, 1994.) Abstract Changes in the production process are always struggles over control. Ostensibly about efficiency and flexibility, the global expansion of capitalism is also about creating contingent employment and coercing cooperation. Relatively predictable markets and relatively stable employment have given way to constantly shifting "niche" markets served by "agile" enterprises. Subcontracting and outsourcing, once the last resort of failing enterprises, are now the marks of world-class firms which have mastered the art of "flexible specialization" and "dynamic stability." Contingency and coerced cooperation come together in the current obsession among management writers with "virtual" organizations, with integrated design and production systems, with Total Quality Management and Business Process Reengineering, and with sophisticated computer-based telecommunications systems. The lessons managers have learned -especially about the control possibilities inherent in selective recombinations of mental and manual labor -- should be of special interest to students of the labor process. Introduction Early in 1993 the Harvard Business Review carried a celebratory article about time-and-motion management methods in a US auto factory (Adler, 1993). The article is less important for its official rehabilitation of Frederick Winslow Taylor than it is for its singular standing: the Harvard Business Review had not published a full length article about traditional production work in years -- and it has not published one since. The Sloan Management Review, Administrative Science Quarterly, the California Management Review, and other trend-setting management journals have shown a similar lack of interest in what managers call "direct" work and workers. Production work and workers have been displaced from the pages of America's premier business press by the most important economic transformation since the end of World War Two, the emergence of global production and global markets. Managers have been quick to recognize the opportunities of globalization: lower wage bills, fewer union members, higher profits and, as a kind of bonus, rapid privatization of public services. There have been costs as well. US firms confront unprecedented levels of competition and the need to produce for constantly changing niche markets. US industry has had to learn how to make a profit by quickly turning out relatively small quantities of high quality products rather than large runs of standard products for protected domestic markets. If the global market confronts US managers with new competitive problems, computer-based technologies promise to provide them with some of the answers. New technologies, especially computer-based information technologies (IT), offer unprecedented opportunities to speed up data collection, analysis, and distribution. IT facilitate process integration, smooth workflows, and the use of unobtrusive embedded command-and-control systems. They therefore offer unprecedented opportunities to redefine work relations. Specifically, US managers have seized upon the control possibilities of IT as a way of satisfying their need to more systematically exploit labor on a global scale (cf. Stalk and Hout, 1990; Porter, 1985, 1990; Simons, 1995). When combined with telecommunications networks, computer-based information technologies offer the prospect of nearly instantaneous communication and therefore the means for continuous surveillance and direct control. All kinds of workers are easily monitored, not just traditional direct workers. Continuous feedback loops in manufacturing, monitoring phone conversations between service representatives and customers, automated testing of computer software and hardware, and automated validating of banking transactions can all be done cheaply and often invisibly twenty-four hours a day, nearly anywhere on earth. In design work, powerful computer-based systems offer managers positively irresistible opportunities. Designers and engineers can model entire finished products and production systems before a single physical prototype is built. Engineers, architects, systems designers, marketing specialists, and accountants can exchange, test, and modify ideas, plans, and images over a telecommunications network without ever meeting each other. So powerful are the new technologies that they have inspired management theorists and futurists to visions of postmodern "virtual" organizations. Enterprises will be held together by telecommunications systems while they adroitly enter, exploit, and then abandon markets over the globe. At the same time they can dispense with the impedimenta of conventional organizations, including buildings, permanent employees, unions, and middle managers. In short, the new computer-based technologies make the global market possible -- and immensely profitable.2 Globalizing Command and Control Global markets mean more than competition between firms. They create competition of a different kind. In a world of "agile enterprises," every agile enterprise is compelled to pursue shorter product development times, lower costs, and fewer defects. For US managers, globalization and computer-based technologies thus intensify an old problem. How can managers rapidly as well as efficiently appropriate the labor of "value adding" workers.3 Value adding workers directly contribute to the process of transforming partially completed products into deliverable commodities (Stalk and Hout, 1990). Those who do not add value, in contrast, only bookkeep or police goods-in-process. In contrast to transformational labor, James O'Connor (1973) has called this work "guard labor."4 It is therefore easy to understand why US management researchers so heavily stress the changing roles of a particular kind of transformational worker, designers and supervisors. The most pressing management problem left is how to make "creative" workers -- Robert Reich's (1991) "symbolic analysts" -- think faster.5 The centrality of "creative" labor is underscored by the pivotal role played by design workers in planning and implementing the computer-based information technologies themselves. Systems designers, programmers, computer hardware engineers, systems engineers, and telecommunications specialists are crucial to making production global. Their work is unique: it produces some of the most valuable commodities of the global economy -- computer programs and hardware, automatic switching equipment, marketing strategies -- which at the same time provide the means to regulate and control the market itself. The intimate connection between design, production, and control is rapidly forcing a fundamental change in the way managers conceive the production process and the nature of transformational labor itself. For US managers there are two important kinds of transformational workers. The first is traditional production workers broadly defined: extractive workers, fabricators, assemblers, data entry clerks, fast food preparers, and other "direct" workers. In some US high tech manufacturing firms such as microcomputer assembly, direct labor can represent as little as two or three percent of total product costs. Traditional production workers therefore present less of a control problem than they did even twenty years ago. There are fewer of them and those who remain are effectively disciplined by the global market itself, i.e., standardized and easily replicable tasks, worldwide sourcing, global labor pools, falling wages, and the insecurities of uncertain employment. With an essentially "union free" workplace and a large supply of legal "replacement" workers, US managers see the battle to control direct labor as essentially over.6 The second kind of transformational worker is Reich's (1991) "symbolic analysts." These are technical and administrative specialists who manipulate ideas or data rather than things.7 For Reich and many management theorists, the "creative" and "data driven" labor of design, engineering, marketing, coordination, communications, and customer service are no longer simply adjuncts to the global value-adding production process: increasingly they _are_ the production process. Coordination delays and conceptual bottlenecks in the work of these symbolic analysts -- the modern equivalent of Taylor's planning department -- are the principal remaining obstacles to the timely production of competitive goods. Eliminating such bottlenecks is the key to maximizing both "flexibility" and profits.8 The Limits of Taylorism: From Command and Control to Control and Inspire For nearly a century US management theory has honored a fundamental taylorist assumption: capitalist commodity production compels managers to spin out the processes of routinization and fragmentation as far as possible. Only those workers whose jobs ultimately resist routinization -- managers, designers, engineers, and other "conceptual" workers -- are permitted inside Taylor's "planning department." The rest check their brains at the door. Even when studies of unskilled workers show them routinely using their imaginations and tacit skills (and when other studies show conceptual labor routinely subjected to rationalization),9 such findings are declared exceptional. In practice the "American System" of production has always relied on pragmatic and constantly shifting combinations of direct and indirect control, of rationalized work and "responsible autonomy." The latter is reserved chiefly but not exclusively for design and administrative workers (Friedman, 1977). The relatively benign attitude towards design workers was possible because, more than most other US employees, engineers, designers, and managers could be expected to display admirable loyalty to the enterprise. When necessary, they could also be induced to heroically self-exploit.10 The most compelling reason to treat such workers gently, however, was neither technical nor ideological, but economic. Apart from some engineering-intensive industries, design and administrative work have been until recently a relatively minor cost of production in the US. Labor-intensive industries producing for protected mass markets-- a description which applied to most US economic sectors between 1918 and 1977-- generally could afford to absorb the costs of intellectual slack time. Employers now see a radically different marketplace. Efficient and rapid product development, intricate coordination, reliable communications systems, and quickly dealing with "exceptions" are the basis for getting competitive products to global markets in a timely way. Managers also see a dilemma: if design, coordination, and dealing with unanticipated disruptions, are the most important "value adding" work in the global production system, how can managers regularize and make predictable the work of employees who deal with "exceptions"?11 Even under the most stable and predictable conditions, taylorist command-and-control systems alone have never been adequate to manage these "creative" workers. The core of taylorism is functional and task decomposition and direct supervision. Taylorism "scientifically" fragments work processes and relentlessly removes "waste": wasted motions, wasted materials, wasted time, wasted people. It does so in order to maximize output and minimize costs, chiefly wages and materials. It is able to do this by invoking the formal right of the employer to control the physical behavior of the worker while on company time. Behavior is controlled by direct monitoring and by the swift and predictable application of carrots and sticks, e.g., variable piece rates, bonuses, and threats of discharge. It is the classic expression of what Gouldner (1954) called "punishment centered" control systems. Taylorism's chief shortcoming is that it can be applied to some aspects of design and administrative work but not to others. When employers try to extend taylorist control techniques to design workers and managers, the results are at best ambiguous.12 The problem is fundamental: managers seeking to extract the greatest value from "creative" workers need to manipulate not only behavior but imagination. They must inspire as well as control.13 But the techniques most favored by American employers to inspire imagination -- cheerleading and appeals to an uncertain professionalism -- have also been the least reliable. The results, even when useful, are short-lived. Cheerleading, motivational seminars, and tent revival inspirational meetings take time. If invoked too often they may also generate cynicism. Similarly, appeals to professional loyalty and "responsible autonomy" may backfire.14 Praising and rewarding engineers for their technical competence or managers for listening to the "voice of the customer" run the risk of divided loyalties. What, then, are managers to do? Management's earliest attempts to regularize imagination were thoroughly taylorist in spirit if not in the literal application of the "one right way." Managers attempted to graft work decomposition techniques onto operations research methods (OR) and management by objectives (MBO) systems. Neither combination was particularly satisfactory. OR is a modeling system for laying out and scheduling the individual tasks broken down by workflow designers. Operations research itself is therefore only an extension of the rationalist assumptions of taylorist fragmentation. It "scientifically" recombines the labor of individual workers whose work had been decomposed earlier by "scientific" work rationalization. It offers little in the way of facilitating design work itself. Similarly, MBO attempted to rationalize the decision-making process for managers and other administrative workers. To facilitate the ability of managers to think innovatively, MBO encouraged them to start with the desired long-term and intermediate outcomes rather than a standard set of administrative procedures and control techniques. MBO would help managers develop creative "problem solving" tools to achieve goals that were constantly changing. MBO, in other words, was the first widely used management system to lay out procedures and guidelines for making choices in ambiguous situations. Analytically, it was thus the opposite of taylorism, which was constrained by the one-size-fits-all practices mandated by Taylor's "one right way." If MBO in practice was a variant of taylorism, it was a reverse engineered taylorism.15 Although OR and MBO did not succeed in systematizing creative decision-making, they did prepare the way for a major departure in the way employers manages design and administrative workers. Operations Research and Management by Objectives were among the first attempts to systematize indirect command-and-control systems aimed at "creative" workers.16 MBO in particular showed managers they did not have to choose between either "hard" (measurement-based) control systems designed by industrial engineers or the uncertain effects of "responsible autonomy" and tent revivalism. Engineering-based measurement of behavior could be combined with carefully constructed "soft" systems of the kind March and Simon (1958) call "bounded rationality." The trick would be to find the right balance. From Rationalization to Obliteration The last two decades have produced a flood of command-and-control systems claiming to provide just that right balance. Among the most well known are continuous quality improvement, or just Continuous Improvement (CI), Total Quality Management -- the now ubiquitous TQM, commonly associated with Crosby, Deming, Ishikawa, and Juran -- and business process reegineering (BPR), whose most famous theorists are Hammer and Davenport. In the spirit of MBO, the goal of these "quality" and "process" movements is to provide managers with reliable ways to predict and control "outcomes" rather than measure and monitor behavior. Their common formal purpose is to reduce non-value-adding time, product defects, process defects, and costs. In this respect, they all owe large and obvious debts to taylorism. TQM and BPR have been enthusiastically adopted by US managers in part precisely because they legitimate and rejuvenate the traditional taylorist obsession with unnecessary motions and unnecessary people. Taylor's cry of "eliminate waste!" finds an echo in Deming's injunction to "shrink, shrink, shrink variation!" and Hammer's command to "obliterate" duplicative and non-value-adding work processes. Like taylorism, the process-centered command-and-control systems assume that all workplaces are systems that can be rationally designed, modified, and made routine, predictable, and efficient. Taylor's "eliminate unnecessary motions" is transformed into "continuous improvement." "Soldiering" is rooted out through kaizan and lean production or some other form of "management by stress" (Parker and Slaughter 1988; 1992; see also Wood, 1993). Their practice, if not their rhetoric, is top down. Like taylorism, all are concerned chiefly with getting fewer workers to produce more in less time and ultimately at lower costs. All rely heavily on techniques that operationalize work flows in order to measure outcomes ranging from time-to-market cycle times to deviations from specifications to customer satisfaction. Increasingly, they rely on a variety of sophisticated computer-based real-time, that is, continuous, measurement systems -- a sci-fi version of the time clock. Time-and-motion studies are supplemented with Statistical Process Control measurements made possible by cheap desktop computers and LAN's (local area networks). For US workers -- whether engaged in "creative" or "routine" work -- the consequences of the new work redesign systems are usually the same as those produced by taylorism: intensified labor, more job competition and job insecurity, and downward pressure on wages.17 The new command-and-control systems also differ from each other in some important ways. Continuous Improvement and Total Quality Management on the one hand and BPR on the other are the extremes of the new management systems. If not exactly the yin and yang of command-and-control, they are its Laurel and Hardy. CI and TQM are the more incremental approaches. They seek steady decreases in product and process variations, i.e., defects, and the constant evolution of "robust" design and production processes. BPR, at the other extreme, demands more fundamental results: a massive restructuring of process and work flows. BPR's goal is more than reducing defects and variations. Its goal is to do so through radical simplification and "downsizing." A reengineered workplace has fewer, not just better designed, tasks, fewer, not just more robust, processes and "multifunctional" -- as well as fewer -- workers. TQM and CI maintain an uneasy silence on speed up and job loss. BPR openly advocates stripping organizations down to their "core competencies" and "core processes" while aggressively outsourcing the rest. BPR is the proud godfather of lean and mean. The Collective vs. Individual Worker If CI, TQM, and BPR share with Taylorism a preference for "hard" command-and-control systems, they are not simply modern variations of work measurement and direct control. Crosby, Deming, Hammer, Ishikawa, and Juran share with Taylor the engineer's credo: measure everything. But CI, TQM, BPR and their variations go considerably beyond simple measurement, task decomposition, and speed-up. TQM and BPR are preoccupied with the design of the production system itself. With taylorism, they seek to maximize output and product quality at the lowest possible unit costs. Unlike taylorism, TQM and other process-centered strategies also seek to eliminate defects in the organization of the production process as a whole. In a radical departure from taylorism, these process-centered systems reject rigid distinctions between design and production and therefore rigid distinctions between design and production workers. On the contrary, CI, TQM, and BPR covet the ideas and tacit knowledge of all workers, including traditional direct workers. If taylorism is about separating mental and manual work and then constructing a rigid hierarchy to administer and police that separation, process-centered management systems are about systematically appropriating ideas and knowledge from all workers through a system Harvey (1990) calls "flexible accumulation." It may just as accurately be called flexible appropriation and flexible control. This is a fundamental difference between the old and new control systems and it warrants further explanation. Taylorist rationalization is linear, serial, and segmented. Taylorist control systems are concerned with three organizationally intertwined but for them conceptually separate and distinct problems. First, what is the best way of decomposing the making of a product? Second, what is the most efficient (that is, fewest, cheapest, and fastest) set of actions required to produce each fragment? Third, what is the best way of combining the fragments into a finished product? Furthermore, from the perspective of taylorist designers, both the finished product and the processes used to design it are givens. Product definition and development (the work of specifying, designing, prototyping, and testing a product) are separate and largely irrelevant activities. So are the design of the technology, tools, and processes used to make the product. More to the point, product design and the design of the production process are the responsibilities of other managers, engineers, or designers. Each fragmented part of the total process is carried out with little regard to what comes before or after it. Management critics of traditional serial organization call this system "over-the-wall" production. In contrast, work systems organized along the principles of continuous improvement, TQM, and BPR stress the identity between the design of the product and the design of the process required to make it. Whether steadily reducing defects or radically reengineering entire processes, managers and engineers must therefore have a tighter and more immediate control over process as well as product variations. Process-centered command-and-control techniques offer managers a way to do this. Over-the-wall design, production, inspection, testing, and service are replaced by "design for manufacturability" and "concurrent engineering." In the language of TQM, it is the difference between quality control on the one hand and "building quality in" on the other, i.e., inspecting finished products for defects rather than reducing variation and defects in the process itself. The shift in focus reflects the world-wide raising of production standards and the need to minimize product design, development, and production cycle times. The obsession with reducing process as well as product variation means that CI, TQM, and BPR must go considerably beyond wringing out "slack time" in direct production or getting design and administrative workers to think more quickly and more effectively. Traditional taylorism assumes that a finished product is the sum of individual tasks and individual "value adding" transformations. An individual's work is decomposed, recombined, and then inserted in a linear process populated by similar individual workers whose behavior is similarly fragmented and coordinated by supervisors. Taylorism, in other words, focuses on the labor process of the individual worker rather than the production process as a whole. The "worker" is an individual from an engineering as well as a social point of view. Hence, taylorism emphasizes piece rates and direct surveillance and control, even for design workers and managers. Control the behavior of the individual worker and you control the "efficiency" of the total process and the cost and quality of the product. Critics of taylorism (and of Braverman's analysis of taylorism) have always pointed out that in practice US managers routinely ignore the presumptions of radical individualism and hyperrationality. Experienced managers know that taylorism often backfires, particularly when crucial parts of the labor process rely on tacit knowledge and illicit coordination among workers. The focus on the individual labor process becomes positively self-defeating when global competition increases the demand for "flexibility," "quality," and speed. Global production requires the coordination and administration of diffuse, constantly changing,and even "virtual" organizations. Under these conditions taylorist hyperrationality can only _increase_ inefficiencies in design, development, production and distribution.18 What was the essence of rationality becomes the source of irrationality and, worse, loss of control. From Obliteration to Incorporation to Empowerment Global production therefore compels managers to conceive of production as a system rather than a collection of individual tasks. The new computer-based information technologies make it possible -- make it necessary -- to act on that assumption. This is precisely the strength of the new process-centered command-and-control systems. Juran, Deming, Ishikawa, Hammer, and Davenport19 -- each in some way Taylor's pupil -- have seized upon a truth familiar to both Roethlisberger and Dickson (1946) and Braverman, if not Taylor: labor may be individual, but production is collective. This new understanding has forced managers to make still another fundamental break with traditional taylorism. Taylorism is about individual workers and decomposed tasks. CI, TQM, and BPR are about collective workers and workflows. Controlling the collective worker requires managers to reject more than the traditional divisions between design and production and product and process. It compels them, however reluctantly, to reject rigid distinctions between conceptual and manual labor. Wringing "unproductive" labor out of the new "agile" production processes means rejecting traditional assumptions about what constitutes transformational labor. For managers, this means acknowledging that work and production are social activities. US managers's reconceptualization of work as a collective activity has pushed them, however reluctantly, to define new truths and to discard older ones that no longer work in the global market. Computer-based process-centered control systems permit managers to concentrate on concurrent (direct and indirect) control systems, process integration, and recomposition rather than merely on task decomposition. In doing so, however, they are forced to rely on normative control mechanisms, e.g., group/peer pressure and identification with the firm. And they must look for new ways of determining the value added by formerly invisible but crucial "indirect" workers, such as systems designers and technical support staff. Product development and data processing are redefined from overhead activities to products with "internal customers," i.e., other transformational workers in the overall workflow of the "value chain." Hence, work that was formerly considered merely overhead, such as communications and warehousing, or peripheral, such as process design and customer service, is now reconceptualized as integral to the larger production process and the overall "value chain." If formerly peripheral work is now part of the "value chain," managers need new ways to measure, monitor, and maximize "value." In the end, this is accomplished by requiring all workers -- old fashioned direct workers and guard labor as well as "creative workers" such as engineers and managers -- to become "design workers." Managers can control value adding workers by making everyone standardize, speed up, measure, monitor and control his or her own work. CI, TQM, and BPR are, in other words, attempts to dissolve the distinction between production, design, process flow, and control. If true, this is the reverse of taylorism: process- centered production control collapses (or claims to collapse) the distinction between conception and execution and between productive and unproductive labor.20 How, then, do managers do it? Controlling and inspiring the collective worker means reinventing an old management tool, the team. Reinventing the Team Management-defined and controlled teams are nothing new. They have been a part of formal management theory since at least the days of Edward Filene and welfare capitalism. Conventional teams, however, are specialized and prone to the same shortcomings of serial, over-the-wall functional divisions of labor. Management critics call them "silos." In contrast, CI, TQM, and BPR teams are simultaneously normative and organizational systems.21 They are normative systems because they formally acknowledge, even reward, the ability of workers-- all workers -- to decompose, recombine, chart, and monitor their own work. If TQM and BPR, with their obsession with facts and data and controlling variation, are "hard" control strategies, at the same time they also use "multifunctional teams" to shift some decision making from managers to "the people who actually do the work." In other words, the process-centered command-and-control systems are prepared to permit some workers some flexibility in fragmenting and recombining some of their work processes under some circumstances. Instead of relentlessly stamping out tacit knowledge -- a cardinal principle of taylorism, with its horror of "rule of thumb" -- TQM and BPR encourage employees to "surface" tacit knowledge in order to systematically incorporate it into "rational" processes. Managers call this transformation of worker knowledge into "fact-based management"_ empowerment_. The practice of "empowerment" depends in turn on two major "soft" innovations. Both are rejections of traditional systems of fragmentation. The first is an emphasis on flexible organizational structures. BPR and TQM in particular provide managers with tools for inventing new "hard" metrics of work effort and "value adding" activity. At the same time TQM and BPR in particular are also strategies to restructure social relations in the work place. Operationally, the goal is to shorten communications times and reduce communications "distortions," that is, to reduce the production of "bad data" or "inaccurate information." This is done frequently through the creation of teams equipped with appropriate computer-based technologies, e.g., computer-numerically controlled machines, email, teleconferencing, computer-aided systems engineering and design tools, and so on. In practice, teams are ways to push traditional policing, coordination, and reporting functions down the production and process chains of command. The decision-making authority offered to "self-directed work teams" is a perfect expression of March and Simon's (1958) bounded rationality. The decision-making authority is the right to choose among a limited range of options that have been designed into the system at a higher level of design or coordination (cf. Perrow, 1986).22 Computer-based technologies, which can be used to monitor in "real time," provide built-in surveillance and control. In some workplaces, technology and bounded rationality combine to produce flexible organizations that are nothing less than surreal: Oticon, the Danish maker of hearing aids, has abolished fixed workspaces for nearly all its workers. Instead, designers, clerks, secretaries, and technicians pull mobile filing carts and work benches from terminal to terminal as production workflows shift. Continuity, communications, and control are provided by local area computing networks and wireless telecommunications devices. Only the CEO has a permanent desk and a private office ( Cf. Bjorn-Andersen and Turner, 1994). The second "soft" innovation is a systematic emphasis on ideology or "culture" as a control mechanism. BPR, TQM, and their variations are comprehensive "philosophies" of workplace relations. There is nothing new, of course, about managers trying to influence the way workers think and feel about their condition and their relationship to their employers. Even taylorism itself is about a deal employers offer to workers: check your brains at the door, do what we tell you, and in return we will share (some) of the increased productivity in the form of higher wages. The normative offensive that accompanies TQM and other process-centered work reorganization is of a different sort. To the extent that the work of imagination cannot be standardized, TQM and BPR are obliged to deploy a cultural strategy, that is, to, rearrange attitudes, mostly about relations of power and control. Normative control mechanisms are intended to alter the beliefs and behavior (the "corporate culture"), chiefly with regard to the definition of "customer" and the functions of middle management. Power and control are given new names like flexibility and cooperation. The notion of customer is broadened to include all transformational workers in the value chain. Middle managers are transformed from police and supervisors to "coaches," "leaders," "facilitators," and "resources" who inspire rather than flog. At the other end of the transformational chain, direct workers, newly empowered on self-directed multifunctional work teams, are informed they are engaged in a "win-win" quest with their employers. Cooperation and jointness replace us vs. them adversarial relations, except, perhaps, when joining forces against a common enemy, e.g., foreign competitors. When necessary, employers can quickly turn such arguments into a case against unions. Finally, TQM, and CI in particular are, in a commonly employed phrase, a "belief system" as well as a philosophy. Invoking a belief system allows managers to "surface" hidden agendas as well as tacit knowledge. The unrelenting pressure of "satisfying the customer" -- that is, to work faster and produce more -- combined with the social pressures of team membership can get team members to reveal their attitudes and opinions, to confess their individual sins, and seek forgiveness from the group. In other words, teams make it easier for managers to apply group pressure to slackers.23 The ideology of "empowerment" reflects the new management conviction that "delayered" or flattened organizations -- organizations in which work teams monitor and intensify their own labor -- can extract value more quickly and efficiently than traditional systems of fragment and flog. This is the most explicitly ideological component of TQM theory. Theorists of process-centered command-and-control systems are quite open about the need to achieve a "world class paradigm shift."24 The following table suggests where the new process-centered control strategies differ from their more conventional predecessors. [TABLE ABOUT HERE] Multi-tiered Control Strategies The real power of CI, TQM, BPR and similar process-centered systems is clearest when they are combined in a multi-tiered control strategy. Together, they continually refine engineering-based control systems for routine processes, radically reorganize whole production systems and work flows, and review the "core competencies" of the enterprise on an ongoing basis. The tiers look something like this: 1. Continuous Improvement for routine processes characterized by simple, repeatable events. Repeatable events range from conventional mass production processes in manufacturing and data entry work to customer service, sales, or other activities relying on scripts or similar limited "decision trees," to assembling "off-the-shelf" modules to make computer software. The goal is to wring out both variations in the product and "slack," mostly slack labor. This is the stuff of traditional taylorism, and it increasingly makes use of IT-based fragmentation and decomposition. The focus is still largely on the individual and individual tasks. 2. Total Quality Management for "problem surfacing" and dealing with less routine and more "creative" work. This is where responsible autonomy meets kaizan to produce "self-directed autonomous high performance work groups." It is also where "hard" (engineering) and "soft" (human resource) managers struggle over the exact balance of measurement and inspiration. The focus is more squarely on systems -- and embedded control structures -- rather than on tasks. The goal is to devise ways to eliminate defects in the production system itself (the "process"). Here "defects" means employing more people than necessary, not just deviations from product specifications. Successful TQM makes "exceptional" work routine. It can then be pushed down the command-and-control chain and managed by continuous improvement techniques. 3. Business Process Reengineering for effecting radical change and solving complex or crisis-level problems. Usually BPR is done under the direction of senior managers and consultants who have a warrant to "downsize" or otherwise sharply reduce costs and cycle times in a hurry. Once an enterprise is radically engineered, its "systems" can be managed by TQM and CI teams and processes. To take Hammer and Champy at face value, BPR provokes fundamental questions about the very purpose of the enterprise. Continuous improvement and TQM focus on the middle and bottom of the process hierarchy. The goal of each separately and in combination is to minimize exceptions -- variations, defects, waste, scrap, anything that does not "add value" -- and to codify and regularize as many "erratic" processes as quickly as possible. Together, these are considerably more than a repackaging of older direct and indirect control mechanisms. The process-centered control systems are no longer interested in rigid distinctions between conception and execution. They are concerned instead with the distinction between routine and replicable labor on the one hand and non-routine and exceptional labor on the other. If the goal is to maximize all value-adding activity, inspiration labor has to be made predictable in the same way routine processes have to be made organizationally flexible. If managers can do this, their dilemma is solved. Conclusion Every opportunity is also a problem. The global competition that has accompanied global markets has forced US managers to reexamine and change long-established production relations. Continuous Improvement, Total Quality Management, Business Process Reengineering, and similar command-and-control systems are as much a departure from taylorist command-and-control systems as taylorism was a departure from craft-based control of production. This has not happened because some forms of crucial "creative" work resist routinization. It has happened because employers have run out of ways of appropriating the "added value" of all workers as quickly as the new global market demands. The original taylorist deal -- when it was honored by managers at all -- was degraded "rationalized" work in return for some of the increase in productivity. TQM, CI, BPR, and especially "empowered" teams are formal acknowledgments by managers that a century of relentless rationalization has reached its limits. They are also a recognition that all work _contra Taylor_ is conceptual and creative. It is therefore a recognition that all labor, sooner or later, formally or illicitly, deals with exceptional rather than routine demands. Even work that is predictable is not necessarily replicable and work that is replicable is not necessarily predictable. It is, finally, a recognition that a more efficient expropriation of the "value-adding" potential of all workers depends on defining and exploiting a fuller range of their conceptual labor. Hence, the new process-centered work organization schemes require more -- if also more embedded -- control, not less. It is the major breakthrough of Deming, Hammer, and others to attempt the intensification of all labor by collapsing social distinctions between design, production, and even service work on the one hand, and management, coordination, and communications on the other. In doing so, they have presented managers with tools to systematically combine "soft" normative and "hard" engineering control systems in remarkably flexible ways. These "hard" and "soft" combinations have little in common with the heavy-handed combinations of human relations inspiration and taylorist task decomposition. CI, TQM, and BPR acknowledge the truth of Braverman's insistence that human labor is simultaneously routine and creative, unique and replicable, instinctive and conscious. That managers have done so in order to extend still further the "ambit of capitalist production" should not prevent us from either appreciating the vindication of Braverman's analysis or the ongoing need to revise it to reflect changing realities. TABLE TRADITIONAL TAYLORISM: CONTROL AND INSPIRE: FORMAL GOALS Management/worker "hearty "Win/win" cooperation" Reduce worker fatigue/injuries Empower employees Increase output/wages Become more competitive ("save jobs") WORK/WORKER FOCUS Individual Workers Collective Workers / Teams Production/clerical workers Designers, engineers, managers "Routine" work All "value adding" work Working hard Working "smart" PRODUCTION GOALS Increase productivity Add value/reduce cycle time Quality Control (inspect for Statistical process control defects) ("build quality in") ORGANIZATIONAL STRUCTURES Top Down ("check brains at Top Down ("empowerment") the door") Integrated organization "Virtual organization," i.e., outsourcing Decomposed labor processes Controlled recomposition of labor process Tightly drawn hierarchy Controlled "delayering" "Permanent" workers Contingent workers MANAGEMENT FOCUS Speed up "Continuous improvement" Root out soldiering Lean and Mean Conflict Problem Management WORKER ORGANIZATION Hierarchical unions/ Independent contractors/"Union "permanent" employees free" contingent workers Individual piece rates "Collective" (team) time-based production Adversarial bargaining Jointness and cooperation SCALE Local/regional competition Global competition Local wage rates Global wage scales PROBLEMS Never applied "right" Never applied "right" Employers "cheat" Inadequate support from top managers Employees "cheat" Employees inadequately motivated Greed Short-term thinking ENDNOTES 1 Here and in other "off the net" extracts I have standardized spelling and grammar. 2 Emergent, virtual, agile, and even dissipative organizations have received a lot of attention from theorists of postmodernity. How much of this is real and how much is wishful thinking is still to be determined. My favorite combination of description and fantasy, at once tacky and insightful, is the collected work of Newt Gingrich's inspirations, Heidi and Alvin Toffler. The most trenchant scholarly analysis of modern and postmodern organizational theories is Harvey (1990). See in particular his examination of control systems in Chapter 10, "Theorizing the Transition." Harvey, alas, wrote The Condition of Postmodernity before the current waves of TQM and BPR. See Roseneau (1992); Boje and Winsor (1993). A much more modest example is Kraft and Truex (1994). See also Lyotard (1987). 3 Slaughter and Parker (1994) among others have properly attacked "phony competition" as a weapon to whipsaw workers. Here I address a different issue: the mechanisms which permit (and, from the perspective of managers, require) extraordinary and concurrent control strategies. 4 The management concept of value-added was popularized by Porter (1985; 1990) in his "value chain" analysis. Some of the most interesting theoretical critiques of "value adding" come from the so-called radical accountancy school. Cf. Yuthas and Tinker (1994). 5 Cf. Simons (1995); see also Gouldner (1954); Perrow (1986); Kusterer (1978); Klein and Kraft (1994a; 1994b). 6 The relative lack of interest does not mean, of course, that routine work and mass production no longer exist. It means only that the displacement of much formerly US-based mass production work to low wage (and often politically repressive) regions makes control of mass production workers less urgent for US employers. Justified or not, employers have shifted their focus. An unusually clear management statement of this is Simons (1995). See Boje and Winsor (1993). Ironically, while union density in the US private sector has declined to its lowest rate since World War Two -- about 10% -- the US telecommunications workers union, the CWA, has actually seen a substantial increase in its membership of workers in the crucial IT sector. See CAW (n.d.). 7 Reich's distinction comes from the "people, data, things" classification scheme long used by the US Bureau of Labor Statistics. 8 Stalk and Hout (1990) call this "time-based competition"; see also Hammer (1990); Hammer and Champy (1992); Davenport and Short (1990); Deming (1984); Porter (1985; 1990); Clark and Fujimoto (1991); Ishikawa (1985); Goldratt (1986). 9 Two very different examples are Kusterer (1978) and Berggren (1992). 10 Cf. Kunda (1992); also Kraft (1977); Friedman (1977); Meiksins (1988); Whalley (1988); Jackall (1988); Kidder (1981). 11 Kusterer (1978), in his early critique of Labor and Monopoly Capital, pointed out that "routine" work routinely calls for imaginative and often illicit solutions from "unskilled" workers. The problem of routinizing imagination, it seems, is important for managers only when imagination is both the process and the product. 12 Kraft (1977); Kraft and Dubnoff (1986); Friedman (1977). 13 Cf. Stalk and Hout (1990); Pine (1993); and Simons (1995). The phrase comes from Gouldner (1954). This is sometimes called the "effort extraction problem" by labor economists: "[W]ork effort is hard for employers to observe, so it must be induced by strategic means" (G. Skillman, 4/4/1995 PEN-L). 14 Cf. Kunda (1992) and Barley and Kunda (1993). See also Meiksins (1988). 15 One of the earliest, and still clearest, statements of MBO is Drucker (1954; 1964). See also Odiorne (1970). 16 This in distinction to older Human Relations efforts to systematize the "psychic reward system" US managers have used to affect worker behavior and attitudes. See note 18. 17 The best description of these new forms of "management by stress" is provided by Parker and Slaughter (1994; 1992; 1988). See also Robertson et al (1993). The Canadian Autoworkers Union has produced a remarkable film, Working Lean, which documents the similarities between "working smart" and taylorist speed ups. At the risk of anticipating the discussion of teams and labor intensification, here is part of a recent post on an Internet TQM discussion list for managers: I am not a consultant or a quality guru. I am just a blue collar worker. I am not stupid .... My company has been in the process of switching to TQM for over two years. My wife went to work in a new factory about a year ago that opened up under the TQM concept....The workers get half the pay and twice the work that they would get if they were employed by the same company in one of their other non-TQM plants. They work long hours under brutal conditions and have absolutely no input in plant decisions although that is what was preached to them during indoctrination. They have no bosses only "resources" who have more authority and are less responsive to the employees than any boss [they] ever had. After all since this is a "team concept" if work doesn't get accomplished it obviously can't be the fault of the "resource" so it must be the fault of the team members....I don't have reams of research to offer you. I can only say that TQM sucks. It is a return to the stone age for the working person. It is nothing but stretch out and company unions wiping out worker gains and worker rights for the benefit of management. It is just another tool for manipulation. Nothing new about that. (QUALITY List). The list was deluged by dozens of responses from sympathetic researchers and practitioners. Their unanimous response was that either the writer's employer hadn't applied TQM "right" -- or the post was some sort of hoax. 18 This raises a rich set of issues. Taylor assumed regional, or at most national, markets, including labor markets. The first modern systematic work rationalization schemes -- those of Taylor, Fayol, the Gilbreths, Gantt, Mogenson, Shewhart -- examined the tasks of individual workers or perhaps a single department in a single site. Fragmented work was redistributed and eventually recombined within the same workplace. "Postmodern" managers confront a different world. Work, including creative work, can now be moved around the globe to suit employers' requirements for low labor costs, predictability and timeliness (cf. Shaiken 1984; Aronowitz and DiFazio, 1994). Skills have been transformed into "knowledge," knowledge is operationalized as "data analysis" and "just-in-time knowledge transfer," and finally retailed as "expert systems." The local perspective also meant dealing with local labor supplies and roughly predictable market cycles. The new systems of work rationalization, although always applied in specific times and places, are conceived globally. The shift has fundamentally changed the way managers conceive the labor process, the workplace, labor markets, and even the definition of work itself. (See Chandler, 1977). Finally, ever since the Western Electric studies, management theorists have understood perfectly well that production is a social activity. As Braverman (1974) points out, management's initial response was to try to capture the subjective bases of worker "resistance" and turn it to their own advantage. The result was human relations theories of various kinds that "scientific" managers generally rejected as too soft and squishy. The process-centered control systems described in this paper go far beyond these crude "contented cows give more milk" manipulations. 19 "Total Quality," in the form of "total quality control," was probably first used by Feigenbaum (1951). The person popularly associated with TQM, Feigenbaum's junior colleague, W. Edwards Deming, refused to use the term. Instead, Deming preferred to talk about "quality" and continuous improvement. Deming's own acolytes and commercial competitors have been less reluctant to use what has by now become a generic label. Cf. Walton (1986) and Berry (1991). 20 Feminist critics of Braverman have pointed out that his use of waged craft work as the model of skilled labor effectively renders women's work invisible as well as unskilled. Management's recognition of the collective nature of work raises still more questions about gender, work, and skill. Cf. "What Gender is the Collective Worker?" (Kraft, in preparation). 21 The following discussion is drawn in part from Klein and Kraft (1994a). 22 For a more detailed discussion, see Klein and Kraft (1994b). 23 The best single analysis of collective pressure on the team-member/collective worker is Parker and Slaughter (1988). See also Robertson et al (1993), the CAW film, Working Lean, and Leary (1995). 24 Note the complaint of a manager in the following melding of engineering measurement and normative control (and the implicit criticism of taylorism) in this suggestion for dealing with affirmative action reporting requirements: Appreciating the need for diverse opinions, ideas, etc. is at the very heart of problem solving. A part of affirmative action also applies in a similar manner as the legislation banning discrimination due to disabilities. Both to maintain effective work and to comply, companies have to be able to (or should be able to) document each job/position by the tasks that when bundled together make up that task. At present most companies don't have a clue why certain tasks are attached to particular jobs. They grow and are or have been passed down through the firm's history. Step number one in most quality and participation processes should be to document and flow chart all the tasks actually done within the workplace. Step two is to determine the value added or not by the task as a whole of the individual steps with the task. If companies/agencies would do this, they would not only be prepared to better deal with becoming more effective and more efficient, they would be able to fully comply with affirmative action policies regarding race, gender or ethnicity, but to disabilities as well. Once the tasks are documented, you can see how they can be rebundled differently to suit the talents in the company and applicants who have been discriminated against in the past." (QUALITY discussion list). WORKS CITED Adler, Paul S. Time and Motion Regained. Harvard Business Review, v. 71, no. 1, Jan-Feb 1993 (pp. 97-108). Aronowitz, Stanley and William DiFazio. The Jobless Future: Sci-tech and the Dogma of Work. University of Minnesota Press (Minneapolis: 1994). Barley, Steven and Gideon Kunda. Design and Devotion: Surges of Rational and Normative Ideologies of Control in Managerial Discourse. Administrative Science Quarterly, v. 37, 1992 (pp. 363-399). Baskerville, Richard, et al, eds. Transforming Organizations with Information Technology. North-Holland (Amsterdam: 1994). Berggren, Christian. Alternatives to Lean Production: Work Organization in the Swedish Auto Industry. ILR Press (Ithaca, NY: 1992). Berry, Thomas H. Managing the Total Quality Transformation. McGraw-Hill (New York: 1991). Bjorn-Andersen, Nils and Jon A. Turner. Creating the 21st Century Organization: The Metamorphosis of Oticon, in Baskerville et al, pp. 379-394. Boje, David M. and Robert D. Winsor. The Resurrection of Taylorism: Total Quality Management's Hidden Agenda. Journal of Organizational Change and Management, v. 6, no. 4, 1993 (pp. 57-70). Braverman, Harry. Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century. Monthly Review Press (New York: 1974). Chandler, Alfred. Visible Hand: The Managerial Revolution in Management. Harvard University Press/Belknap Press (Cambridge, MA: 1977). Clark, Kim and Takahiro Fujimoto. Product Development Performance: Strategy, Organization, and Management in the World Auto Industry. Harvard Business School Press (Boston: 1991). Communications Workers of America. Changing Information Services: Strategies for Workers and Consumers. Washington, D.C. (no date). Davenport, Thomas H. and James. E. Short. The New Industrial Engineering: Information Technology and Business Process Redesign. Sloan Management Review., v.31 no. 4, Summer 1990 (pp. 11-27). Deming, W. Edwards. Out of the Crisis. MIT Press (Cambridge, MA: 1984). Drucker, Peter F. Managing For Results: Economic Tasks and Risk-Taking Decisions. Harper & Row (New York: 1964). ____________. The Practice of Management. Harper (New York: 1954). Feigenbaum. Armand V. Quality Control: Principles and Administration. McGraw-Hill (New York: 1951). Friedman, Andrew. Industry and Labour: Class Struggle at Work and Monopoly Capitalism. Macmillan (London: 1977). Goldratt, Eliyahu M. The Goal: A Process of Ongoing Improvement (rev. ed.). North River Press (New York: 1986). Gouldner, Alvin. Patterns of Industrial Bureaucracy. Free Press (Glencoe: IL: 1954). Hammer, Michael. Reegineering Work: Don't Automate, Obliterate. Harvard Business Review, Jul-Aug 1990, pp. 104-112. _______and James Champy. Reengineering the Corporation: A Manifesto for Business Revolution. Warner Books (New York: 1992). Harvey, David. The Condition of Postmodernity: An Enquiry into the Origins of Cultural Change. Blackwell (Cambridge, MA: 1990). Ishikawa, Kaoru. What is Total Quality Control? The Japanese Way. Prentice-Hall (Englewood Cliffs, NJ: 1985). Jackall, Robert. Moral Mazes: The World of Corporate Managers. Macmillan (New York: 1988). Kidder, Tracy. Soul of a New Machine. Little, Brown (Boston: 1981). Klein, Heinz, and Philip Kraft. Social Control and Social Contract in Networking: Total Quality Management and the Control of Knowledge Work, in A. Clement, P. Kolm and I. Wagner, eds., NetWORKing: Connecting Workers In and Between Organizations, pp. 45-53. North-Holland (Amsterdam: 1994a). ________________________. Social Control and Social Contract in NetWORKing: Total Quality Management and the Control of Work in the United States. Computer Supported Cooperative Work (CSCW), v. 2 (1994b), pp. 89-108. Kraft, Philip. Programmers and Managers: The Routinization of Computer Programming in the United States. Springer Verlag (New York: 1977). ________and Steven Dubnoff. Job Content, Fragmentation, and Control in Computer Software Work. Industrial Relations, XXV:2 (Spring 1986), pp., 184-196. ________and Duane P. Truex. Organizational Emergence and the 'Postmodern' Enterprise, in Baskerville et al, pp. 113-128. Kunda, Gideon. Engineering Culture: Control and Commitment in a High-Tech Corporation. Temple University Press (Philadelphia: 1992). Kusterer, Ken. Know-How on the Job: The Important Working Knowledge of "Unskilled" Workers. Westview (Boulder, CO: 1978). Leary, Elly. Back to the Future:Workers and the New 'American Plan.' New Directions Education Fund (St. Louis: 1995). Lyotard, Jean-Francois. The Post-modern Condition, in After Philosophy: End or Transformation? in K. Baynes, J. Bhoman and T. McCarthy, eds. MIT Press (Cambridge, MA: 1987). March, James A. and Herbert Simon. Organizations. Wiley (New York: 1958). Meiksins, Peter. The "Revolt of the Engineers" Reconsidered. Technology and Culture, v. 29 (Apr. '88), p. 219-46. O'Connor, James. The Fiscal Crisis of the State. St. Martin's Press (New York: 1973). Odiorne, George S. Management by Objectives: A System of Managerial Leadership. Pitman Publication Corp. (New York: 1965). Parker, Mike and Jane Slaughter. Choosing Sides: Unions and the Team Concept. South End Press (Boston: 1988). ________________________. Management by Stress: The Team Concept in the US Auto Industry, Science as Culture, v. 8, pp. 27-58. Perrow, Charles. Complex Organizations: A Critical Essay. (3rd ed.). Free Press (New York: 1988). Pine, B. Joseph. Mass Customization: The New Frontier in Business Competition. Harvard Business School Press (Boston: 1993). Porter, Michael. Competitive Advantage: Creating and Sustaining Superior Performance. Free Press (New York: 1985). ___________. The Competitive Advantage of Nations . Free Press (New York: 1990). Reich, Robert B. The Work of Nations: Preparing Ourselves for 21st-century Capitalism. (A. A. Knopf (New York: 1991). Robertson, David et al. The CAMI Report: Lean Production in a Unionized Auto Plant. CAW-Canada Research Department (Willowdale, Ont: 1993). Roethlisberger, F. J. and William J. Dickson, with the collaboration of H. A. Wright. Management and the Worker. Harvard University Press (Cambridge, MA: 1946). Roseneau, Paula Marie. Post-Modernism and the Social Sciences, Princeton University Press, (Princeton: 1992). Shaiken, Harley. Work Transformed. Automation and Labor in the Computer Age. Holt, Rinehart and Winston (New York: 1984). Simon, Herbert A. Administrative Behavior (3rd ed.). Free Press (New York: 1976). Simons, Robert. Levers of Control: How Managers Use Control Systems to Drive Strategic Renewal. Harvard Business School Press (Boston: 1995) Slaughter, Jane and Mike Parker. Working Smart. A Union Guide to Participation Programs and Reengineering. Labor Notes (Detroit: 1994). Stalk, George and Thomas M. Hout. Competing Against Time: How Time-Based Competition is Reshaping Global Markets. Free Press (New York: 1990). Toffler, Alvin. Powershift: Knowledge, Wealth, and Violence at the Edge of the 21st Century. Bantam Books (New York: 1990). Walton, Mary. The Deming Management Method. Putnam Publishing (New York: 1986). Whalley, Peter. The Social Production of Technical Work: The Case of British Engineers.State University of New York Press (Albany, NY: 1986). Wood, Stephen. The Lean Production Model. Background paper prepared for the Lean Workplace Conference (Canadian Autoworkers Union), Port Elgin, Ontario, Canada (September 1993). Yuthas, Kristi and Tony Tinker. Paradise Regained? Myth, Milton, and Management Accounting. Typescript (1994). ***** [End of file]