The specialists had failed to make an enduring dent in the organization of the card center's labor processes. As outsiders, they had not gained the support of line managers and made only a minimal impression with their recommendations. In part, then, the installation of a new level of management was an attempt to gain supervisors' trust and overcome their resistances to reorganizing different production contexts.

On their own, the line supervisors themselves had very little time to "innovate," as Len put it. They were under tremendous pressure to begin increasing productivity; the center was swarming with efficiency groups and external consultants who were hired to evaluate the card center work and management processes. It was in this context that a new level of middle management was put in place. New section managers were to manage a "quick turnaround" of the center, to make sure new systems would meet specific production demands. Divisional-level management made this commitment in the belief that the new managers, located in work units on a quasi-permanent basis, stood a greater chance of being integrated into the management structure and therefore would possess greater legitimacy to act as agents for the larger restructuring agenda.

Thus the new level of middle management had a very explicit role in the transition to a newly organized credit card division. Not only would they monitor supervisors and workers under them and communicate between higher divisional management and the line; they would actively seek ways to reorganize the labor process, both in its current unautomated and its future automated versions. This mission lent them a significant degree of autonomy with which to shape the course of a new productivity regime. It especially gave them a basis on which to reject more coercive aspects of that regime. Their rejection was strengthened significantly because the negative side of past growth strategies-the total neglect of technology and organization in the card center-limited strategic management's credibility in gaining managers' consent to applying coercive norms.

Card Center Managers and Managing Change

Like other middle managers, managers in the card center were critical of strategic management's role in shaping the current disarray of their division. They had a very clear view of how the turnaround task before them was adversely affected by the inadequate investment strategies of the past. But managers in the credit card unit had a different position in the new corporate agenda. Recruited by strategic management to participate in the overhaul of the center, these managers had an autonomy from which they could set limits on certain demands emerging from the restructuring process. When card center managers were pressured to raise the productivity bar and to use higher production levels as a justification for managing out, they responded by wielding their discretion to their own ends-to increase productivity within a different set of parameters. Because they believed that productivity problems originated in labor-process conditions rather than in workers themselves, they used their knowledge about the work process of their units, past management practices, and strategic management's decision making to fend off intense speedups.

When Ralph Holstein began managing the customer-inquiry-filing unit on the cardholder side of the center, his manager recommended that Ralph push the supervisor below him to manage out several file clerks in low-grade positions. Ralph's filing unit created files, kept track of files, and retrieved them in response to customer inquiries about credit card accounts. File-tracking was separate from the actual work of answering phones to handle the inquiries; file clerks' work was set in motion by requests from customer representatives. The area was extremely paper-intensive, and access to the files was far from routinized.

When Ralph first took this managing assignment his employees could find only 70 percent of the files they needed at any given time. To give a sense of the proportion of this operation, in the three months before our interview, 140,000 files had been created, all of which were kept in metal file cabinets and cardboard boxes. Because of minimal organization in this unit, there was no system to indicate where a file had gone once it was removed, and employees could not effectively track the file and ensure its replacement. Several of the clerks in the unit had received low rankings and were perceived as slow and inefficient, an impediment to intensified productivity goals.

In Ralph's perception, the chaos of the existing labor process imposed insurmountable limitations on the productivity of the clerks. A high level of employee anxiety worsened this chaos. Many of Ralph's employees were worried about the effects of what he called the "aggressive more toward automation" on their jobs. He blamed the way the unit had been managed in the past for much of this chaos. Employees had been unwillingly subjected to the disorganized character of the card center, he argued; a manager was justified in working to manage them up to higher productivity levels.

In his position of reorganizing the labor process without computerized systems to handle the file documentation and retrieval electronically, Ralph's first major project had been to systematize the filing procedures so that clerks would have ready access to any given file or, if the file had been drawn, would know precisely where it was.

Ralph's innovation was at once a dramatic improvement with respect to the existing system and startlingly simple with respect to organizational sophistication. Under the new system, clerks replaced files with one-page forms on which the name and account number of the file, the name of the clerk withdrawing it, and the date of withdrawal would all be recorded. These forms thus provided information about the location of the file; under the old system the whereabouts of a file had remained a mystery until it was replaced.

The new system raised the accuracy of retrieval from 70 to 90 percent and simultaneously increased the productivity of the clerks. Ralph was also generating productivity statistics that had never before been maintained. These statistics were used to manage up employees. He was gradually transferring all information on work flow and output-information on numbers of inquiries, the types of files being pulled, the number of files different departments were bringing in per day, week, and month-to a small computer; tallies of the information would be made available to employees. He called the old system, under which no one had up-to-date information on production activities, a hit-or-miss system inadequate to the rapidly changing production environment.

That managers use statistics on employee work activities to monitor, control, and manage employees may seem a most obvious and even insidious truth to students of the labor process. Typically, workplace observers consider primarily the negative implications of such detailed monitoring procedures for workers. Under the conditions described in this study, however, in which middle managers were challenged to raise productivity levels and maintain some degree of consent to ongoing productivity objectives in the context of restructuring, I would argue that managers used such statistics to set limits on the degree to which workers could be pushed, from higher up in the corporation, to achieve new productivity levels. Ralph was devising a system for ultimately managing up the production employees over whom he held responsibility; but he maneuvered this system in opposition to more extreme demands, from higher divisional-level management, for increasing productivity and getting rid of employees who could not reach greatly increased production levels.

While Ralph turned his attention to innovating the existing organization of work in the file unit, he simultaneously protected his own prerogative for managing his employees. The lack of a viable way to measure productivity under the old file system further strengthened Ralph's ability to generate new productivity statistics that conformed to the changing labor process and the relevant constraints on productivity. He simultaneously opposed demands from higher levels of the division for productivity and staff reductions.

Diane Timbers was a "master" at manipulating numbers and variables to maintain control over her unit of customer service representatives. The seventy representatives, whom she managed through five managers, spent their time plugged into IBM PCs and headphones, answering questions about credit card accounts. A firm believer in the idea that numbers enabled workers to define and participate in achieving productivity goals, Diane's weapon of management was a massive statistical volume in which she documented multiple bases of productivity. She had systematically generated numbers on all aspects of staffing, such as number of full- and part-time employees, classified by hours worked, productivity, grades, ranks, and salaries; customer inquiries, stratified by credit products, type of credit inquiry, number of complaints, and marketing costs; and productivity measures such as the number of incoming and outgoing telephone calls, their length, and resolution time frames.

Diane had been brought in to do "systems enhancement": to find better ways to organize this particular customer inquiry unit. Before Diane's arrival the "reps" had worked two to a terminal. Now each rep had his or her own PC. (The reps were predominantly, but not solely, women.) One of her pet projects was to help design new screens for the inquiry process that would enable the reps to solve customer problems better.

The level of uncertainty in the work of these employees was staggering. The rules and regulations pertaining to credit cards changed continually as new credit products were introduced and chargeback time frames and procedures changed. Customer service jobs had a six-month learning curve in which reps attended about ten hours of training sessions a month to familiarize themselves with handling the 3,000 to 5,000 daily customer inquiries.

Diane had hired two extra supervisors to juggle the demands of what she described as a "fast-paced environment." In addition to managing the work flow, supervisors were actively involved in handling calls that representatives could not resolve. Thus Diane was plugging in staffing gaps, using supervisors to increase the general productivity of the unit.

Despite the complex and increasing work load incurred as more cardholders were brought into the bank's card program, divisional-level management made several specific demands for higher productivity in Diane's unit. For example, one of the objectives written into a recent PPCE was that Diane would manage an overall reduction in the costs of the unit's phone calls and therefore in the amount of time it took her representatives to resolve customer inquires. Ostensibly, that objective meant that she and her supervisors would pressure their employees to speed up their telephone transactions, with the ultimate goal of reducing the number of representatives needed to staff the unit.

As it turned out, the costs of her unit's calls increased in that period. It did so, however, because of the addition of a new phone system in the unit, a fact that Diane was able to establish through detailed calculations of a variance analysis. That system, an automatic call distributor, more efficiently and rapidly forwarded phone calls to the numerous representatives. By factoring in the costs of the new phone system, as well as giving the reasons for maintaining current staffing levels (which she was able to justify with her stratified measures of incoming calls and types of calls as a result of new credit products), Diane was able to justify the total telephone costs incurred in the period. She used an elaborate set of equations because, in her words, she refused to have her unit penalized for what she saw as an "arbitrary intervention," wherein someone, somewhere higher in the division decided on a new area of cost cutting.

In that particular case she used the numbers to argue on her own behalf with her manager about what was and was not possible with regard to cutting both costs and staff. More to the point, Diane used those production numbers to protect her unit against the confusion surrounding the installation of new technology systems. Her statistical manipulations allowed her to invalidate new demands by showing how her work area was negatively affected by the new systems.

Diane viewed the new concept of managing out as similarly arbitrary. In her view, employees who fell on the bottom 15 percent of the ranked curves performed their jobs quite adequately. She also insisted that "managers want to manage up. It's in our interests to get people to perform better. Managing out is only a solution if managing up doesn't work."

Diane claimed that it was critical to set boundaries on what higher-level management could squeeze from those production areas. Their attention was only focused on profitability, without considering the negative and undermining implications of heavy-handed approaches to achieving profitability. "Top management now sees the card center as very profitable. But in terms of the care they're taking for our long-run success in pushing through new programs it's like we're living hand-to-mouth. If the unit can make profits top management wants us to make even more profits."

Sarah Fleischman saw the entire card center as completely out of control, a state stemming from past neglect of automation and productivity standards. As a section head on the merchant side of the center, Sarah managed six supervisors and through them seventy indirect reports. Her unit processed chargebacks: whenever a cardholder or cardholder's bank disputed a charge made by American Security's participating merchants, the amount was charged as a debit to American's account until proper liability had been established and American Security had been cleared of the charge. It is worth looking at this particular unit in some detail, for it reveals the pressure of processing information in an environment in which every second funds are unaccounted for is a drain on company profits.

For the period during which a chargeback was unresolved, the bank was liable for the figure under dispute and lost the additional interest that would accumulate on that amount. In this extremely labor-intensive production process the bank's risk of high and uncontrolled losses ran higher the longer it took to account for these debits.

The employees in Sarah's unit adjusted claims as soon as a customer disputed the account. Their role was to document and investigate each claim meticulously. They obtained both the draft of the account from "file builders" in another unit and documentation from the cardholder about why the charge was disputed. The adjuster was to consider all aspects of the dispute and make a decision about proper bank action. Any number of factors could cause a customer to initiate the chargeback process. A cardholder's name might appear on a "Warning Bulletin" listing individuals prohibited from using their credit cards: if the merchant did not refer to the bulletin but nonetheless accepted a customer's card, the cardholder's bank could challenge American Security's charge. Or a merchant might claim they had billed a cardholder for an order placed over the phone, although the cardholder denied placing the order.

Whatever the nature of the dispute, it was the responsibility of the file adjuster to provide accurate data for settling the case, optimally in favor of American Security, by removing the charge from the liability column. When Sarah entered this position in early 1985, the unit had an enormous backlog of chargebacks. The rows of adjusters' desks stacked high with photocopies of documentation at every stage of every account under investigation were vivid testimony to the lack of effective measures for both "capturing money" and "capturing data."

Sarah had been promoted to section head to facilitate this chargeback unit's transition to an automated environment. She noted that in some locations, on the cardholder side of the center, employees were able to bring up customer accounts handily onto a terminal screen for quick responses to inquiries; on her side, however, terminals and PCs had not yet been introduced. She claimed that "Mr. 'X' [the past CEO] wouldn't approve a budget for investing in new technologies in the card center. He also wouldn't go outside American Security Bank for software. It was a situation where things were left alone and management didn't deal with these issues." When she took on her large chargeback unit she found it very difficult to get it to move: "When I first came people didn't know quite how to measure things. Managers hadn't known how to do budgets or any kind of staff forecasting. They didn't have time to step back and ask how can this be done better. I inherited a very messy situation."

By eliminating some duplication of cross-referencing procedures and by instituting measures that enabled the chargeback unit to reduce the error ratio from 20 to 5 percent, she felt that she could move the unit beyond the enormous backlog of accounts to be resolved.

Nevertheless, she had been given new quotas for increasing volume and decreasing staff that were, in her view, "no good." In response to these projections she wrote a staffing and planning forecast that was more realistically aligned with the current capacities of the chargeback unit, factoring in new productivity potentials resulting from the preliminary reorganization strategies she and her supervisors had introduced.

Sarah's role as a representative for corporate change was enacted within parameters she developed about how to negotiate the existing inadequacies of the labor process with new pressures for productivity. Her philosophy was in distinct opposition to the philosophy of "opaque management," in which managers cajole workers to do something on the basis of individualistic concepts such as achieving one's personal potential through stretching, rather than because of the stark facts of pressures for corporate change. One manager liked it, she said,

Because I tell them [workers and managers] what's going on, what the rationale is for what they're doing. I don't just say "do this," or try to hide from them the political and economic pressures on our work. You need to explain or people don't get what's going on. Downward communication is very important. In this situation there's lots of political pressure, there are a lot of people leaning on us; it's especially bad when the bank is experiencing these bad losses.

Like Diane and Ralph, Sarah summoned the expertise she had gathered in her unit and managed her supervisors and workers within a different set of standards from those proposed by divisional-level management. She simultaneously used her leverage to manage divisional-level demands and to press for cooperation from those in the chargeback unit. In this sense, frank and open management was a form of persuasion and manipulation that served Sarah's role in maneuvering the unit toward higher productivity levels. She gained some degree of legitimacy from her oppositional position to top management, which she used to further her own aims in the production unit.

These managers undertook their mission with the belief that it was the organization of the labor process, rather than individuals themselves, that was responsible for productivity deficiencies. In reorganizing the labor process they worked to manage people up to new productivity levels. However, there were important limits to their view that the organization of the labor process alone was to blame for the lack of productivity. Managers in the card center felt strongly that workers and supervisors had to understand and cooperate in the demands for reorganization that were pressuring everyone. As the branch managers attempted to get branch employees to participate in increasing productivity, the card center managers tried to use their numbers to buttress their position vis-à-vis workers as well as divisional-level management. Ralph, for example, summarized and organized, in a central terminal, the productivity figures for his supervisor's filing unit so that workers themselves could develop a baseline standard for new achievements.

Diane's adamant posture on having comprehensive and accurate figures with which to guide her unit to new productivity levels was reflected as well in her belief that workers must take more responsibility for meeting higher goals in the unit. For this reason, she posted productivity figures, for the unit as well as for individuals, on a wall for all to see. She used her aggressive position vis-à-vis strategic management to strengthen her latitude with her employees. Her open-office policy and her ability to summon the data necessary to command her unit were weapons in winning the cooperation of her supervisors and their employees. Diane's opinion of employees' responsibility for their fate in the future of the card center reflected her unwillingness to bear the full brunt of responsibility:

I post the standards daily for the reps to see. The reps want to know what they are being measured against. Performance shouldn't be a mystery! The employees here have to make the choice about whether or not they want to be managed up. We give them that choice but they have to run with it. And frankness is the only way to deal with it, that's how we can decide if the employee is ultimately good for the job or not. It may be that there is a skills mismatch and that some kind of lateral mobility is the answer.

Those examples in which managers used computerized reports to accumulate statistics on output and productivity run counter to claims that computers inevitably deskill managers. In the card center, managers generated reports for divisional-level management and their employees to protect their units and their autonomy in decision making. Whereas numbers and reports stood as weapons for the expropriation of expertise from branch managers, numbers and reports were weapons of autonomous management in the card center.

Because of the different position held by the card center in the bank's restructuring agenda, higher-level management was limited in its ability to push or coerce those managers. Compared with its treatment of the branch system, strategic management had taken a hands-off approach to card center production sites, giving hands-on authority to managers to control and to initiate changes. Thus middle managers were given the authority to make restructuring decisions within their production sites. In the branch system, on the other hand, authority for restructuring decisions was external to production sites and was held, not by middle-level production managers, but by area management groups and the strategic-management-level policy team that was uniformly and by fiat organizing the reduction of branches.

Finally, this hands-off approach gave card center middle managers more room to protect themselves with their criticisms of strategic management for past strategies-specifically, the failure to update and automate the labor- and paper-intensive work processes. Strategic management recognized its own past policy deficiencies and sought the cooperation of middle managers to transform this large and potentially very profitable production unit. Under these conditions, middle managers' arena of discretion was actually enlarged, whereas in SystemsGroup discretion was circumscribed and in the branch system it was expropriated.

The new level of management in American Security Bank's credit card center was ushering in the corporate restructuring processes, but in a gradual way, so as to increase productivity within the constraints of existing labor process conditions. We could conjecture that strategic management's ultimate vision of the card center was of a work site that, once automated, would operate with far fewer supervisors and middle-level managers. The new technological systems, with their potential for centralized control, could ideally minimize the need for direct supervision and higher-level coordination of production groups. In this sense these managers might, in the long run, be managing themselves out of positions by reorganizing and preparing different work sites for new technologies. Indeed, Len Cordova speculated that the section head level of management was transitional: "They [section head managers] are there to speed up the conversion process. They make sure the controls are in place as new processes are instituted. Middle management may then be phased out."

At the same time, there may be limits to such an ideal picture, insofar as it underestimates the role of middle managers in coordinating workers and supervisors to use materials, machines, and time. When I asked Diane whether and how managers were necessary in her case, where workers were plugged into centrally controlled terminals, she responded, "The monitor may give numbers but numbers achievement depends on people and coordination. Also, you can have all the data in the world but you need someone to make sense of them. And I spend a lot of time on the floor. I have to be there to fight fires."

By 1987, middle managers had acquired an even larger role in negotiating between the ongoing productivity of work sites and new technological systems. Len Cordova had by that time been promoted to managing four "cost centers" on the merchant side of the center and was indirectly managing sixty-five employees. He described how strategic management's attempts to catch up rapidly to state-of-the-art systems and rectify decades of neglect had had a whole new set of unforeseen costs. Major systems glitches had occurred. The inability of units to assimilate all aspects of the new technologies had led to significant losses of electronically captured data, and thus to real monetary losses, and there was much uncertainty about the best strategies for setting the center on a more stable and routine automated track.^[1] The possibility of eliminating managers seemed remote at that point. Although strategic management may have viewed the new level of management as transitional, middle managers seemed to have been thrust into an even more critical role in staying a long course of conversion.

Conclusion

Labor-process theorists may find nothing remarkable about the conclusions drawn in the last three chapters. Middle managers deploy their discretion to achieve certain corporate ends. They attempt to draw workers and other managers into intensified efforts to increase productivity. Those claims may provide ample proof of a one-sided view in which middle managers willfully execute restructuring, possessing a predictable set of interests by which they manage, in tandem with top management, in the coercive corporate interest.

Examining middle managers working within different production contexts and facing different sets of constraints demonstrates how, in fact, middle managers' actions were shaped by an alternative sense of the corporate interest. American Security's managers contested the ways that strategic management's new corporate policies would make workers and managers pay the costs of both past and present policies. Middle managers refused to execute important facets of the new managerial agenda, such as ranking, arbitrarily raising productivity bars, and obscuring the import of centralization and contraction, tools that would coercively cut the numbers of employees in the firm and alter a historically paternalistic set of employment policies.

Nor does the dispute over means constitute evidence of the irrationality of middle managers, their self-interest in historically accumulated power, or their fear of conflict, as many management theoreticians and consultants have been quick to claim. It is, rather, evidence of a logical response to the constraints that different levels of managers faced in the form of production politics and resources, for which no quick or superficial turnaround solution exists. It is also evidence of a strong rationality, insofar as middle managers perceived that their actions had the power, at the simplest level, to keep operations running.

Managers rejected top management's particular definition of entrepreneurial management-using their judgment to downsize the firm-but they did use their judgment extensively to reshape their own workplace. Whether innovating group work processes (as in the branch system), measurement and evaluation (as in SystemsGroup), or the production process itself (as in the card center), middle managers actively wielded their expertise in a set of oppositional management practices, both to organize the consent and cooperation of employees in their units to achieve long-term productivity objectives and to defend their units from the coercive aspects of the new agenda.

A one-sided interpretation of middle management fails to consider the autonomy and expertise lodged in the ranks of middle levels of managers and thus cannot adequately comprehend the limitations of top management's control over middle managers. Nor does it ask how managers organize production (specifically, within what constraints middle managers must maneuver) or how managers organize cooperation and consent. This study suggests that middle managers cannot be seen solely as agents of capital and top management and that the terrain on which managers organize consent may become a terrain of struggle within corporate management. It also suggests, contrary to the notion that managers resist simply to protect their turf or to stave off any kind of change, that managers undertake change, but it is not necessarily of the sort that top managers desire. Further, managers may have substantial reasons for refusing to manage out their own functional expertise.

Middle managers' capacity to respond to the new agenda differed according to "sectoral differentiation" (Baron and Bielby 1980), details of which are summarized in Table 1. The branch sector was being dramatically transformed: as production processes and the control over them were rationalized and externalized, the domain of middle managers' responses was restricted to struggling within the branch to organize existing resources and personnel. Because branch managers' control over ranking processes had been centralized, the tool of ranking did not become an object of struggle as it did in SystemsGroup. Branch managers' efforts were thus devoted almost exclusively to organizing and mobilizing branch employees.

Card center managers similarly worked to reorganize and mobilize workers to higher productivity. Card center

managers, however, resorted to offensive strategies, unlike the defensive actions of branch managers; they protected their jurisdiction as managers and in turn used their ability to maintain autonomous positions with respect to divisional-level management as a weapon for eliciting cooperation from card center employees. These managers controlled the productivity numbers and used them to forge an even more autonomous vantage point for managing.

Part of that strategy revolved around one critical variable: discretion over production-process innovation and production decisions had been moved from branch managers' jurisdiction to area management, whereas it had been enlarged for card center managers. Branch changes were imposed from outside the branch unit and were of a comparatively unequivocal nature. In the card center, section managers dictated the course of organizational change, acting on the autonomy of their new position and their knowledge of production processes. The organizational space in which to defend autonomous management practices was therefore different in these areas. Whereas branch managers' negotiations were centered on their own work sites, card center managers' negotiations were vertical, directed toward higher levels of management.

One more critical variable is needed to explain the ability to protect autonomous management: the functional importance of the different areas. Strategic management targeted the credit card center as an increasingly significant profit center, whereas it targeted the branch system as a less significant source of bank profitability. Profitability considerations per se did not affect group systems managers in SystemsGroup. SystemsGroup's functional role in the restructured bank was, like that of the card center, a critical one. But even in SystemsGroup-a research and development division that in noncrisis conditions may have been immune to cost cutting, rationalization, and centralization-the autonomy and decision-making processes of middle managers themselves were targeted for greater control. For SystemsGroup managers the evaluation and compensation process became the domain of struggle.

Like the card center managers, SystemsGroup managers had a degree of leverage for organizing consent and cooperation, but they did not have a role in reorganizing production, and they lacked detailed productivity figures to fend off top management and elicit consent from below. Measurements of actual SystemsGroup output were minimal compared with the measurable work processes of the card center. But the main target in SystemsGroup was the total wage bill of the division, and ranking was a primary mechanism for slowing down wage increases and eliminating allegedly superfluous workers. We can best understand group systems managers' interests in reranking as a response to being targeted as the agent for cutting the wage bill.

In part, managers' resistance was based on efficiency considerations. They were under scrutiny themselves, and their ability to achieve new productivity quotas depended on whether they would be capable of gaining consent to new objectives. Their concerns for increasing productivity, however, were tempered by a set of considerations about consent and about past and present conditions of the workplace that strategic management would have preferred to ignore.

In this regard, middle managers' resistance to the coercive aspects of the new agenda even benefited workers in the restructuring process. Some may want to argue, in contrast, that the ways in which managers managed, documented in the foregoing three chapters, simply evidence coercion and not protection of workers. Inducing workers to throw their weight behind middle managers' objectives obviously involved some degree of manipulation, a manipulation made all the more inequitable by the fact the relationship between managers and the managed was fundamentally inegalitarian. But managers' oppositional practices, and their effects on workers, had a decidedly more complex character. The point is not that managers were, or even necessarily perceived themselves as, the humane arbiters of corporate change. Rather, in defending their own domain of expertise, to further their particular vision of what best served the corporate interest (a vision that included a unique interpretation of the social relations of the workplace), managers also defended workers from the more extreme aspects of strategic management's productivity and downsizing demands.

Strategic management's vision of middle managers as the agents of their own demise and of the downgrading of their employees neglected a simple and real fact: these managers played a critical role in organizing other managers and workers to achieve productivity objectives, and top management faced severe constraints in turning that relationship on its head. Middle managers do not simply collect and process information up and down the management hierarchy; nor are they a mechanical apparatus of control that simply and mysteriously makes employees conform to rules and regulations.

My research thus challenges another orthodoxy about American middle management: the notion that middle managers can be automated out of existence as their "information processing" function is absorbed into computerized systems. They play a legitimating role, combining technical and social relational expertise; it is they who must absorb the contradictions of restructuring, and they deploy their unique knowledge about specific production sites to do so.

Managers' resistance was not only pragmatic. It was political, in that managers were angry about strategic management's handing them a set of tasks that could not compensate for the real and fundamental problems of the firm. Their belief that layoffs should be initiated (even knowing their own positions could be in peril); that top management's failure to modernize the firm and invest in certain technologies had hurt the firm's prospects for profitability; that strategic management was adopting Band-Aid policies by forcing all managers to attend "corporate culture" training, were all ingredients of an anti-strategic-management politics. In that politics, although middle and strategic managements may share a vision of corporate objectives, significant differences exist over the appropriate means for achieving these ends.

The specific content of coercive autonomy revolved around the attempt to deploy managerial discretion to the goal of corporate downsizing and restructuring and a new regime of intensified productivity and savings. Strategic management could order managers to manage out as a form of disguised cutbacks, but the micro-level execution of techniques to achieve cutbacks resisted precise specification or rationalization. Middle managers consequently were exhorted to take more aggressive action, to use their managerial judgment in an autonomous fashion. Nevertheless, as we have seen, the exercise of judgment, innovation, and entrepreneurial behavior were subjected to strict surveillance.

In proffering the strategy of coercive autonomy, strategic management was proposing a new solution to a historically variable problem: managing management, or managing the domain of action that rests between the firm's need for centralized control (necessary because of increased size and complexity) and its need to extend discretion (necessary because of the imperative to delegate authority) (Bendix 1956, p. 336). Under the particular conditions of profit squeeze, contraction, and restructuring, top management both scrutinized and exploited discretion in a historically unprecedented way.

But discretion is not an autonomous capacity that workers or managers carry around with them to exercise uniformly in any production context. Rather, discretion is shaped and developed within specific work contexts; it is valuable because it is developed as a "way of knowing" over time. Thus discretion cannot be codified. Furthermore, production processes and relations at the lower levels of the firm cannot be arbitrarily and abstractly manipulated without regard to prior organizational practices. In American Security Bank, middle managers toiled not merely to preserve traditional practices but to build on established practices to create the conditions for a renewed corporation. This is certainly one dimension of the corporate interest that remains undertheorized.

In the corporate agendas that emerged in the 1970s and the 1980s, corporate leaders would ideally use middle managers to transform the employment and production structures of contemporary American industry. In the current international context of competition and contraction, however, many of the real solutions to profitability rest in more fundamental and enduring restructuring policies. In an important way, strategic managements are facing the consequences of decades of specific capital-maximization strategies.

Whether or not top managers will accept the responsibility for these consequences remains open. American Security's strategic managers attempted to decentralize and obscure the responsibility for remedying one of the major problems of the bank. But middle managers' criticism of strategic management ultimately forced top management to adopt more direct, visible means for downsizing and cutting back. Strategic management's failure to gain managers' compliance to absorb the costs of restructuring was evidenced in the progression of events over 1986 and early 1987.

By 1986 there had been little significant reduction in the ranks of American Security Bank's employees. Reductions proceeded very slowly; reports indicated that only 8,000 employees left the bank over a four–five-year period as a result of the combined forces of redeployment, normal attrition rates, the sale of company assets, and retirement. Much speculation circulated about strategic management's failure to scale down American's operations dramatically. Some commentators even suggested that Wedgewood was constitutionally incapable of taking deep cuts in the corporation.

In fact, Wedgewood attempted something quite bold by banking on the thousands of managers below him to get rid of allegedly superfluous employees. Wedgewood and his top committee threw their weight behind the "managing up or out" strategy, behind culturally reforming the managerial style to be more productive and aggressive.

Those broad reform efforts were overshadowed by the bank's mounting loan losses. Strategic management was forced to bolster American Security's loan loss reserves, thus lessening profits. Even the sale of properties and entire subsidiaries to reduce the effect of profit loss through one-time capital gains failed to halt declining earnings.

American Security's situation was further complicated when the Federal Reserve Board and the Comptroller of the Currency stepped up their intervention in American's affairs.^[2] As a result of pressure from federal agencies, American Security was unable to use lending as a growth tool. The regulators refused to allow American Security Bank to expand loans on its equity base, ordering strategic management instead to redirect capital into building the equity base and loan loss reserves to a higher level, protecting the bank and bank depositors against accumulating losses.

Even more significant, the Comptroller pressured American Security Bank to reclassify many more loans as "nonperforming," resulting in increasing loan write-offs. Spurred by earlier bank collapses, federal authorities targeted American Security early on as potentially troubled. Its basic core of deposits and good loans were viewed as solid, but the breadth of American Security's financial influence meant that all authorities were on alert for damaging economic effects of the loan losses.

Finally, after months of public and private skepticism about the quality of American's management, the board of directors came under intense pressure to make a change in strategic management, whether the effect was symbolic or real. Wedgewood left the bank in the mid-1980s.

After his departure, strategic management was forced to announce that they would quickly undertake widespread layoffs of up to 5,000 employees. Regional newspapers carried front-page headlines about the future layoffs, and less than a year later, strategic management reported that they had in fact exceeded the originally anticipated reductions. They had aimed, in that one-year period, to cut staff by 5,000 and had eliminated over 6,000.

There is evidence, although unofficial, that even here American Security's strategic management did not directly cut thousands of workers from its payroll. Projections of future cuts notwithstanding, actual large layoffs apparently never materialized. According to the banking correspondent for the regional newspaper, much of the attrition at American Security Bank resulted from the combined effects of the fear of layoffs and the active promotion by the bank of a severance package. Significant numbers of employees had been driven to take advantage of the package rather than wait for notice that their position would be eliminated.^[3]

Managers' resistance to managing out did not cause this eventual outcome. In other words, just as middle managers' practices cannot be blamed for American Security Bank's decline in the early 1980s (despite strategic management's attempts to scapegoat middle management as the major impediment to productivity and competitiveness), neither can their responses to the corporate restructuring processes be blamed for the more recent dramatic events of the bank: asset sales, vast scaleback of its operations, large profit losses. Their reactions did, however, have an important constraining effect on the domain of top management action. By refusing to accept responsibility for what they perceived as an unjust and inviable agenda, American Security's middle managers eventually forced strategic management to take greater and more direct responsibility for the painful task of restructuring its own labor force.