Christopher A. Bartlett
Professor of Business
Administration
Harvard Business School
Sumantra
Ghoshal
Professor of Strategy and
Management
INSEAD
Chandler's study of the adoption of the multidivisional organization structure by fifty of the largest companies in the United States, and his detailed examination of four of the pioneers of that revolutionary structural form, led him to develop his highly influential strategy/structure thesis. His central conclusion was that companies being driven by market growth and technological change to develop greater diversity in their products and markets, were able to manage their new strategies efficiently only if they adopted a multidivisional organizational structure - the so-called M-form.
The reason this structural form proved so powerful was because it defined a new set of management roles and relationships that emphasised the decentralization of responsibility to operating divisions whose activities were planned, coordinated and controlled by a strong corporate management - the general office in Chandler's terms - which also made the company's "entrepreneurial decisions" about resource allocation. He showed how the management process created by this organization allowed companies to apply their resources more efficiently to opportunities created by changing markets and developing technologies.
Before highlighting the differences, we should note that many of the attributes of ABB's organization, management and even its change process fit well with the model proposed by Chandler. First, the process by which the company's new organization and management approach was designed and implemented parallels his description of changes in the companies pioneering the M-form. In both Asea and Brown Boveri - the two companies that were subsequently merged to form ABB - an existing management of "insiders" had been unable to change the structure sufficiently to create efficient operations out of their respective diversified portfolios of electrical and power-related products, blocked by what Chandler described as "the psychological hazards of adjusting to new ways, new tempos, and new duties" (1962:320). Yet both companies had the rational, analytic and engineering-driven "institutional ethos" that Chandler had found conducive to organizational change. More importantly, in Asea there was a top level manager with exactly the profile of Chandler's classic organizational change agent - a young, analytically inclined outsider. Brought in initially to turn around the troubled Swedish company, the 39 year old Percy Barnevik had attracted considerable attention when, within six years, he had increased Asea's sales four fold and its profits by a factor of ten. He was the natural candidate to lead the company created by Europe's biggest ever cross-border merger.
At the broadest level, some of the structural elements on which Barnevik built ABB's new organization follow the design principles of the classic M-form. At the corporate center, he replaced the functionally oriented staff groups that had dominated the headquarters of the two founding companies with a team of capable group executives supported by highly specialized staffs. More important, the strong principle of decentralization of responsibility that Chandler saw as central to the operation of his model, is also central to ABB's management philosophy. As defined in the company document, ABB's Mission, Values and Policies, "our guiding principle is to decentralize the Group into distinctive profit centers and assign individual accountability to each." (p. 25).
Finally, a core element of ABB's approach for coordinating and controlling its diverse operations is in keeping with Chandler's prescription for creating reliable information and data flows to support the lines of authority. The fully automated ABACUS system (Asea Brown Boveri Accounting and CommUnication System), designed and installed within twelve months of the merger, is the centerpiece of the formalized information flow. It not only provides accurate and timely data to the field operations, it is also explicitly designed to help the group executives evaluate performance. As the company's much used Mission, Values and Policies document states, "A decentralized organization will only work effectively with a good reporting system that gives higher level managers the opportunity to react in good time." (p. 42).
While the broad principles of delegation and control have their roots in the organizational model Chandler described, the structural form by which they are implemented begins to diverge from the classic M-form. ABB is structured not in the traditional multidivisional form, but around a business/geography matrix. Figure 1 illustrates this structure. Don Jans, the general manager of ABB's relays business in the United States, reports to both Ulf Gundemark, the Sweden-based head of the company's worldwide relays business area (BA in ABB terminology), and Joe Baker, who has overall responsibility for ABB's power transmission and distribution businesses (of which the relays business is a part) in the United States. Gundemark, in turn, reports to both B-O Svanholm, the Group Executive responsible for all of ABB's businesses in Sweden, and Goran Lindhal, another Group Executive heading the company's worldwide activities in the power transmission sector. Similarly, Baker reports to Lindhal on the business axis and to Gerhard Schulmeyer, the Group Executive responsible for all ABB operations in North America, on the geographic axis.
Yet a closer examination of this structure, by now quite common among large global corporations, reveals it to be simply a more complex form of the same basic divisionalized model described by Chandler. As such, it probably serves more to confirm the underlying strategy- structure thesis than to demonstrate a substantive difference. In an industry environment driven by technological and competitive forces requiring companies to be globally integrated, while market and governmental pressures demand a more nationally responsive approach (Prahalad and Doz, 1987), ABB's strategy requires it to manage both product and geographic diversity. To do so, it has structured its organization around a matrix with dual reporting channels linking front-line operating units to their global business area organization on one side and to their national or regional managements on the other.
Despite these similarities, however, there are some core elements in the structure and management of ABB's organization that clearly differentiates it from the model Chandler described. Most immediately obvious are the differences in the extent of decentralization of assets and delegation of responsibilities that both reflect and reinforce a philosophical divergence concerning the locus of entrepreneurship within the organization. It is this philosophical difference that lies at the heart of ABB's structural divergence from the classic M- form.
Barnevik has organized ABB around a principle he describes as "radical decentralization". This is more than just a difference of degree from the basic philosophy underlying the creation of the divisionalized organizations at du Pont or General Motors. It is a fundamental difference that is rooted in a very different organizational logic. The ABB organization is based not so much on sub-dividing the corporate entity into separate divisions to be directed and coordinated from the top, as it is on a concept of providing "self contained and manageable units with overview." Translated into action, this philosophy has resulted in an organizational structure and management system in which the basic building block is not the company's seven business segments (groups in Chandler's terms), nor even its 65 business areas (the equivalent of Chandler's divisions). Instead ABB is organized as "a federation of companies" - 1300 of them that are structured as separate and distinct businesses and, to the extent possible, as free-standing legal entities. The U.S. relays business headed by Don Jans (see figure 1) is an example of such a company. On average, each of these companies employs 200 people and generates $25 million in annual revenues. In creating such a large number of small entities, Barnevik hoped that employees would lose "the false sense of security of belonging to a big organization" and would develop the "motivation and pride to contribute directly to their unit's success".
Equally striking, and in marked contrast to the classic M-form organization, is the extent to which ABB's heirachical structure above the individual company level has been stripped to a leanness that Chandler would not recognize. In contrast to the eight or nine layers of management in its predecessor companies, in ABB there is only one intermediate level between the corporate executive committee and the managers of the 1300 front line companies. Also, the staff groups supporting that shallow structure are very thin. Barnevik's rule of thumb for restructuring the traditional organizations of the merged and acquired companies from which ABB was created was to remove 90% of all employees at each level above the operating companies. Those individuals or groups that were not required by the companies to increase their self sufficiency were either set up as independent service centers charging market rates or eliminated. As a result, the entire corporate headquarters of this $ 30 billion company including the CEO, the seven group executives (such as Lindhal and Schulmeyer) who comprise the executive committee, and the various corporate staff groups together number less than 100 people. Similarly, at the business area level, managers like Ulf Gundemark are supported by a staff of three to five - typically a controller, atechnical director and a business development or marketing director, none of whom have any assistants.This change in basic philosophy to radical decentralization is supported by an equally radical redeployment of human, technological and financial resources. With the corporate HR staff consisting of only one manager, human resources have to be recruited and developed at the level of the front-line companies. Similarly, rather than concentrating its technology in centralized corporate laboratories, or even at large group or division-level research facilities, ABB allocates more than 90% of its substantial R&D budget ($ 2.4 billion in 1992) to support technology development in the front-line companies' centers of excellence whose expertise are then linked with those of centers located in other companies and leveraged broadly across the entire organization.
Finally, each front line company is given responsibility for its complete balance sheet, and it is explicit corporate policy to allow managers to inherit results over the years through changes in their company's equity. As a result, each unit manages its own treasury function including cash management, foreign exchange exposure and responsibility for borrowing. Coupled with a dividend policy that permits each company to retain a third of it's net profits, this policy gives front-line management substantial financial independence by limiting their need to rely on corporate management for funding.
Barnevik's primary objective in making such a radical change to the classic multidivisional structure he inherited was to create a context in which the entrepreneurial activities could be much more widely distributed that in most contemporary companies or indeed in the classic model Chandler described. In Chandler's model, the corporate level executives were indisputably "the key men in any enterprise" due to their central role as resource allocators. It was this role that led Chandler to dub them the company's "entrepreneurs" while those lower in the organization were termed "managers" and were assumed to focus on the operating activities required to implement the corporate executive's entrepreneurial decisions. Through the radical decentralization of resources to 1300 front line units and through the dramatic reductions in the machinery of control, Barnevik has redefined these management roles to create a very different entrepreneurial process that represents the first level of ABB's evolution beyond the classic M-form model.
This entrepreneurial process, framed and institutionalized by ABB's new organization structure and management philosophy, is embedded in a redefined set of management roles and relationships. Front-line managers - the heads of the 1300 little companies - have evolved from their traditional role of implementers of top-down decisions to become the primary initiators of entrepreneurial action, creating and pursuing new opportunities for the company. Middle-level managers like Baker and Gundemark are no longer preoccupied with their historic control role, but instead have become a key resource to the front-line managers, coaching and supporting them in their activities. And top management, having radically decentralized the resources and backed them with strong delegated responsibility, focus much more on driving the entrepreneurial process by developing a broad set of objectives and by establishing stretched performance standards that the front-line initiatives must meet (see figure 2).
- Figure 2 about here -This fundamental difference between the ABB organization and the classic M-form is brought sharply into focus with the changes that occur in organizational processes and management roles every time ABB acquires a company operating in the more traditional mode. With its acquisition of Westinghouse's North American power transmission and distribution business for example, ABB had to overlay its radically decentralized structure and its philosophy of front line entrepreneurship on a much more hierarchically controlled management system. The impact on the management of the relays division were typical of the changes occurring across this whole acquired business. Because of Westinghouse top management's unwillingness to support investments in what was seen as a mature business, Don Jans - then heading the relays division for Westinghouse - had been unable to convince his bosses to invest to any significant degree in solid state and microprocessor relays technologies, despite the fact that his aging electro-mechanical product lines were coming under incresing competitive threat. Under ABB's ownership, however, management roles and relationships changed radically. The relays operation was structured as a separate company and was expected to take responsibility for its own long term future. Jans, whom ABB retained as the general manager of the company, quickly used his new-found strategic freedom and financial flexibility to start a small-scale effort to develop microprocessor-based products. As the initial experiments yielded some positive results, the scope of the activity was gradually broadened and the company was able to develop innovative new solid state products in time to head off an aggressive new competitor who had begun to take advantage of the product-market gap.
The stagnation occuring in Westinghouse's power transmission businesses is not unusual. As has been documented by several observers of large global companies, the hierarchical and staff-led processes in the mature M-form organization tend to dampen front-line initiatives and innovation (Kanter, 1983; Peters, 1992). Evidence over the 1980's suggests that what Williamson (1975) described as the corrupted multidivisional organization is perhaps not an exception but the state to which the internal dynamics of the M-form inevitably leads (see also Chandler, 1991).
One way to prevent this erosion of front-line entrepreneurship is represented by the conglomerate organization labelled by Williamson as the holding company model or the H-form. While the entrepreneurial process developed in ABB is explicitly designed to counter the M-form's tendency to reduce front-line managers to the role of operational implementers, it differs significantly from the management processes of the H-form primarily because of the roles that are played by the company's middle and top managements. Unlike in conglomerates, both these groups in ABB remain involved in the operational realities of the businesses and contribute directly to the entrepreneurial process. Their actions not only differentiate them from the hands-on substantive decision-making role of the M-form, but also from the hands-off financial control role of the H-form.
The role that Gundemark and Lindahl played in shaping and supporting Jans' initiative can illustrate the point. As head of ABB's worldwide Relays Business Area (BA), Gundermark saw his role as supporting Jans' initiative on microprocessor based products, incorporating it into the BA's global strategy to lead this technological conversion. As a member of the steering committee that acted as a board for Jans' local company, Gundemark was able to ensure that the proposal was supported and funded within a few days rather that the months or even years it would have taken in Westinghouse for the capital request to work its way up to corporate headquarters for approval and back. Beyond funding approval, Gundemark provided Jans with advice and support in pursuing a strategy that involved a risk of short term profit decline. For example, when a U.S. revenue shortfall led Jans' geographic boss, Joe Baker, to push for cuts in the $1.5 million Jans had budgeted for microprocessor product development, Gundemark arranged to provide the technical support from the Swedish relays company.
At the corporate level, Lindahl worked at building a shared commitment from his management team to the ambition of "conquering the globe in power transmission". Expecting his BA heads and company managers to take the leadership in developing their business strategies, he saw his role as questioning and challenging the robustness of those strategies by testing them against key issues like environmental legislation or asking them to develop scenarios for dealing with different trade barriers. Beyond such macro exercises, ABB's top management also lived by the motto "what gets measured gets done". Yet Lindahl saw his role as much more than defining financial targets; his primary focus was on embedding performance standards that would stretch the organization to achieve extraordinary results. This led him to adopt a style he described as "fingers in the pie" management, which he contrasted to the "abstract management" approach of controlling agglomerated units through sophisticated but remote systems. Managing in this way, he had no problem in making direct contact with Jans or other front-line managers to expand their horizons, encourage their initiatives, or offer help when performance was slipping off-track.
The review, support and coaching provided by experienced managers like Gundemark raises the levels of discipline, analysis and judgment in the entrepreneurial initiatives of front-line managers like Jans. Similarly, the strategic mission and performance standards defined by top managers like Lindahl ensures that those initiatives are aligned with the company's overall strategic priorities. This discipline, guidance and support provided by ABB's senor management groups was clearly less constraining than the much more direction- and control-oriented roles played by similar management levels in Westinghouse. At the same time, however, they were much more involved than the remote, measurement-focused management that had allowed front line entrepreneurs to lead many H-form companies into their incoherent expansions in the 1960s and 1970s (Chandler; 1991). These changes in the roles and tasks of senior managers not only distinguish the entrepreneurial process in ABB from those in both divisionalized and conglomerate organizations, they also lead to a very different process for integrating the knowledge, resources and capabilities lodged in different parts of the company. It is to an analysis of this integration process in ABB that we now turn.
Bower's intensive study of business planning and investment decision making in four divisions of a large diversified company he identified as "National Products" led him to develop a process model of the firm that enriched Chandler's work by challenging and expanding on several of its descriptions and conclusions. He presented a less heroic view of top management than Chandler's description of corporate entrepreneurs determining strategy through their control over resources. Instead, Bower described a process in which the shaping of new strategic initiatives and the investment proposals to support them ("definition" in his terminology) were initiated by front-line managers. Furthermore, in his view, it was the middle-level managers who typically made the resource commitments, since projects reaching the executive committee level with their support (Bower used the term "impetus") were "almost never rejected" (1970:57). Thus, the major source of top management's power in Bower's model lay in its control over what he termed the "structural context" - "the set of organizational forces that influenced the processes of definition and impetus" (1970:71)
.This model of the core management processes in a large, complex organization provides a good yet partial insight into some of the management roles and relationships in ABB. Certainly Percy Barnevik's priorities as he set the course for the newly merged operations were very much oriented towards the context-defining role that Bower described as the top management job of "constitution writing" (1970:2). Within days of the merger on January 5, 1988, Barnevik convened a three day meeting of his top 300 managers in Cannes to describe in great detail the operation of the matrix structure, the management philosophy of radical decentralization with strict accountability, and the behavioral expectations based on speed, flexibility and an action orientation. His presentation, communicated through 198 overhead transparencies, was formalized in a widely distributed 21 page Mission, Values and Policy booklet referred to inside the company as the "policy bible". During the years that followed, he and the other members of the corporate executive committee continued to spend an enormous amount of time elaborating on the management philosophy, refining the structural model, and developing the corporate policies.
Goran Lindahl, member of the corporate executive committee, characterized the resulting management philosophy as one of "decentralization under central conditions". In this system, he described the top management role as being "to provide a framework" within which those lower in the organization could operate and make decisions. It is a philosophy that closely parallels Bower's view that top management's primary role is to define and manage the organization's "structural context".
Similarly, at the front-line management level, Bower's description of managers who were the primary developers of new strategic ideas and investment proposals also rings true in ABB, and detailed evidence of this was presented in the previous section. As the "policy bible" confirms, "ABB's improvements are generated by thousands of actions taken by both large and small profit centers... Every ABB manager must be a driving force for change and development" (P18). One manager described his role running a profit center as "white water rafting management", but was clear that it was his initiative that drove the agenda of his two matrix bosses. Only partly in jest, he declared, "We are the master of two slaves!"
It is at the critical level of the middle-level manager - like Don Jans' "two slaves", Joe Baker, and Ulf Gundemark - that the practices at ABB differ the most from Bower's process model. In Bower's observation, as in Chandler's, the growing size and complexity of organizations required middle managers to focus most of their time and effort on managing the two vital processes of business planning and resource allocation. The challenge at the center of the division manager's job was to resolve the problems caused by the asymmetry of information in large complex organizations by communicating the corporation's objectives and standards down to front-line managers and transmitting business needs and opportunities up to corporate level managers. As Bower explained:
Once the gap between the kinds of information, analysis and choice that goes on at the two extreme levels is recognized, it is obvious that some manager has to provide a consistent integration of the parts.... The detail necessary for conducting an evaluation makes it unfeasible for top management to accomplish the taskÓ (1970:288).But Bower recognized that this task involved more than being a two-way communcation conduit, and he saw the middle manager's key role being focused on the selection, screening and interpretation of information. Like Chandler, he also viewed this task as legitimizing the mutli- tiered, staff-supported hierarchies that were typical of the organizations both researchers studied. While Chandler described the primary role of staffs at department, division and corporate levels in analytic terms - "an independent check on requests, proposals and estimates" (1962:310) - Bower saw their role more from the political perspective that lay at the heart of his model. At the division level, staff played a key role in helping the general manager "evaluate projects against their understanding of corporate criteria for evaluating managers" (1970:78, emphasis added). Thus, the middle management challenge of "placing bets" and "building a track record" required them to focus not only on the content of the projects, but also on the process by which they were developed and the credibility and commitment of those proposing them. To ensure that project initiators would "defend their projects and plans vigorously" (1970:341), Bower's division general managers needed an expert staff whose primary role was "to perform the adversary role in the planning process" (1970:342)
The centrality that both Bower and Chandler gave to the resource allocation process in their description of the managerial task was appropriate in describing the organizations they were studying. In the rapidly expanding economy of the 1920s and again in the booming markets of the post-War decades, opportunities far exceeded many companies' ability to finance them. Yet, in today's highly competitive, technologically driven environment, the scarce resource that constrains the growth and strategic success of companies like ABB is not so much capital as it is specialized knowledge and expertise, and the organizational capability that embeds it within the company.
Unlike capital, knowledge is a resource that is difficult to accumulate at the corporate level and allocate according to top management's evaluation of strategic need. In ABB and companies like it, there is an increasingly clear recognition that those with the specialized knowledge and expertise most vital to the company's competitiveness are usually located far away from the corporate headquarters - in the front-line units' research laboratories, marketing groups or engineering departments. By decentralizing assets and resources into these small specialized operating units, these companies are trying to create an environment in which this scarce knowledge can be developed and applied most appropriately. However, this creates a greater need for a powerful horizontal integration process to ensure that the entire organization benefits from the specialized resources and expertise developed in its entrepreneurial units. Forced to free up its middle managers to create and support such horizontal linkages, ABB found it had to reduce the demands placed on them by the intensive vertical information processing tasks and the complex politically driven decision-making processes that were at the heart of Bower's model. This led the company to create a set of alternative organizational mechanisms to address the problem of heirarchical information assymetry and the need for information screening.
First, the problem of information assymetry is considerably alleviated by ABB's commitment to system-wide information sharing. As Barnevik described, "you don't inform, you over inform". It is a philosophy that has been implemented both through the company's formal information systems and its informal communication processes. At the foundation of the company's commitment to keep managers at all levels equally well informed lies ABACUS - ABB's sophisticated information system. Governed by strict rules concerning definition, format and timing, this democratic system ensures that managers around the company receive the same information at the same time regardless of their hierarchical level. Coupled with an organizational norm that puts value on managing content and not just process - reflected in Goran Lindahl's "fingers-in-the-pie management" - this information sharing approach has created a context in which top management continuously remains in touch with front-line operations, thus reducing the need for middle management to constantly play its upward intermediating role.
Similarly, the company's intensive informal communication processes greatly reduces middle management's need to close the information gap in the other direction. Starting with the mission, values and policies detailed in clear, direct and often tough language in the "policy bible", senior management sees the communication, interpretation and elaboration of ABB's objectives, priorities and philosophies to the lowest levels as a core task. Barnevik sets the standard and provides the role model for this activity, meeting annually with about 5000 managers with whom he discusses corporate objectives and priorities which he communicates using some of the 2000 overhead transparencies he carries with him. Again, the impact is to reduce the burden on middle management to ensure that corporate objectives and standards are properly transmitted down through the organization.
Beyond managing information assymetry, ABB has had to develop mechanisms to provide an alternative to the vertical information screening process provided by the successive layers of staff- driven challenge and refinement in Bower's model. Compared to the multi-level hierarchies where investment proposals could be defined by managers "seven or eight levels below the corporate president" (Bower, 1970:9), the ultra lean ABB structure has only one level between the executive committee and the front-line managers running the independent operating companies. As a result, middle level managers like Gundemark typically have 10 to 20 companies reporting directly to them, yet must manage this wide span of control without sufficient staff support to provide the independent expertise and adversarial challenge that drove the information screening process in Bower's model.
To replicate the discipline and refinement of the M-form's challenge and evaluation processes, ABB has built into its structure a sophisticated system of internal tension that achieves a similar effect. For every key decision, from establishing strategic priorities and setting budget targets to making key investments and restructuring business operations, managers must present, defend, and adjust their proposals to meet the often conflicting interests of two matrix bosses. Through such a process, ABB subjects the overall planning and resource allocation decisions to a filtering and refinement process at least as powerful as the staff driven challenges of the traditional model. Equally important is the company's extensive use of self-monitoring and self-regulating mechanisms. For example, BA managers routinely create "performance league tables" ranking all their businesses' operating companies according to the measures or standards that reflect the priority objectives. This creates an environment of internal competition that requires front-line managers to measure their performances and defend their proposals against the standards set by the best performers within the BA.With their vertical information processing tasks reduced through such mechanisms, middle managers at ABB are able to focus much more of their attention on the horizontal integration tasks. In contrast to the planning and budgeting dominated role they played in Bower's model, this management group spends considerably more time and effort managing activities such as internal benchmarking, best practice identification, and technology transfer - all aimed at linking and leveraging the company's widely distributed resources and capabilities. The development and management of this intensive horizontal integration process is the second major characteristic distinguishing this organization from the classic M-form.
Although this vital capability-building and -leveraging task is driven by the middle managers, it requires intensive interaction with and support from managers at all levels of the organization (see figure 3). In ABB, the middle managers' pivotal horizontal linkage role is supported by a top management that creates a value-based context to support and reward collaborative behavior, and by a front-line management that exploits the personal networks such horizontal relationships facilitate. It is a very different process than the vertical planning and resource allocation process at the core of Bower's model, but a vital one for a company competing in an increasingly knowledge- based economy.
- Figure 3 about here -At the top level, for example, ABB's group executives devote enormous effort to create a sense of shared organizational identity - what Barnevik calls a "glue" - to bind disparate efforts, as well as organizational norms that value collabaration - a "lubricant" in Barnevik's terminology - to facilitate the linkages that intensive knowledge transfers require. Without such an organizational context, the centrifugal forces driving independent entrepreneurial units can quickly result in fragmentation, isolation and inter unit competitiveness to create barriers and defenses against internal flow of resources, knowledge and expertise.
Consequently, the first section in ABB's statement of its values is listed as "Corporate Unity". Elaborating on this core value, the "policy bible" defines clearly the expectation that individuals and groups would interact "with mutual confidence, respect and trust... to eliminate the we/they attitude,... and to remain flexible, open and generous". Barnevik and his colleagues in the Executive Committee carry the primary responsibility for translating those values into action through their selection and promotion of individuals reflecting such behaviors, their imposition of sanctions against those violating the values, and through their own role modeling actions.
These strongly ingrained corporate norms and values have created an environment which encourages front line managers to develop personal networks and operating interdependencies, and rewards them for doing so. Their regular horizontal contacts across formal organizational boundaries create spontaneous transfer of knowledge and expertise that has become a valuable capability of the ABB organization. The linkages are well illustrated by events following the introduction of a time-based management program by the US relays company as a means of increasing its overall competitiveness in the market. This innovative program, which required intensive cross functional coordination within the company, was highly successful not only in increasing customer service levels but also in reducing working capital at the same time. At a regular meeting of quality assurance managers held in the premises of the U.S. relays company, the hosts conducted a plant tour as a side event, explaining their just-in-time inventory system and other changes made to reduce throughput time. Without any corporate directive, this program was transfered to almost every other major relays company in ABB within a year, solely at the initiative of the functional managers and their informal networks.
However, ABB and companies like it cannot rely on such "spontaneous combustion" to drive the itensive knowledge sharing required in most modern corporations. While the top level context setting and the front line personal networks can provide the enabling conditions for his vital horizontal process, it is the middle managers who are the best placed to facilitate these cross-unit linkages. Yet, historically, it has been at this level of the organization where the highest barriers to horizontal transfer of knowledge and expertise has existed, largely due to the structure and management philosophy behind the design of the M-form. By creating divisional managers to carry out the cross functional integrating task, this successor structure to the functional or U-form organization allowed the top management to devote themselves to the broad corporate-level strategic priorities. In this process, however, the M-form fragmented the company's functional capabilities, while simultaneousely creating a norm of divisional autonomy that constrained the reintegration of those capabilities across the divisions.
The integration process framed by ABB's organization structure and management philosophy attempts to overcome this problem of fragmentation of capabilities by avoiding the pitfalls of both the U-form's centralization and consolidation of functional departments and the M-form's decentralization of functions to ensure divisional self-sufficiency. Instead of dividing the company into a small number of independent divisions, ABB has structured itself into 1300 small companies none of which can ever expect to fully develop or control all required functional capabilities itself. Freed of many of the demands of managing the intensive vertical planning, control and resource allocation processes, and equipped with an intimate knowledge of all the operating companies, the BA managers in ABB are ideally placed to devote substantial time to managing the resulting horizontal interdependencies.
For example, in order to leverage the knowledge and expertise developed in various entrepreneurial units around the world, Gundemark has created a series of additional communication channels and decision making forums at several levels of the organization. At the lowest level, he has organized functional specialists from the front-line operating companies in the U.S., Sweden, Finland and Germany into various functional councils, encouraging them to transfer best practice from the leading edge units to the others. At the next level, he has formed a steering committee for each operating company, nominating to these local boards individuals from the BA management team, the regional staff and others within ABB who could provide the local company with relevant advice or experience. And at the business area level, he has created a BA board with membership drawn from the general managers of his key operating companies, to pool their expertise in shaping the BA's global strategy, management policies and operating objectives. Each of these new structural mechanisms has been managed to maximize its effectiveness as a cross-unit communication channel or decision making forum, and the company has dramatically increased its use of internal benchmarking and transfer of best practice.
The creation of this broad portfolio of task forces, teams and committees have also had a broader impact on the organizational and management processes: they have served to develop management perspectives and relationships in ways that have helped to prevent isolationism and to break down parochialism, and they have become forums in which managers can negotiate differences and resolve conflicts that are inherent in the matrix structure. These changes in management motivations and behaviors and their consequences for the organizations' goal-setting and learning processes come into sharp focus in the following analysis of ABB from the perspective of Cyert and March's behavioral theory of the firm.
Using clinical field data, experimental studies and computer simulations, Cyert and March formulated a behavior-based theory of the firm that challenged many of the economists' classic assumptions. By viewing organizations as coalitions of participants with disparate demands, they developed a notion of goal formation in firms that was based on an internal process of bargaining among coalition members. The objectives developed through this process were given stability by internal control mechanisms, particularly by standard operating procedures, yet were adaptable to changes in the external environment and to internal changes in the coalition. To explain decision making behavior, Cyert and March developed four key concepts about the way in which firms set objectives, manage expectations, and make choices. In their model, behavior was driven by the quasi-resolution of conflict, uncertainty avoidance, problemistic search, and organizational learning.
In a global corporation that has very consciously divided itself into 1300 operating companies which are further fragmented into more than 3500 profit centers, the notion of an organization as a coalition of diverse interests has a good deal of face validity. And, particularly with regard to the setting of performance targets, Cyert and March's description of organizational objectives being established through a bargaining process provides an accurate description of ABB's system. On the basis of each business' global strategic plan, ABB's executive committee negotiates macro- level targets for each BA and each geographic area, defining them in terms of broad measures such as growth, profit and return on capital employed. As these macro targets are transmitted through the organization, each BA and regional manager allocates to each operating company objectives that are defined in terms of more micro-level targets such as new orders, revenue, profit, working capital and headcount. Operating in the matrix, a company manager typically receives different and sometimes conflicting objectives from a BA manager whose priorities may be to maximize global market share and to rationalize worldwide sourcing, and a regional manager who may be working to increase short term profitability and protect local employment. Forced to agree on a single budget, these three managers then engage in a negotiation that exhibits many of the characteristics of quasi-resolution of conflict. They establish goals on the basis of local rationality, using decision rules based on satisficing rather that maximizing objectives.
Similary, ABB's management seems to recognize very clearly and explicitly that the tensions in its fragmented organizational system needs a strong set of stablizing mechanisms to reduce complexity, guide action, and control behavior. Rules of behavior and guidance on decision making make up a large part of the company's "policy bible". As Cyert and March suggested, these clearly defined "rules of thumb" and "standard operating procedures" affect the goals and perceptions of individual ABB managers, as well as the alternatives they consider and decision rules they use in the company's organizational choice process.
For example, it was largely because of its detailed and clearly understood policies and procedures that the company was able to make a large number of vital rationalization decisions in a concentrated two year period. The speed with which the decisions were made was even more impressive since issues such as the allocation of production capacity and export markets among the companies that had historically competed against one another were obviously highly divisive in nature. Such strategically important and politically sensitive decisions could be agreed on and implemented so quickly only because there was a clear and detailed set of rules laid out in the "policy bible" relating to such matters as the full utilization of existing facilities before proposing new investments, the principle of locating investments as close as possible to long term markets, and the mechanics of determining internal sourcing patterns and transfer prices.
Finally, the dynamics underlying ABB's demonstrated market responsiveness and technological adaptability can be captured, at least partially, in Cyert and March's notion of organizational learning. There are scores of examples of the way in which the firm has adapted its practices in response to changes in the highly complex and dynamic global power equipment market. Many of these adjustments have been subsequently institutionalized as new policies, guidelines or rules. To retrive an earlier example, in 1988, the rule of thumb that guided ABB's major restructuring effort was that every staff group above the operating company level be cut to 10% ot its previous size. The prescribed guideline for the remaining 90% was that 30% be retired or laid off, 30% be assigned to independent, self-sufficient service centers, and 30% be redeployed to the decentralized operating companies. Four years later, reflecting the experience of several businesses able to make even deeper cuts, the rule of thumb had changed, and managers were being required to cut the staffs assigned to service centers from 30% to 15%, those in profit centers from 30% to 20%, and the remaining corporate staffs from 10% of their original level to 5%.
Yet, as insightful as many of these concepts are in describing some of the micro-level processes in ABB, they clearly do not capture the main thrust of its management decision making. In particular, Cyert and March describe an organizational learning process that is fragmented in approach ("with shifts tending to reflect the expansionist inclinations of subunits rather than systematic reviews by top management" [p.112]), short-term in focus ("so long as the environment of the firm is unstable and unpredictably unstable, the heart of the theory must be the process of short-run adaptive reactions" [p.100]), and incremental in nature ("because many of the rules change slowly, it is possible to construct models of organizational behavior that postulate only modest changes in decision rules" [p.101]). This fragmented, short-term focused and incremental learning process is assumed to be captured in and institutionalized by the company's standard operating procedures (SOPs) - "the memory of an organization" (p 101). In contrast, our observations suggest that the simplifying nature of SOPs, and the slowness with which they change often make them impediments rather than facilitators of effective renewal in the highly competitive, technologically sophisticated global environment in which companies like ABB operate. As a result, the process of incremental learning at the micro level tends to be framed by a much more macro process in which top management plays a key role.
Barnevik's ability to continually shake up the internal operations of ABB has, over the years, stimulated a much more dramatic and strategically focused process of organizational learning than the incremental process described by Cyert and March. For example, his decisions to merge the previously separate power generation, power transmission and power distribution businesses, to introduce a company-wide customer focus program, and to acquire companies that would leverage ABB's environmental protection technologies were all initiated to force such internal adjustment and learning. As he stated, "These efforts do not represent a one time project, but will push ABB in the direction of becoming a continuous learning organization". (President's comments, ABB 1991 Annual Report, P9).
The framing of this more macro process of organizational learning occurs within a goal-setting process that is also at variance from the Cyert and March model in which a company's objectives are defined in terms of bargained constraints, its aspirations determined by individuals' past experiences, and its actions primarily driven by a short-term problem solving process. In contrast to this view that firms "solve pressing problems rather than develop long range strategy" (1963:119), ABB managers at all levels place a great deal of importance on developing clear and shared strategic objectives. This process is rooted in a corporate vision that Barnevik presented at the senior management meeting in Cannes marking the formal merging of the two companies. While based on detailed analysis, this vision was hardly the outcome of internally bargained trade- offs. Instead, it reflected Barnevik's strongly held personal beliefs about the power industry and his assessment of the resources and capabilities within ABB. Some of these beliefs, such as his strong commitment to the then stagnant power equipment industry, were indeed contrary to the widely held views among ABB managers that the company needed to diversify away from this business. Yet, that vision became the basis for a massive acquisition activity, a major R&D investment program, and a variety of related strategic commitments.
To ensure the implementation of this overall blueprint, the company has created new structures and processes to allow managers at all levels to translate the broad objectives into specific business and market strategies. First, within the corporate executive committee, then at BA board meetings, and finally in the operating companies' steering committees, managers examine the implications of Barnevik's vision for their particular areas of responsibility. Overall, these processes of goal formation and organizational learning appear to be much closer to the "entrepreneurial" and "consensual" models that Cyert and March rejected. Furthermore, at the business level, the discussions to develop a consensus about strategies and priorities do not result in an outcome that Cyert and March feared would be "rather vague objectives" under which would lie "disagreement and uncertainty about subgoals" (1963:28). In fact, the annually updated business strategies become the basis for annual budget targets broken into very specific subgoals that are negotiated and agreed between business, geographic and company managements.
In sum, ABB has created a much more rational macro framework for goal setting and learning, capturing them both in a third core organizational process we describe as the renewal process (figure 4). Like the entrepreneurial and integration processes, it is defined by a set of top, middle and front-line management roles that not only drive the process but also embed it in the organization's ongoing activities. But, while the front-line managers are the key drivers of the entrepreneurial process and middle management provides the anchor for the integration process, it is the top management of the company that takes the lead in inspiring and energizing the renewal process.
- Figure 4 about here -This process of purposeful corporate renewal is very different from the individually based, negotiation driven process by which Cyert and March assumed that companies established their goals and aspiration levels. In fact, many of Barnevik's earliest actions were designed to break the coalitions and redefine the policies that had made ABB's precedent companies operate in such a politically negotiated manner. By reassigning managers, stripping out organization layers, redefining SOPs, and shifting the locus of power towards the operating companies, he reshaped coalition membership and redefined the relationships among them. Rather than waiting for the new organization to develop and institutionalize new objectives and behavioral norms through what Cyert and March described as "accidents of organizational genealogy" (1963:34), he intervened strongly to define a new common purpose and ambition to shape those objectives and standards.
In doing so, Barnevik seems to reject Cyert and March's assertion that "people have goals, collectives of people do not" (1963:26) as well as their complementary belief that the organizational aspiration level is determined by the attainable goals aspired to by individual coalition members according to their perceptions of the past. Instead, he has focused ABB on a very streched, future oriented sense of corporate mission that is defined with sufficient conviction and communicated with sufficient intensity that it can result in a simultaneous raising and convergence of individual aspiration levels within the company. It is an activity to which he has devoted an enormous amount of time, firstly in developing a clearly articulated sense of purpose, and then in overlaying it with a more dynamic sense of challenge.
While many companies have aspired to the voguish objective of "creating a shared vision", Barnevik seems to have moved well beyond the rhetoric towards making this an operating reality at ABB. He has done so through several means. The company's mission statement is designed to engage organization members worldwide and to align their interests with ABB's broadly defined corporate purpose "to contribute to environmentally sound sustainable growth and make improved living standards a reality for all nations around the world" (P6). With such a statement, Barnevik has created the sense of "public interest" and "social welfare" that Cyert and March saw as providing the common goal of political institutions (1963:28). However the mission statement then expands that unifying altruistic purpose into a more managerial objective: "to increase the value of our products based on continuous technological innovation and on the competence and motivation of our employees... becoming a global leader - the most competitive, competent, technologically advanced and quality minded electrical engineering company in our fields of activity" (P6). This translation of the broad mission to a strategic objective lends it a sense of organizational reality and legitimacy, and gives it more managerial power and relevance. Barnevik has further operationalized the broad vision by expressing the goals in financial performance terms - "10% operating profit and 25% return on capital employed by the mid 1990s". Such statements have not only provided a common aspiration level, they have also moved the mission statement beyond the "ambiguous goals" and "vague objectives" that led Cyert and March to conclude that such solutions to the goal setting problem were "misdirected" (1963:28).
Beyond creating a shared purpose, Barnevik has taken on a major role in providing the organization with a sense of ambition that, on the one hand, legitimizes the company's stretch targets and, on the other hand, creates sufficient strategic turmoil to stimulate organizational learning. For example, recognizing that ABB was too dependent on a stagnant European market, Barnevik set an objective of expanding its North American sales to represent 25% of the company's total. To initiate change, he acquired Combustion Engineering and Westinghouse's power transmission and distribution business in the United States, grafting these operations on to ABB's existing BAs. Not only did such a move boost the North American share of ABB's sales from 12% to 18%, it also exposed the Europe dominated business to North American technical developments (such as process control automation) and management practices (such as time based management) that were previously neither well understood nor effectively practiced in the ABB system.
The dual impact of a clearly defined anchor of corporate purpose juxtaposed against a slightly dissonant corporate ambition is felt particularly strongly in the front-line units. It is here that managers confront the gap between current performance and the aspiration level defined by the visionary purpose statement and the stretching ambition. If this process is to remain in balance, however, this tension created in the front-line units must be resolved. And without a high level of credibility and trust, such resolution can easily degenerate into horse trading or grand compromises. Creating and maintaining credibility and trust in the system, therefore, is a key requirement and these tasks define the main responsibilities of middle management in the renewal process.
One important means for executing these responsibilities lies in their stewardship of the various horizontal coordination mechanisms we have described. In contrast to Cyert and March's assumption that "firms do not resolve potential conflict... by a procedure of explicit mediation" (1963:27), managers like Ulf Gundemark create and manage a portfolio of boards, committees and project teams that serve not only as channels for the communication and transfer of knowledge, but are also explicitly designated as forums where managers can negotiate differences and resolve conflicts in an open and legitimate manner.Middle managers also ensure the legitimacy and credibility of this tension-filled process by creating a decision making context which is both participative and transparent. To ensure the credibility of top management's ambitious objectives, for example, Ulf Gundemark formed a Relays Vision 2000 Task Force of nine operating managers from various front-line companies, and charged them with the task of translating Barnevik's broad vision of ABB's place in the global power equipment industry into specific strategies for the relays business. After six months of intensive effort this task force developed a set of self-funding proposals that ranged from strategic investments to expand into new products (like metering) and new markets (like telecommunications), to plans for increasing employee motivation and organizational effectiveness. With a legitimacy and credibility that came from being defined by peers rather than being imposed by the top management, the strategy was accepted by front line managers like Jans who began trying to meet the ambitious challenges by expanding beyond their existing business boundaries into the new products and markets legitimized by Vision 2000.
The behavioral theory of the firm is, in essence, premised on the absence of leadership, while ABB's renewal process is clearly driven by highly effective leaders at the company's top level. In this specific attribute, the Cyert and March model is fundamentally different from Chandler's conceptualization of the classic M-form organization, which was premised on a strong leadership role of the top management. Alfred Sloan's towering presence in General Motors is perhaps the best known illustration of this role, as reflected in Chandler's (1962) description of him as the architect of GM's strategy, the builder of its structure, and the designer of its systems. Even since Sloan and his pioneering contemporaries discovered how the then emerging strategy of diversification was facilitated by the divisional structure and how that structure could be supported by some tightly designed planning and control systems, this strategy-structure-systems link has become an article of faith that has shaped top management roles in most large M-form organizations.
While it is clear that the role played by Barnevik and other top managers at ABB is not well captured by Cyert and March's model, it is important to understand how it also has diverged from the top management role described by Chandler. Given their starting point in the midst of a fusion between two large worldwide companies, Barnevik and his colleagues in ABB's group executive management initially focused on these tasks of formulating the merged company's strategy, structure and systems. Over time, however, their roles have evolved to be significantly different from these classic responsibilities of leadership in M-form organizations. From being formulators of corporate strategy, they have become the shapers of an instititutional purpose with which all employees can identify and to which they can commit. Instead of being the architects of formal structure, they have come to see themselves as the developers of organizational processes that can capture individual initiative and create supporting relationships. And, rather than being the designers of systems, they have refocused on the individual as the primary unit of analysis in the leadership task and have become the moulders of people.
Leadership in the M-form is fundamentally based on the view of companies as economic entities: managers in ABB and in most of the other companies we studied have premised their role on the recognition that large corporations are also complex social institutions. To capture the energy, commitment and creativity of their people, these managers are replacing the hard-edged strategy-structure-systems paradigm of the M-form with a softer, more organic model built around purpose, process and people. The renewal process we have described challenges the behavioral theory of the firm by the very existence of the leadership role; but it equally challenges Chandler's model of the M-form because of this basic change in the content of that role.
Chandler, Bower and Cyert and March provided three different and generally complementary views of the same phenomenon - the multidivisional organization that had first emerged in the 1920's, growing to become the corporate model of the 1950's and 1960's when all three studies were undertaken, and continuing to dominate the assumptions on which most large companies structured their operations into the 1990's. In examining the ABB organization through the perspectives of these three pioneering conceptual models, our objective was not only to develop a richer understanding of the subject of the analysis, but also to derive insights and ideas that would provide the basis for a new model of the large, complex firm more appropriate for companies like ABB.
This new organizational model we have proposed is a product of two inter-related but separate factors, and in this concluding section of the paper we will examine the implications of both. First, it reflects changes in the phenomenon being studied. Over the last decade or so, the operating environments of large global firms have changed significantly, requiring managers not only to rethink the structures and processes of their companies but also their own underlying management philosophies. The model we have presented here describes these emerging organizational characteristics and the new managerial assumptions on which they are built. In other words, the new model represents a real change in the organizations of large firms caused by the changes in their environments and task demands.
But our model also reflects a different research perspective. Despite the obvious fact that organizations are social structures that shape and are shaped by the relationships among actors within their social systems, organizational analysis has historically focussed on abstract generalizations of relationships represented by its formal structure. In contrast, in figure 4, we have defined our model in terms of three core processes that are built around a specific set of relationships among the front-line, middle and top managements of a company. In this way, we have presented a conceptualization of organizations, not as a scheme for dividing the overall corporate activities among a group of sub-units, but as a cluster of roles and their inter- relationships. From this perspective, it is the behaviors and actions associated with each of these roles that collectively define the social structure of a company within which its management processes are embedded. Thus, the new model is also the product of a different way of viewing and conceptualizing organizations.
In the late 1980's and early 1990's there were several new forces in the environment that were driving firms to develop a different management approach. In many industries, growth was occuring more slowly than earlier optimistic expectations, leaving most firms to contend with serious overcapacities and intensifying their need to reduce costs as a way to protect profits. At the same time, technological change was accelerating, reconfiguring industry structures and boundaries, and increasing the need for speed and flexibility at the firm level. Simultaneousely, information processing technologies were transforming internal organizational activities and management roles.
The combined impact of this slowing market growth, accelerating technological change and transforming organizational process was to shift the focus of many firms from allocating capital to managing knowledge and learning as the key strategic task. And, in this overall context of an increasing dependence on knowledge and expertise, the ability to attract and retain the best people was increasingly becoming a key source of competitive advantage. Collectively, these factors created the "urgent needs and compelling opportunities" that have forced companies like ABB to challenge the basic assumptions about corporate structure and management processes that lay at the core of their multidivisional organizations.
Concerned by the cost of supporting the heavily staffed, multi-tiered configuration into which such organizations had evolved (for example, Bower's research site showed eight layers of general management from the CEO to the product group manager; ABB's predecessor company, Brown Boveri, had nine), and frustrated by the slowness and inflexibility it caused in decision making, the most obvious target companies focused on was their basic organizational structure. Managers like Percy Barnevik began to challenge the conventional wisdom that the classic multidivisional hierarchy was the structural model of choice for large, diversified companies. In doing so, they began to question the basic philosophy underlying that venerable organization form.
Because the M-form typically succeeded the functionally organized U-form company, it built on a management model in which responsibility and authority were concentrated at the apex of the organization in the hands of its only general manager, the CEO. In creating the multidivisional structure, companies were literally dividing the single corporate entity into several entities appropriately called divisions. The new structural units were given legitimacy and power through the allocation of resources and delegation of responsibility to general managers created at this level. In short, the M-form was created through a philosophy of devolution of assets and accountability from the corporate to the division level.
The new organizational model we have described, however, represents a radically different structural philosophy. Conceptually, it is based on an assumption that the organization needs to be developed and managed on a principal of proliferation and subsequent aggregation of small independent entrepreneurial units from the bottom up, rather than one of division and devolution of resources and responsibilities from the top. This structural philosophy implies a significantly different distribution of corporate assets and resources. In contrast to the classic M- form where control over most resources is held at the corporate level, in the new model, resources are decentralized to the front-line units which operate with limited dependence on the corporate parent for technological, financial or human resources, but with considerable interdependence among themselves. In turn, this approach allows a drastic reduction in supervisory layers (like Barnevik at ABB, Jack Welch's objective has been to reduce the number of organizational layers at GE from 10 to 4) and the size of staff groups each level of management needs.
Beyond driving these changes in the macro-structure, the environmental shifts have also led managers in the new organizations to focus on a very different kind of organizational process compared to those that dominated the operations of traditional divisionalized companies. Chandler's structural perspective concentrated his attention on a set of processes he identified as strategic and tactical decision making. Bower's model of the firm as a political organism led him to focus on hierarchical information processing activities that influenced resource allocation decisions. And Cyert and March's view of the firm as a coalition of participants negotiating to achieve their disparate objectives drew their attention to the micro-level processes of setting and adjusting objectives. While shaped by the perspectives of their research, this focus on planning, resource allocation, and objective-setting also reflected the emphasis managers of multidivisional companies placed on these processes. The common characteristics of all these processes was that they were administrative in their focus and function, and were designed primarily to reflect the vertical information processing and decision making capabilities of the M-form structure. Yet, eventually it was the complexity and inefficiency of these internally-focused planning, allocating and controlling activities that forced many companies to make radical changes to their organizations.
In the emerging organizational model, the elaborate planning, coordination and control systems have been drastically redesigned and simplified as management time and attention has shifted towards the creation and management of processes more directly related to adding value than on facilitating internal administrative activities. It was this shift in focus that led to the widespread implementation of process-based activities such as total quality implementation, new product development, or customer focussed management for example. At the most strategic level, this has resulted in the development of the three core processes that are at the heart of our model. In many ways, these three processes provide a necessary counterbalance to the biases and limiting assumptions that were increasingly causing difficulties in the classic M-form organization. The entrepreneurial process highlights the need to supplement top management's direction and control with front-line initiative and flexibility; the integration process creates a horizontal information processing capacity at least as efficient as the previously dominant vertical mechanisms; and the renewal process imposes a dynamic tension into the organization that preventes it from developing strategic commitment at the cost of organizational adaptability. It is a process management orientation quite different from the classic one of ensuring that the allocation of resources matched the strategic priorities and that operating performance met the budgeted objectives
.Just as the largest companies of the time like General Motors, du Pont, Sears and Jersey Oil. were the pioneers of the new M-form organization in the 1920's and 1930's because they experienced the emerging environmental shifts most acutely, so too are the largest global companies of today like General Electric, ABB , 3M, Toyota and Canon acting as the pioneers of the new organizational form we have described. They are often the first to experience the strains because their existing organizations are stretched so thin by the "urgent needs and compelling opportunities" arising from the emerging internal and external environmental demands.
Documenting a change that had occurred a few decades earlier, Chandler could both describe the multidivisional organization quite precisely and provide some clear evidence of its advantages over the functional form. Not having the benefit of such a retrospective perspective, our description of the new organizational model is considerably less rigorous. However, as we will show in a forthcoming monograph, we observed several companies in our sample independently developing similar organizational characteristics, leading us to conclude that the changes being made at ABB are not an ideosyncratic experiment. Our hypotheses is that these companies and others are laying the basic foundations of an emerging organizational form for large complex corporations.
Over the last two decades, organizational theory has come to be increasingly dominated by a remote, reified and pessimistic view of large organizations. The remoteness is manifest in theories such as population ecology and transaction cost analysis in which the focal level of analysis either lies in gross aggregations such as organizational populations or in the micro-level detail of specific transactions. With such distant foci, in these theories, organizations have been reduced to reified abstractions, stripped of all the concreteness of social actions and relations that define them. At the same time, these theories have presented a view of individual-organization interactions that is grounded in the assumption that the human role in organizations is essentially passive and pathological. This negative assumption about human agency is manifest in the extreme determinism in population ecology (Hannan and Freeman, 1977), the powerful forces of isomorphism that have virtually replaced the role of leadership in institutional theory (contrast, for example, DiMaggio and Powell, 1983 and Selznick, 1957), the denial of purpose and direction in behavioral theory of the firm (Cyert and March, 1963) and the assumptions about shirking, opportunism and inertia in organizational economics (Williamson, 1975; Alchian and Demsetz, 1972). The organizational model we have presented here represents a different perspective on both organizations and on individual-organization interactions.
An organization is fundamentally a social structure. Even though actions of and within organizations may be motivated by a variety of economic and other objectives, they emerge through processes of social interactions that are shaped by the social structure. Over 35 years ago, Merton provided a framework for the analysis of such interactions based on the definitions of statuses and roles within the social structure. As argued by Merton, "status" implies a position in a social system involving designated rights and obligations and "role" defines the behavior oriented to the patterned expectations of others. "In these terms, status and roles become concepts serving to connect culturally defined expectations with the patterned conduct and relationships which make up a social structure" (Merton, 1957:110). Weick (1969) has similarly defined the structure of an organization in terms of "interlocking behaviors" which are reciprocal actions that are repeated over time. As summarized by Nelson, "if behaviors between two or more parties are repeated over time to the point where those behaviors are taken for granted as normal or expected, the resulting relationship or tie has become part of the organization structure" (1986:75). Despite this clearly relational nature of the concept, however, an increasing preoccupation with structural forms has resulted in most organizational analysis focusing not on the network of roles and relationships that define a social structure but on constructs such as centralization, formalization or divisionalization which, at best, represent some broad generalizations about those relationships.
In contrast, in figure 4 we have defined the new organizational model in terms of three core positions within a company's management structure - the front-line, middle and top-level managers - and the behaviors and actions associated with each of them. Each of the three core pro cesses is structured around a specific set of relationships across these three roles; the three processes coexist because of the overall symbiosis within and across those roles. In this way, we have defined the structure of the organization, not in terms of how sub-units are composed and decomposed but as a cluster of statuses and associated roles that collectively define the social structure of a company within which its core management processes are embedded.
In explicating its differences from the M-form's structure, processes and decision-making mechanisms, we could have built our arguments not by focusing on the three processes in the new organizational model but by highlighting the very different roles that front-line, middle and top- level managers play in companies like ABB (see table 1). In Chandler's model, the top management is the entrepreneur and resource allocator, middle managers are the administrative controllers, and front-line managers are the operational implementers. For Bower, the top-level manager is the creator of a company's structural context, the middle manager is the vertical information broker, while the front-line manager is the initiator of operating decisions. In Cyert and March's analysis, top-level management is the dominant coalition that establishes the standard operating procedures and resolves conflict, middle managers are the champions and advocates of sub-unit preferences based on local rationality, while front-line managers are the organization's problem solvers. In the organizational model we have described, while all the three management groups have key roles to play in each of the three core processes, top-level managers are primarily the creators of organizational purpose and the challengers of the status-quo, middle level managers are the horizontal integrators of strategy and capabilities, and the front-line managers are the organizational entrepreneurs.
- Table 1 about here -The second difference in our research perspective relates to the assumptions that are made about the nature of human motivations and interpersonal behavior within organizations. Particularly in comparison to economic theories that assume a human propensity towards shirking, opportunism and inertia (Williamson, 1975; Alchian and Demsetz, 1972), the managers in the firms we studied took a much less pathological view of human nature. The new model developed from management perspectives reflects that orientation.
The more positive view of human nature on which this model is built, however, does not rely on a fundamental premise of altruism as a stable human disposition. Instead, it reflects more of a relativist view of personal attributes, and more of a situationalist perspective on human behavior (see Kenrick and Funder, 1988 for a recent review of this perspective). Its assumption, based on our interpretation of the views most managers we interviewed expressed, is that human beings are capable of both initiative and shirking, that they are given to both collaboration and opportunism, and that they are constrained by inertia but are also capable of learning. Within a firm, actual behavior is determined in part by the prior disposition of actors (i.e., the location on the spectrum of personal characteristics of the particular individuals involved) and in part by the situation they face (i.e., how the firm's context influences their behavior) (see Figure 5)
- Figure 5 about here -This model assumes that companies ensure positive individual characteristics both by selecting and promoting those whose personal characteristics predisposes them toward the desired norms of behavior, and by creating an internal context that encourages people to act in the way they would as a member of a functional family or a disciplined sporting team. Thus, the entrepreneurial process is built on the assumption that individuals have the capacity of personal agency and initiative, then creates the selection devices and support mechanisms to elicit and encourage such behavior. Similarly, the integration process both assumes and shapes collaborative behavior, and the renewal process is designed to capitalize on the human motivation to learn while creating a context that drives them to do so. In short, the same managerial actions that drive the three processes also help create an organizational context that reinforces the effectiveness of the processes by inducing organizational members to take initiative, cooperate and learn. This interactive and mutually reinforcing development of management action, organizational context and individual behavior is of central importance to the model we have presented here and we will elaborate and illuminate these interactions more thoroughly in a forthcoming article (Ghoshal and Bartlett, in press).
We call ours a managerial model for two reasons. First, it describes the firm in terms of the managerial roles necessary to develop and manage the three core processes. Equally important, however, is the fact that the model was developed by analyzing the operations of a firm from the perspective and in the language of the managers who live within the system. In this sense, our model is very different from currently dominant theories of the firm which are grounded in the perspectives and languages of different social sciences. The managerial perspective in our model, we believe, is at least as useful for illuminating organizational phenomena as those generated by economists describing the firm in the context of market failure or social psychologists describing it as a higher order aggregation of individual and group behavior that are typically their focal level of analysis. It is this managerial perspective that we would like to see legitimized in a new theory of the firm that would focus on the distinctive characteristics of large business organizations and illuminate issues that managers of such firms perceive to be important. As we acknowledged in the introductory section, clearly a lot more work has to be done before this initial effort of describing a new organizational form can progress to the stage of presenting an alternative theory of the firm. Our hope is that the model we have proposed in this article might provide some energy to that agenda.
Alchian, A.A. and H. Demsetz (1972), "Production, Information Costs, and Economic Organization", American Economic Review, 62(5), pp 777-795.
Allison, G.T. (1971) Essence of Decision. Little, Brown, Boston, MA.
Barnard, C.I. (1938) The Functions of the Executive.. Harvard University Press, Cambridge, MA.
Bower, J.L. (1970 Managing The Resource Allocation Process. Division of Research, Graduate School of Business Administration, Harvard University, Boston, MA.
Chandler, A.D. j.r. (1962) Strategy and Structure. The M.I.T. Press, Cambridge, MA.
Chandler, A.D. j.r. (1991) "The Functions of the H Q Unit in the Multibusiness Firm", Strategic Management Journal, Vol 12, pp 31-50.
Cyert, R.M. and J.G. March (1963) A Behavioral Theory of the Firm. Prenctice- Hall, Englewood Cliffs, NJ
.DiMaggio, P.J. and W.W. Powell (1983) The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields, American Sociological Review, 48 ,pp 147-160
Ghoshal S. and C.A. Bartlett, "Linking Organizational Context and Managerial Action: The Dimensions of Quality of Management" forthcoming in the special issue of the Strategic Management Journal edited by C.K. Prahalad and Gary Hamel, Summer, 1994.
Hamel G. and C.K. Prahalad (1993) "Strategy as Stretch and Leverage", Harvard Business Review, March-April, pp 75-84.
Handy, C. (1990) The Age of Unreason. Harvard Business School Press, Boston, MA.
Hannan, M.T. and J. Freeman (1977) The Population Ecology of Organizations, American Journal of Sociology, 82, pp 929-964.
Hill, C.W. (1988) "Internal Capital Market Controls and Financial Performance in Multidivisional Firms", The Journal of Industrial Economics vol 37, no 1, pp 67-83.
Kanter, R.M. (1983) The Change Masters, Simon&Schuster, New York. P>
Kenrick, D.T. and D.C. Funder (1988) "Profiting From Controversy: Lessons From the Person-Situation Debate", American Psychologist, vol 43, no 1, pp 23-34.
Merton, R.K. (1957) "The Role-Set: Problems in Sociological Theory", British Journal of Sociology Vol 8; pp 106-120.
Nelson, R.E. (1986) "The Use of Block modelling in the Study of Organization Structure: A Methodological Proposal", Organization Studies, Vol 7, No 1, pp 75-85.
Peters, T. (1992) Liberation Management, Alfred A. Knopf, New York.
Selznick, P. (1957) Leadership in Administration, Harper and Row, New York.
Taylor, W. (1991) The Logic of Global Business: An Interview with ABB's Percy Barnevik, Harvard Business Review, March-April, pp 91-105.
Weber, M. (1964) Basic Concepts in Sociology. The Citadel Press, New York.Weick, K.E. (1969) The Social Psychology of Organizing, Addison-Wesley, Reading, Mass
Williamson, O.E. (1975) Markets and Hierarchies: Analysis and Antitrust Implications. The Free Press, New York.