Strategy, capabilities and corporate coherence: Exploring some dynamics of learning

This paper explores the notion of corporate coherence by relating it to models of experiential learning. It is suggested that corporate coherence is a determinant of strategic direction and competitive advantages of firms. Moreover, the composition of learning activities implies the range and extent of corporate coherence, due to various problems associated with myopia of learning (Levinthal and March 1993). The paper specifies a model for analyzing corporate coherence in a learning framework. The paper suggests that corporate coherence, while being a necessary component in understanding strategic management, is a multi-faceted concept. It implies multi-level learning, where learning substitution (due to simplification and specialization) takes place both within each learning activity, as well as between activities. Introduction At the core of research in strategic management lie the research questions of what determines the direction of corporate strategies and how chosen strategies can lead to sustainable competitive advantages. As an example of the importance of the latter issue, the previous decade witnessed a, sometimes fierce, debate between resourcebased and market structure-based answers to questions related to sustainable competitive advantages (cf. Grant 1991; Barney 1991; Porter 1996: Foss 1999). In regards to the former issue, the strategic management research of the 1980s had one of its focuses on explaining corporate diversification (cf. Penrose 1959; Rumelt 1972). One avenue suggested for probing forward towards an explanation of the rate and direction of strategies and how corporate strategies relate to performance is the further investigation of organizational capabilities (cf. Dosi, Teece and Winter 2000; Helfat 2000). This links issues of strategic management to economic theories of the firm. This paper continues in this direction by investigating the learning heuristics of organizational capabilities. It suggests that strategies based on the organizational capabilities of a firm should be understood as activities of learning. Moreover, we propose answers to the questions of direction and advantages of competitive strategies by using the notion of corporate coherence (Dosi, Teece and Winter 1992; Teece et al 1994). The first aim of the paper is to show how the dynamics of learning shape strategic behavior in large industrial corporations. Second, using the theory of learning simplification and learning specialization of Levinthal and March (1993), the paper tries to establish the relationship between the organizational learning dynamics and the corporate coherence of industrial enterprises. Third, the paper tries to address some issues derived from this framework that are related to research in business history. The paper is structured accordingly. The first section relates research in strategic management to organizational capabilities, more specifically discussing the concepts of organizational learning and corporate coherence. The second section outlines a model of corporate strategy based on the idea of corporate coherence. It is suggested that corporate coherence is determined by various dynamics related to experiential learning. A third section describes a methodology for investigating this idea empirically, while the fourth section of the paper provides some empirical examples of the issues discussed based on a longitudinal comparative study of the heavy electrical engineering industry and suggest further research issues. Strategy, capabilities, learning and corporate coherence Two major themes in research into strategy have been the creation of sustainable competitive advantages firms (cf. Porter 1980) and the direction of firms’ strategic decision-making (cf. Eisenhardt 1989; Rumelt et al 1994). One emerging theme in the strategic management literature in response to these major themes has been the capabilities’ framework (cf. Helfat 2000). As pointed out by Dosi et al (2000) such a framework, relying on the notion of organizational capabilities (Chandler 1977), interconnects scholars in evolutionary economics, strategic management, technology and organization, business history, and economic growth. In the area of corporate strategy, however, the development of the concept of capabilities has most frequently been associated with a resource-based theory of strategy (Eisenhardt and Martin 2000). Grant (1991) pointed to some of the implications of a resource-based theory of strategy. He there repeated the message of e.g., Barney (1986), Lippman and Rumelt (1982) and Wernerfelt (1984) that a firm’s capabilities may give rise to sustainable competitive advantages due to the difficulty for competitors to imitate such capabilities. Moreover, and of special interest for this paper, Grant showed how the capabilities of a firm provide a base for decisions on future strategic decisions and actions. By thinking of strategy in terms of a resource-based theory of the boundaries of the firm, it is hence possible to derive an answer to why e.g., related diversification is more common and more successful than unrelated diversification (cf. Penrose 1959; Chandler 1962; Rumelt 198?). This answer seems to lie in the nature of organizational capabilities: these are essentially a product of organizational learning (Nelson 1991; Chandler 1992; Lazonick and Mass 1995; Dosi, Nelson and Winter 2000). As stated by Eisenhardt and Martin (2000, p. 1106): “...well known learning mechanisms guide the evolution of dynamic capabilities and underlie path-dependence.” However, this answer begs the question of how the dynamics of organizational learning is related to organizational capabilities and strategic management. It is to this question that this paper seeks to explore some potential answers. An important building block in providing such answers is the notion of corporate coherence, introduced by Dosi, Teece and Winter (1992). Corporate coherence has primarily been studied on the level of products or technology (e.g., Teece 1982, Granstrand and Sjölander 1990; Patel and Pavitt 1997; Piscitello 2000). In this paper we will argue that corporate coherence, as a core concept for strategic management, should be concerned with the coherence of the corporation as a whole, thus also involving e.g., organizational and market dimensions. An extension of the corporate coherence framework This paper revisits the learning model provided by Levitt and March (1988); Levinthal and March (1993) and discusses it in relation to the notion of corporate coherence (Dosi, Teece and Winter 1992; Teece et al 1994). Dosi et al (1992, p. 188) defined corporate coherence accordingly: “A firm exhibits coherence when its line of business are related, in the sense that there are certain characteristics common to each. A firm’s coherence increases as the number of common technological an market characteristics found in each product line increases.” Moreover, these authors (ibid. p. 190) insisted that a theory of the boundaries of firms should consider learning, path-dependencies, technological opportunities, selection and complementary assets, and not merely transaction-costs. In this paper we will argue that the first of these considerations, namely learning is particularly important. Moreover, it will be proposed that corporate coherence is not just an issue at the product level, but, more generally, coherence is a feature of the learning activities of the firm as a whole. We will revisit a search model of learning under bounded rationality and evolving preferences, with positive reinforcement mechanisms, and populations of learning agents (cf. Levinthal 1995; 2000). The general model has been, in various modes and forms, elaborated on by authors such as Cyert and March (1963), Simon (1976), Nelson and Winter (1982), Dosi et al (1995), Levinthal (1995), and Arthur (1994). An initial assumption in this model is that experienced-based learning is far from easy, as stated by Levinthal and March (1993, p. 96): “Experience is often a poor teacher, being typically quite meager relative to the complex and changing nature of the world in which learning is taking place. Many of the cognitive limits that constrain rationality also constrain learning.” Levitt and March (1988) survey literature on organizational learning using three basic assumptions derived from a behavioral perspective on the firm. First, they argue that organizational behavior is based on routines (Nelson and Winter 1982), significantly procedural or based on a logic of appropriateness (March and Olsen 1989). Second, Levitt and March suggest that routines are interpretations of the past. Third, and based in bounded rationality, they propose that organizations are target-oriented or only intentionally rational. In essence, this refers to the dependence of organizational behavior upon the relation between observed outcomes and the level of organizational aspirations. They distinguish between two types of learning from experience: i) trialand-error experimentation, as the probability of routine usage increases if it is associated with success in meeting a target; and ii) search, organizations draw upon a pool of alternative routines, if they discover better ones, these are adopted. Search and learning in this model tend to be local, for a number of reasons. Interpretation of experience is difficult as lessons are to be drawn from a relatively small number of observations in a complex and changing environment of learning organizations (Levitt and March 1988, p. 323). This makes it difficult to localize causality and draw correct inferences. Organizations and organizational members exhibit systematic biases in interpretation, since they attend overwhelmingly to recent and salient events, are insensitive to sample size, attribute overly importance to intentionality; and use simple and linear algorithms. Collective understanding of history creates unreflected frames or paradigms, in which learning occur. Moreover, success is ambiguous, since there are problems in evaluating outcomes as well as in the unambiguous interpretation of goals. These dynamics lead to “superstitious learning” i.e., when subjective experience is compelling but relations between actions and outcomes are disconnected with proper causation. A related dynamics of learning is what Levitt and March (1988) and Levinthal and March (1993) call “competency traps”: “The effectiveness of learning in the short-run and in the near neighborhood of current experience interferes with learning in the long run and at a distance. Knowledge and the development of capabilities improve immediate performance, but they often simultaneously reduce incentives for and competence with new technologies or paradigms. Learning has its own traps.” (Levinthal and March 1993, p. 97) The occurrence of competency traps is due to multilevel learning. If each routine is regarded as a collection of routines (i.e. a capability) learning takes place at several nested levels, and needs to both discriminate among routines as well as learn within them. This problem leads to specialization of routines, since those routines that are associated with success become more frequently used. Competency traps then occurs as organizations lack the knowledge of causal relationships, and obtain favorable performance with an inferior procedure. This causes organizations to focus on that routine and not giving sufficient attention to other, sometimes superior, procedures. “Competency traps are particularly likely to lead to maladaptive specialization if newer routines are better than old ones.” (Levitt and March 1988, p. 323) Organizational learning becomes particularly cumbersome in ecologies of learning organizations. The simultaneous adaptation implies that routines yield different outcomes at different times. Furthermore, due to endogenously changing environments, systematic comprehension of learning is complicated to model (Levitt and March 1988, p. 331). Levinthal and March (1993) argue that organizations use two generic mechanisms to facilitate learning from experience. The first strategy is simplification or buffering. This mechanism simplifies experience, minimizes interactions, and restricts effects of the spatial and temporal neighborhood. Simplification is important since organizational learning requires the interpretation of experiences; a simpler learning environment facilitates interpretation, and isolated units can learn effectively while simultaneous learning is difficult. Organizations thus create buffers, either through departmentalization, which segregates experience in units or through sequential attention to goals. In this way buffering allows for local search, and problem arises in tightly coupled organizations which can be expected to be better in error detection than error diagnostics (cf. Lindkvist, Söderlund and Tell 1998, p. 944). The second mechanism proposed by Levinthal and March (1993, p. 99) is specialization and the principle of learning substitution between different learning mechanisms. Deriving from the adaptive model of boundedly rational organizations, one can infer that success of adaptation at one part of an organization will cause: i) a relief of pressures for adaptation of another part; and ii) a relative increase in adaptive capacity of this part of the system compared to the part not being used. Hence, specialization of learning competence (cf. Levitt and March 1988, p. 322). Levinthal and March (1993, p. 100) suggest that organizations can respond to success or failure by: • Varying the intensity of search. • Varying the level of organizational slack. • Varying the target (aspiration level) of performance. Success implies a decrease in search and an increase in slack and targets, while failure causes increase in search and decrease in slack and targets. These three mechanisms serve as substitutes for each other. The substitution is complicated because of the interwoven structure of learning levels. “Learning experience is nested. That is, learning occurs at several different but interrelated levels at the same time. An organization simultaneously learns which strategy to follow and how to operate within various alternative strategies [...].” (Levinthal and March 1993, p. 100, italics in original) In accordance with the reasoning of Nelson and Winter (1982), there is a nesting of “knowing how to choose” and “knowing how to do”. However, learning at one level may effectively substitute for learning at another level, which causes learning rates to differ. “Fast adaptation at one level in an organization leads to slow learning at other levels” (Levinthal and March 1993, p. 101). This implies that either do organizations refine existing knowledge (exploitation, March 1991) or search for better knowledge elsewhere (exploration, March 1991). Interestingly, organizations will tend to substitute learning at the operating level for learning at higher levels. “Lower-level adaptation is a sensible activity that tends to enhance an organization’s position in its present environment. In the long run, however, such first-order learning can not substitute for second-order learning of new routines and strategies.” (Levinthal and March 1993, p. 101) This analysis of organizational learning points to the myopia of organizational learning. Levinthal and March (1993, pp. 101-105) suggest that the myopia of organizational learning has three effects, all related to the “localness” of learning. First, it causes organizations to overlook distant events. Second, it leads organizations to overlook distant places. Third, it implies that organizations overlook failures. A proposed methodology: Learning activities in the industrial enterprise In order to further investigate the implications of learning effects on organizational capabilities and corporate coherence, we here suggest three learning activities, management-based activities, technology-based-activities, and market-based activities, as particularly pertinent in the formation of organizational capabilities in science-based industries. In providing this “three-pronged” partitioning of activities, we adhere to some previous work made in the study of industrial and corporate change (e.g. Teece and Pisano 1994; von Tunzelmann 1995; Pavitt 1998; Dosi et al 2000). However, the study most akin in both conduct and content, is perhaps business historian Bernard Carlson (1991), who uses the partitioning into technology, organization and markets when describing and analyzing the early history of the American electrical manufacturer Thomson-Houston. Management-based activities Like Lazonick and Mass (1995), it is possible to argue that one important activity in an organizational capability is the way management through different means differentiates and integrates productive knowledge. In particular, the ability of firms to integrate knowledge becomes important. Investment in management-based activities does not only refer to the elaboration of a managerial hierarchy, but also to the hiring of professional managers trained in planning, accounting, and organizing. Management knowledge and judgment is tightly associated with the operations of a firm and not easily transferable e.g., to other industries. Following the arguments pursued in Penrose (1959), Nelson and Winter (1982), and contemporary theorists associated with the knowledge-based theory of the firm, it seems adequate to recognize that investments in management is basically investments in knowledge accumulation in managerial routines. Such routines involves strategizing and organizing, which can be manifested in decentralization or centralization of decision making authority, associated provision of incentive schemes and the elaboration of guiding visions. Therefore we suggest that the development of management-based activities as one important sub-component of organizational capabilities in the large industrial firm. Technology-based activities In Chandler’s (1977; 1990) historical analysis investments in manufacturing play an important role. It is in manufacturing that we find the major advantages of scale and scope. Despite its ingenuity, this analytical category somewhat narrowly depicts production as primarily manufacturing, and says little on how firms come about to make improvements in manufacturing and generate new products. Conceiving organizational capabilities as an evolving system, the Chandlerian view implies primarily learning by doing effects (Arrow 1962) by knowledge exploitation and technological knowledge accumulation (Dosi 1988; Lazonick and Mass 1995). However, it leaves out much of the innovation induced by technology i.e., the ability of the modern enterprise to endogenously “shift its production function” by investing in R&D; or by acquiring new technology. A conspicuous detail of the modern large enterprise, particularly in science based industries, is its ability to generate new technologies that improves processes and products (Dosi 1988, p. 1158). An account of organizational capabilities in manufacturing should, particularly in industries that are technology or science-based, include the development of capabilities in the realm of both constructing and implementing technology. In the literature on technological and economic change (e.g., Dosi 1982; 1988; Malerba and Orsenigo 1993; Nelson and Winter 1982; Pavitt 1984) the tendency of technology to form history-dependent trajectories has been discussed. This theorizing interconnects with theories of the firm in that they view the firm and its members as important in shaping such trajectories e.g., through the establishment of technological paradigms (Dosi 1982). There thus seems to be activities in industrial firms that not only contributes to technological development and e.g., increasing complexity of technologies and products, but also activities aiming at integrating and make use of such complex technologies. As science based technologies may form quite complex systems, there are reason to believe that such activities can be quite important for the viability of the firms competitiveness. Market-based activities Chandler (1990) rounds off his analysis of the emergence of managerial capitalism by emphasizing the part played by marketing and distribution investments. The firm is however not only responding to changing market conditions (Dosi 1988, p. 1142; Tell 1997, p. 30). Rather, the knowledge-accumulating firm is quite inert, moreover it tends to construct much of its knowledge from the “inside”. However, as pointed out by Dosi (1988), innovation studies show that market conditions exert influence on the search performed by firms, providing “focusing devices” (Rosenberg 1976, pp. 111112). Moreover, as shown in von Tunzelmann (1995), it is the transformation of technological and market knowledge by firms that has shaped industrial progress during the last 150 years. Hence, in building organizational capabilities, market needs and preferences are important. As argued by Penrose (1959, pp. 5; 42), firms’ environments (e.g., technology or market) are primarily in the eyes of entrepreneurs and managers that use their knowledge in recognizing productive opportunities, and not strictly determining external forces. Loasby (1996; 1998) argues that an important activity in the formation of organizational capabilities thus becomes investing in markets, since they too are capabilities. Such investments may take the form of financing potential consumers of the firm’s products. The large industrial enterprise is also able to affect its market through actions pursued, this goes for inter-firm relations established (both horizontally and vertically) as well as the creation of demand by distribution networks and commercial advertising. Another feature indicating the marketing dimension of organizational capabilities is knowledge-based is the observed tendency of industrial firms to “lock-in” on customer preferences, which may lead to missed technological opportunities (Christensen 1997). It is by market-based activities that firms “link up” with market changes and opportunities, and that such activities make up an important part of the concept. Strategy and learning dynamics the history of electrical manufacturing The following discussion is based on a historical study of the heavy electrical manufacturing industry. Some of the largest electrical manufacturers in the Western world during the period 1878-1990 were studied. The basis for such a study was a 1 For instance Hertner (1993, p. 168) comments on the early development in the German electrical manufacturing industry: “[...] the ownership advantages of the German firms such as Siemens & broad survey of existing secondary literature on the electrical manufacturers in the U.S., Germany, the U.K., France, Switzerland and Sweden (Tell 2000). The study was primarily focused on their activities in the power transmission segment. The partitioning into the three learning activities suggested above was used for analyzing some implications for corporate coherence. The story in brief According to Glete (1983, p. 13), there are (or have been) three main categories of electrical manufacturers, based on their ambition in the industry. The first group of firms consists of those firms that are large and comprehensive with a full range of products sold on international markets. Secondly, there are medium-sized companies striving for the manufacturing of the basic products of the industry for their domestic market, not aiming for the technological frontier. The third category of electrical manufacturers is specialist firms with a narrow range of products, in which they are able to sustain a world-class technological level. The focus of this study is almost exclusively the first category, or the growth of firms belonging to group two or three into the “top-tier league”. The electrical power transmission industry grew out of the scientific findings in electromagnetics in the mid-1800s. With the invention of incandescent lighting, Thomas Alva Edison in the U.S. designed a system for generation and distribution of electricity for lighting in cities, and the first system – the Pearl Street Station – was inaugurated in New York in 1882 (Hughes 1983). In order to develop and manufacture his inventions in electricity, Edison had established the Edison Electric Company, which changed name to Edison General Electric in 1889 as it merged with some other companies. Also in the U.S., George Westinghouse founded the Westinghouse Electric Company (Westinghouse). Edison General Electric merged with Thomson-Houston Electrical Company, the third major electrical manufacturer in the U.S. in 1892 and General Electric (GE) was formed. With the establishment of a universal system for power transmission, there were also two major competitors left on the American continent: GE and Westinghouse. Halske and AEG before the First World War lay not only in their superior management and technology, but their marketing and after sales services” In Europe, Germany became the center for electrical manufacturing. Siemens & Halske was established already in 1847, exploiting Werner Siemens’ inventions for developing the pointer telegraph. The company also invented and refined the dynamo, enabling it to compete in electrical power transmission. The major transfer of technology from the U.S. to Germany, however, occurred under the supervision of Emil Rathenau, and his company subsequently known as the Allgemenie ElektricitätsGesellschaft (AEG). Through different licenses AEG was able to use the technology developed by Edison for the German and European market. These two companies and their American counterparts rapidly grew into four industrial giants on international basis (Chandler 1977; 1990; Freeman and Soete 1997). Except for these four companies, there were many “national champions” in electrical engineering throughout Europe. In Switzerland Oerlikon and Brown Boveri (BBC) were early technological leaders, In Sweden, Allmänna Svenska Elektricitetsaktiebolaget (ASEA) was the prime company. Albeit a bit later, General Electric Corporation (GEC), English Electric, and Allied Electric Industries (AEI) grew stronger in Great Britain. In France, the major manufacturers were Compagnie Générale d’Electricité (CGE) and Alsthom, but here, as in Britain, the influence from foreign manufacturers was significant. However, these companies, as well as the many others competing with equipment for electrical power transmission, were dramatically smaller than the four giants were. Noteworthy, however, is that all these smaller companies still had a capability of manufacturing and designing equipment for complete power transmission systems. They were, however, only “small giants” compared to the four “large giants”. ...Siemens and AEG, along with General Electric and Westinghouse, continued to dominate one of the world’s most significant industries during the whole period from the 1880s to the 1940s. (Chandler 1990, p. 464) This situation further accentuated and strengthened until well after World War II. The American and German firms also diversified into a number of other areas, not at least due to the capabilities that they had developed in R&D; laboratories (Wise 1985; Chandler 1990). An important part of the strategy was to produce equipment for consumers that needed electricity, hence causing demand for more electricity and power transmission. Moreover, especially the American firms began managing their diversified companies in a more decentralized way. The divisionalized form and profit-center based organization were pillars upon which Strategic Business Units (SBU) were built. After World War II a ban was placed on scientific research for German companies that was not lifted until 1955. Gradually thereafter, Siemens restructured and centralized its organization. In 1969, the creation of the Kraftwerk Union (KWU) and the Transformatoren Union (TU) meant that the two German companies merged their power engineering capacities in a joint venture. AEG ran into severe financial difficulties however, and Siemens acquired most of its parts in the joint venture in 1977, with the result of only one German giant remaining. Until about 1960, the manufacturers, through cartels and quotas, controlled the global market for electrical power transmission equipment. The Americans exploited their vast home market, while the Europeans also drew on domestic markets but also had extensive export to developing countries. The American firms were discouraged to continue these agreements after an indictment of violating the Sherman Act in the beginning of the 1960s (Sultan 1974). However, the patterns of trade for this period clearly show that especially the American firms focused on their domestic market (Surrey and Chesshire 1972). It was not only the German firms that were in turmoil at the late 1960s. In Britain the major three competitors GEC, AEI, and English Electric merged within a few years, creating one strong electrical manufacturer: GEC. The new organization, under the head of Arnold Weinstock, was pursuing the profit center style of management even more forcefully than the American firms were. The trend of concentration furthered, as the Swiss firm BBC acquired its minor compatriot Oerlikon in 1967 (Catrina 1991), and as CGE became a major shareholder of Alsthom in France in 1969. In the 1980s, the time for change had come to the U.S. GE moved away from what used to be one of the pillars of its business. Power transmission was no longer considered a core business of the company and in 1987 the plant for large transformers (the very heart of a high voltage power transmission system) was closed down (Blalock 1998). The company sought more fertile and profitable soil in emerging high-technology industries and service business. Westinghouse performed poorly throughout the 1970s and 1980s, the company responded by pursuing portfolio-style management acquiring and divesting businesses at a rapid pace, and little was invested in R&D; for heavy electrical engineering. In 1989, Westinghouse sold the Power Distribution and Transmission division to a new Swiss/Swedish giant in the power transmission industry. In August 1987, it had been announced that ASEA and BBC would merge, effecting January 1, 1988. The company’s name was to become Asea Brown Boveri (ABB). The next step towards creating another new heavy electrical giant was taken when it was made public in 1989 that GEC was to team up with the French electrical manufacturer Alsthom, and create a new power engineering venture: GEC-Alsthom. GEC had realized that they were to small to compete in the power industry. The French counterpart was twice the size of GEC, but the ownership of the venture was split 50/50. Hence, by 1990, the industrial scene in the business of power transmission equipment had been completely changed, and only Siemens remained of the big four. The American firms had gone for other industries, AEG had been forced to an almost complete withdrawal, however, old but small European firms had concentrated and were now a dominant factor in the industry. 1913 1929 1953 1975 1990 Sales Assets Sales Assets Sales Assets Sales Assets Sales Assets General Electric (GE) 71.5 (1910) 232 (1917) 376.2 (1930) 494 (1930) 2,234 (1950) 1,177 (1948) 14,105 58,414 Westing -house 165 (1917) 246 (1930) 1,448 (1955) 694 (1948) 6,164 10,700 (1987) AEG 463 202 579 294 658 4,596 451 (1962) Siemens 324 138 406 699 (1957) 740 6,381 60,000 GEC GEC 8 GEC 63 English Electric 31 AEI 42 GEC 232 English Electric 347 AEI 367 (1955) GEC 200 English Electric 196 AEI 254 GEC 1,867 Alsthom CGE 16 (1914) CGE 51 (1930) CGE 135 (1952) CGE 1,823 (1971) GECAlsthom 8,874 GECAlsthom 13,193 2 For a resembling account of the heavy segment of the electrical manufacturing industry, see Björkman (1999). Brown Boveri (BBC) 241 3,094 ASEA 5 7 13 48 212 (1955) 209 (1955) 1,886 2,382 ABB 26,700 ABB 32,038 Table 1. Firms and their sales, assets (in US $ millions) and number of employees (in thousands) in the evolution of the business of electrical transmission equipment (Hautsch 1982; Glete 1983; Rapp 1985; Chandler 1990; Sörman 1991; Catrina 1991; Frost 1991; Brummer and Cowe 1998; Annual reports GEC, AEI, and English Electric) Table 1 illustrates the size of the companies described in this study during the 20 century, by using assets as a proxy for firm size. The idea of this illustration is not to be precise on exact size or relative strength of each company, but rather to illuminate the changes discussed in brief above. Some interpretations As March and his associates point out, organizational learning tends to be difficult, in particular if viewing it from a Panglossian perspective. Learning tends to be local, and the electrical manufacturers showed systematic biases in interpretation, not at least when it came to interpreting technological evolution or market demand. Moreover, such local learning led to the creation of unreflected interpretation schemes. Such schemes were manifest in the “obviousness” of the superiority of technological systems. Thomas Edison was first adamantly opposed to direct current (DC) systems, and finally driven out of the heavy electrical industry, through his loss of control in Edison Electric (Hughes 1983; David 1992). Equally unrivalled was the “universal system”, i.e. alternating current, introduced by Westinghouse and other firms, as the final solution to this technical problem (Prout 1921). In Britain, engineers stood firm on arguing the superiority of direct current solution, and there was substantial inertia before the AC system could break through (Byatt 1979; Hughes 1983). However, there was not in Britain the DC system would revive again for heavy electrical purposes, but in Germany and Sweden. Another area where learning has promoted myopic behavior, is in the way various managerial principles have emerged. Organizational structures promoting independent and short-term behavior in firms like GE, GEC and Westinghouse. In dynamic environments of learning organizations, as the one of the electrical manufacturing industry, a pertinent finding is the ambiguity of success. Financial and technological success did not always go hand in hand. Moreover, the lack of “perfect” competition in the industry made it difficult for management of the electrical manufacturers to assess the efficiency of operations. Hence, it is not hard to understand why the electrical manufacturers exhibited significant inertia in their behavior. In general, they adhered to strategies, organizational structures, preferred markets, and technologies longer than what, looking in the rear view mirror, may have been optimal for longterm profitability and survival. Moreover, the three-partitioning of organizational capabilities into different activities indicates multilevel learning. Within each activity, a number of specialized routines emerged in response to problems in need of solutions. For instance, Siemens in the early history of the industry paid attention primarily to dynamo technology rather than lighting technology, GE developed products and organization suited for the demand of American utilities and consumers, and GEC (under Hirst) developed an organizational system where manufacturing and marketing departments “traded” rather than were centrally coordinated. However, the firms were in large successful, and promoted the further use of previously instigated routines. It is possible that superior routines could have been discovered, but why change a winning team? When difficulties arrived (e.g., the depression in the early 1930s or the decreasing demand in the 1970s), no other knowledge on causation was developed than the one based on the utilization of previously successful routines. It has been possible to discern patterns of competency (or success) traps in the electrical manufacturing firms in, for instance, the promotion of a universal technological system, the excess use of Strategic Business Unit organization, or reliance on particular customer segments. Hence such specialization implies that substitution effects are possible. Learning facilitation through learning specialization and substitution has been particularly pertinent, and point directly to the learning activities involved in organizational capabilities. Thinking in terms of the three basic activities proposed in this general framework, management-based, technology-based, and market-based activities may substitute for each other in the organization’s development of its organizational capability. Being technologically strong (e.g., in terms of patents), for example, during some periods may relieve firms from developing their market-based and management-based activities. This implies that the organization develops an “absorptive capacity” (Cohen and Levinthal 1990), within this particular activity, conducive for faster learning. However, such learning then takes place at the expense of learning in other activities. Such “imbalances” in learning dynamics may create “incoherence” of the industrial firm, as there are discrepancies in learning rates and direction of different activities. Complementing learning substitution, simplification is another means of facilitating learning from experience, which also induces localized learning. Here, we may observe the partitioning of organizations into different specialized departments as an indicator of this strategy. The principle used for departmentalization e.g., functional grouping (ASEA and BBC) or market/product (GE and GEC) tends to focus the direction of learning. Departmentalization creates boundaries and buffers restricting the searchable area of knowledge necessary. Within functionally organized units expertise knowledge can be enhanced, while the focus of market grouping increases pressures for responsiveness in system-wide sub unit problem solving (cf. Lindkvist, Söderlund and Tell 1998). In the history of the electrical manufacturing industry the British company Associated Electric Industries AEI kept its two major brand names Metrovick and British Thomson-Houston (BTH) in separate organizations until 1960. Both companies were considered technological leaders in Great Britain, had well established channels to their markets and were making money. As a strategy for enhancing the technologybased and market-based activities on a local level, the strategy of AEI seemed defendable. However, the general management of AEI was weak e.g., it was lacking a management accounting systems for the group as a whole. The potentials of knowledge integration efficiencies and scope attainable on a global level was not realized as “a learning curve” on these activities had not been embarked upon. A similar reasoning is applicable for the Swiss manufacturer Brown Boveri (BBC), which strove for technological excellence and subsidiaries with large home markets, but with little central coordination. The story of Westinghouse in the late 1970s and 1980s is quite a different story, where the development of refined managerial systems substituted for a coherent technology strategy and solid attempts of reaching global markets. Finally, the Swedish manufacturer ASEA during the 1960s and 1970s focused very much inwards on internal efficiency of its organization, increased its spending on R&D;, but lost its close cooperation with Vattenfall and did not build an organization directed toward international sales. There are at least two forms of learning substitution related to corporate coherence that seems to have been taking place in the electrical manufacturing industry. The first may be denoted vertical substitution, and refers to instances where “lower” and “higher” levels of routines become substituted. This was for instance the case in Westinghouse during the 1970s and 1980s, where operational focus and synergies where lost in favor of grand strategic projects. An opposite case is Swedish manufacturer ASEA at about the same time, which was highly focused on operational efficiency, but lacked strategic ability. The second type of learning substitution that tend to occur is horizontal substitution, i.e. the substitution of one set of routines on one level for another set of routines at the same level. The problems of sticking to management, market, or technology focus in respect to a certain product area may be one example of such dynamics. To sum up, the concept of coherence as discussed by Dosi, Teece and Winter (1992) does thus not only refer to the coherence in products per se, but also involves the balancing of learning activities. In this sense, products can be regarded as the outcome of intricate organizational learning processes (Rumelt 1994). The issue of balancing indicates that coherence entails a “juggling” of learning activities (cf. Pettigrew and Whipp 1991), for instance with regard to trade-off between the exploitation of current routines and the exploration of new routines (March 1991). Of course the learning perspective suggested here suffers from serious limitations. For instance it does explicitly not consider such important dimensions as institutional (regulatory) dimensions and access to finance, and power. However, the perspective points to one, very Chandlerian, way at looking at the development of corporate strategy in business history. An explicit learning agenda on the study of business history, and in particular when studying science-based firms producing complex product systems, as is the case for the heavy electrical manufacturing industry, there are a number of issues to be dealt with. First, there is a question of how learning dynamics and corporate coherence affect the ability of keeping the firm together whenit can be considered a system of distributed knowledge (Dosi and Marengo 1994;Tsoukas 1996). When highly specialized areas of expertise are to be managed, inparticularly in a context of innovation, it would be very interesting to further study (ina historical perspective) the role of centralized and decentralized knowledge, and therole of general as well as specific learning. Building on the suggestions in this paper,there are reason to suspect that highly specialized learning trajectories will evolveover time, creating problems of knowledge coordination. Another important issuerelates to the role of deliberation in learning processes. To what extent have firms’strategies in, for instance heavy electrical engineering been the outcome of planned,conscious knowledge accumulation and decision-making and what is the role of moreautomatic (routine-based) forms of learning (Gavetti and Levinthal 2001; Prencipeand Tell 2001; Zollo and Winter 2002). Extending that question, there is alsoimportant studies to be made of “dynamic capabilities” (Teece and Pisano 1994;Eisenhardt and Martin 2000), i.e. the development of an ability for renewal andreconfiguration of organizational capabilities. ReferencesArthur, Brian (1994), Increasing Returns and Path Dependence in the Economy, AnnArbor, MI: The University of Michigan PressByatt, I. C. R. (1979), The British Electrical Industry 1875-1914: The economicreturns to a new technology, Oxford, UK: Clarendon PressCarlson, W. Bernard (1991), Elihu Thomson and the rise of General Electric, 1870-1900, Cambridge, UK: Cambridge University PressChandler, Alfred D. Jr. (1977), The Visible Hand: The Managerial Revolution inAmerican Business, Cambridge, MA: The Belknap Press of Harvard UniversityPressChandler, Alfred D. Jr. (1980), The Growth of the Transnational Industrial Firm in theUnited States and the United Kingdom: A Comparative Analysis, EconomicHistory Review, Vol. 33, pp. 396-410Chandler, Alfred D. Jr. (1984), Comparative business history, in: Coleman, D. C. andPeter Mathias (eds.), Enterprise and history: Essays in honour of Charles Wilson,Cambridge, UK: Cambridge University Press Chandler, Alfred D. Jr. (1988), Business History as Institutional History, in: McCraw,Thomas K. (ed.), The Essential Alfred Chandler: Essays Toward a HistoricalTheory of Big Business, Boston, MA: Harvard University PressChandler, Alfred D. Jr. (1990), Scale and Scope: The Dynamics of IndustrialCapitalism, Cambridge, MA: The Belknap Press of Harvard UniversityCohen, Wesley M. and Daniel A. Levinthal (1990), Absorptive Capacity: A NewPerspective on Learning and Innovation, Administrative Science Quarterly, Vol.35, pp. 128-152Cyert, Richard M. and James G. March (1963), A Behavioral Theory of the Firm,Englewood Cliffs, NJ: Prentice HallDavid, Paul A. (1992) Heroes, Herds and Hysterisis in Technological History:Thomas Edison and ’The Battle of the Systems’ Reconsidered, Industrial andCorporate Change, Vol. 1, No. 1, pp. 129-180Dosi, Giovanni and Luigi Marengo (1994), Some Elements of an EvolutionaryTheory of Organizational Competences, in: England, Richard W. (ed.),Evolutionary Concepts in Contemporary Economics, Ann Arbor, MI: TheUniversity of Michigan PressDosi, Giovanni, David J. Teece, and Sidney Winter (1992), Toward a Theory ofCorporate Coherence: Preliminary Remarks, in: Dosi, G., R. Giannetti, and P. A.Toninelli (eds.), Technology and Enterprise in a Historical Perspective, Oxford,UK: Clarendon PressDosi, Giovanni, Richard R. Nelson and Sidney G. Winter (2000), Introduction, in:Dosi, G., R. R. Nelson and S. G. Winter (eds.), The Nature and Dynamics ofOrganizational Capabilities, Oxford University PressEisenhardt, Kathleen M. (1989), Making fast strategic decisions in high-velocityenvironments, Academy of Management Journal, Vol. 32, No. 3, pp. 543-576Eisenhardt, Kathleen M. and Jeffrey A. Martin (2000), Dynamic Capabilities: WhatAre They?, Strategic Management Journal, Vol. 21, pp. 1105-1121Foss, Nicolai J. (1999), Research in the Strategic Theory of the Firm: ‘Isolationism’and ‘Integrationism’, Journal of Management Studies, Vol. 36. No. 6, pp. 725-755Gavetti, G. and D. Levinthal (2000) Looking forward and looking backward:Cognitive and experiential search. Administrative Science Quarterly, 45, 113-137Granstrand, Ove and Sören Sjölander (1990), Managing innovation in multi-technology corporations, Research Policy, 19: 35-60 Helfat, Constance E. (2000), Guest editor's introduction to the special issue: theevolution of firm capabilities, Strategic Management Journal, 21: 955-959Hughes, Thomas P. (1983), Networks of Power: Electrification in Western Society1880-1930, Baltimore, MD: The John Hopkins University PressKönig, Wolfgang (1996), Science-Based Industry or Industry-Based Science?Electrical Engineering in Germany before World War I, Technology and Culture,Vol. 37, pp. 70-101Levinthal Daniel A. (1995), Strategic management and the exploration of diversity,in: Montgomery, C. A. (ed.), Resource-Based and Evolutionary Theories of theFirm – Towards a Synthesis, Boston, MA: Kluwer Academic PressLevinthal Daniel A. (2000), Organizational Capabilities in Complex Worlds, in: Dosi,G., R. R. Nelson and S. G. Winter (eds.), The Nature and Dynamics ofOrganizational Capabilities, Oxford University PressLevinthal, Daniel A. and James G. March (1993), The Myopia of Learning, StrategicManagement Journal, Vol. 14 (Winter Special Issue), pp. 95-112Levitt, Barbara and James March (1988), Organizational Learning, Annual Review ofSociology, Vol. 14, 319-340Lindkvist, Lars, Jonas Söderlund and Fredrik Tell (1998), Managing ProductDevelopment Projects – On the Significance of Fountains and Deadlines,Organization Studies, Vol. 19, No 6, 931-951March, James G. (1991), Exploration and Exploitation in Organizational Learning,Organization Science, Vol. 2, No. 1, 71-87Nelson, Richard R., and Sidney G. Winter (1982), An Evolutionary Theory ofEconomic Change, Cambridge, MA: The Belknap Press of Harvard UniversityPressPavitt, Keith (1984); Sectoral patterns of technical change: Towards a taxonomy and atheory, Research Policy, Vol. 13, pp. 343-373Pavitt, Keith (1998), Technologies, Products and Organization in the Innovating Firm:What Adam Smith Tells Us and Joseph Schumpeter Doesn’t, Industrial andCorporate Change, Vol. 7, pp. 433-452Patel, Pari and Keith Pavitt (1997), The technological competencies of the world’slargest firms: complex and path-dependent, but not much variety, Research Policy,26: 141-156 Pettigrew, Andrew M. (1990), Longitudinal Field Research on Change: Theory andPractice, Organization Science, Vol. 1, No. 3, pp. 267-292Pettigrew, Andrew M. and Richard Whipp (1991), Managing Change for CompetitiveSuccess, Oxford, UK: BlackwellPiscitello, Lucia (2000), Relatedness and coherence in technological and productdiversification of the world’s largest firms, Structural Change and EconomicDynamics, 11: 295-315Porter, Michael E. (1980), Competitive Strategy – Techniques for AnalyzingIndustries and Competitors, Free Press, New YorkPorter, Michael E. (1996), What is Strategy?, Harvard Business Review, Vol. 74, pp.61-78Prencipe, Andrea and Fredrik Tell (2001), Inter-project learning: processes andoutcomes of knowledge codification in project-based firms, Research Policy, Vol.30, No. 9, pp. 1373-1394Prout, Henry G. (1921), A Life of George Westinghouse, New York, NY: TheAmerican Society of Electrical EngineersRosenberg, Nathan (1982), Inside the Black Box – Technology and Economics,Cambridge, MA: Cambridge University PressRumelt, Richard P. (1972), Strategy, structure and economic performance,Cambridge, MA: Harvard University PressSimon, Herbert A. (1976), Administrative Behavior – A Study of Decision-MakingProcesses in Administrative Organzation, 3rd ed., Free Press, New YorkTeece, David J. and Gary Pisano (1994), The Dynamic Capabilities of Firms: AnIntroduction, Industrial and Corporate Change, Vol. 3, No. 3, pp. 537-556Teece, David, Richard Rumelt, Giovanni Dosi and Sidney Winter (1994),Understanding corporate coherence: Theory and evidence, Journal of EconomicBehavior and Organization, 23: 1-30Tsoukas, Hardimos (1996), The Firm as a Distributed Knowledge System: AConstructionist Approach, Strategic Management Journal, Vol. 17 (Winter SpecialIssue), pp. 11-25von Tunzelmann, G. N. (1995), Technology and Industrial Progress, Aldershot, UK:Edward ElgarZollo, M and S. G. Winter, 2002. Deliberate learning and the evolution of dynamiccapabilities, Organization Science, forthcoming