A PROCESS MODEL OF MNC EVOLUTION: THE CASE OF SONY CORPORATION IN THE UNITED STATES

 Working Paper 95-9

The authors gratefully acknowledge the Division of Research, Harvard Business School, and the Carnegie Bosch Institute for Applied Studies in International Management, for their generous funding of this research.
 

We use an embedded case study of Sony Corporation in the United States to develop a process model of the evolution of the multinational corporation (MNC). We model evolution along two dimensions: functional migration within a given line of business (LOB), and sequential entry by additional LOBs. These, in turn, are driven by intra-organizational processes of capability development. We describe MNC evolution as a series of recurring loops of capability development, with capabilities accumulated in early loops making possible subsequent capability development. Impediments and facilitators to the process of MNC evolution, and questions for further study, are identified.

INTRODUCTION

 Most MNCs begin by operating in a single country, and evolve into complex, multicentered organizations. Effective management of MNC evolution is likely to have important performance implications, as firms that develop strong multinational positions more fully and more quickly can better take advantage of operational flexibility (Kogut, 1985) and take advantage of the capabilities of each subsidiary (Bartlett and Ghoshal, 1986). Empirical evidence has shown that MNCs with geographic and product diversity frequently exhibit superior returns (Geringer, Beamish, and de Costa, 1989). Understanding the process of MNC evolution, and identifying impediments and facilitators to this process, is therefore important. Yet as noted by Malnight (1995), little attention has been devoted to the process by which MNCs become differentiated networks. Whereas it has long been recognized that MNC evolution is a lengthy and difficult process (Perlmutter, 1969), little research has focused on the managerial processes that guide MNC evolution. 

 This article uses an embedded case study design to develop an evolutionary model of the MNC. We seek to identify the process by which foreign subsidiaries develop from initial entry to become complex and internally differentiated players within the broader MNC network. We identify two dimensions of evolution: first, the migration of functions performed within a given line of business, and second, the addition of new lines of business. Evolution along these two dimensions is made possible by the accumulation of two kinds of subsidiary capabilities: those associated with a specific line of business, which make possible functional migration, and those specific to competing in the host country, which make possible the addition of new lines of business. We describe capability development in MNCs not as a one-time transfer or as a linear process, but as a series of recurring loops. By describing this process in detail, we seek to identify some of the factors that can facilitate or impede MNC evolution. 
 
 

THEORETICAL FRAMEWORK AND RESEARCH QUESTIONS

 A great deal of research has focused on the ways that firms expand abroad, known as foreign direct investment (FDI). Research has examined the conditions that lead to FDI (Dunning, 1988; Magee, 1981), patterns of FDI by industry (Caves and Mehra, 1986; Kogut and Chang, 1991) and by host country (Ajami and Ricks, 1981), the choice of entry mode (Kogut and Singh, 1988; Hill, Hwang, and Kim, 1990), and choice of subsidiary ownership structure (Anderson and Gatignon, 1988). These studies have identified many of the factors which explain FDI, but they have not examined the challenges of managing foreign subsidiaries following entry. 

 Considerable research has also been conducted into the management of MNCs as complex organizations. These studies have examined the varying roles played by national subsidiaries (Bartlett and Ghoshal, 1989; Jarillo and Mart'nez, 1990), knowledge flows among subsidiaries (Gupta and Govindarajan, 1991), headquarter-subsidiary governance structures (Ghoshal and Nohria, 1989), diffusion of innovation among subsidiaries (Ghoshal and Bartlett, 1988), and inter-unit communication (Ghoshal, Korine, and Szulanski, 1994). This body of research mainly includes cross-sectional studies of firms that are already multinational, but does not shed light on the process by which they became differentiated networks. Even typologies of subsidiary roles that imply a developmental trajectory, (e.g., Bartlett and Ghoshal, 1989; Jarillo and Martinez, 1990) offer little explication of the process by which a subsidiary evolves from one role to another. 

 Others have focused explicitly on the process of MNC development, known as internationalization theory. Johanson and Vahlne (1977) surmised that MNCs tend to expand from lesser to greater commitment, usually beginning with sales and marketing organizations, and building toward fully-integrated manufacturing. More recently, Malnight (1995) showed that functions increased in importance, first acting as appendages, and eventually becoming fully integrated within the MNC network. These and other views, while important, tend to focus on evolution within a single line of business. Yet most MNCs not only increase their commitment within a given line of business, they also add new lines of business in a foreign country. Chang (1995) provided empirical support that MNCs enter foreign markets in their core lines of business (LOB), then expand sequentially into other LOBs for which they have less competitive superiority versus local firms. 

 If we combine these two dimensions--migration of functions on one hand, and sequential LOB entry on the other--we can describe the broad outline of MNC evolution. Yet none of these offer a glimpse into the management process that drives evolution. Our focus, then, is on the process of management decisions and capability development that drives MNC evolution. The questions that motivate our study include these: 1) What enables MNCs to increase the functionality of subsidiary activities? What process of capability accumulation leads to functional migration within a given LOB? 2) What allows MNCs to enter in new lines of business? What process of capability development makes sequential LOB entry possible? 3) What impediments are confronted during the evolutionary process? 4) What actions can be taken by managers to facilitate the process of MNC evolution? 

 In a sense, our approach is related to the work of Kogut (1983), who first discussed foreign direct investment as a sequential process. Later, Kogut and Zander (1993:640) noted that "the sequential expansion of a firm's activities after first entry is a expression of the evolutionary acquisition and recombination of knowledge." We share that view, and seek to extend it by describing the process by which MNCs acquire knowledge and develop capabilities. Our approach is also similar to Doz and Prahalad (1991), who maintained that a process perspective, emphasizing organizational learning, held special promise for our understanding of the MNC, because a key feature of these organizations was their ability to learn from multiple products and geographic markets, and to transfer that learning across organizational subunits. Furthermore, their reference to "DMNCs"--diversified multinational corporations--explicitly brings into focus the line of business level, a key element in our approach. Unlike either of these antecedent works, however, we seek to elaborate a detailed process model. 
 
 

METHOD

 We decided to narrow the scope of study by examining multiple entries by a single MNC in a single foreign market. We chose the United States as the host market, and looked for a MNC that had a large and differentiated position in the United States in recent years, so that we might interview managers who had been directly involved in the evolution. Based on these criteria, we selected Sony Corporation, the Japanese-based electronics and entertainment firm. Sony is a mature global firm, with worldwide sales in 1993 of $34 billion distributed almost equally among North America (30.4%), Europe (26.0%), and Japan (25.8%). Its brand name ranks along with Coca-Cola and McDonald's among the most recognized in the world. 

 In the United States, Sony Corporation employs more than 20,000 people at many sites and in many distinct lines of business, including televisions, magnetic tape, semiconductors, and precision audio equipment. A summary of Sony's U.S. activities is shown on Table 1. Sony's many U.S. divisions play a variety of roles, some handling local procurement and manufacturing, while others are engaged in research and development, and play a role of worldwide leadership. 

Table 1
Sony Engineering and Manufacturing of America, 1994

1. Administrative headquarters, Park Ridge, NJ

2. Product divisions    

   Picture tube, color television, and display products activities      

       San Diego, CA        Employees: 2,000      Est.  1972    
       Tijuana, Mexico      Employees: 2,200      Est.  1987    
       Pittsburgh, PA       Employees:   800      Est.  1990    

   Audio Manufacturing Division, hi-fi speakers and audio racks 
       Delano, PA           Employees:    80      Est.  1975    

   Sony Professional Products Company, professional audio and video equipment
       Boca Raton, FL       Employees:   250      Est.  1982    

   Sony Microelectronics, semiconductor manufacturing   
       San Antonio, TX      Employees:   700      Est.  1990    

   Factory Automation Division, factory automation solutions    
       Orangeburg, NY       Employees:    20      Est.  1989

3.      Research and development centers

   Monterey, CA:   Advanced Computer Architectures Research Laboratories

   San Jose, CA:   Sony Microelectronics Design Center 
                   Semiconductor and Systems Laboratory
                   Sony Intelligent Systems Research Laboratory 
   
   San Diego, CA:  Display Systems 
                   Television Business Group of America 

   Boulder, CO:    Data Storage Laboratory

   Boca Raton, FL: Sony Professional Products Company   

   Orangeburg, NY: Factory Automation Division

Source: Sony company documents

Data Gathering

 Our meeting with the president of SEMA was fruitful in three ways: it resulted in permission to conduct the study, it identified key sites and provided introductions to key managers for subsequent interviews, and it provided initial interview data to complement the archival data already gathered. Over the next months, we conducted additional interviews at SEMA headquarters, and visited Sony divisions in San Diego, CA; Pittsburgh, PA; San Antonio, TX; and Tijuana, Mexico. A final series of interviews were conducted at SEMA headquarters in New Jersey and at Sony Corporation headquarters in Tokyo. 

 All data were gathered in 1994 and 1995. We interviewed more than forty Sony managers, some more than once. They included a cross-section of managers: Americans and Japanese; senior executives, division managers, and technical professionals; and included functions including R&D, manufacturing, human resources, finance, and business planning. Almost all interviews were conducted with both authors present. This proved to be important for thorough note-taking, for asking the best follow-up questions, and for subsequent synthesis and analysis. In conducting field interviews and reviewing our data, we found that two heads were indeed better than one. 
 
 

Data Analysis and Review

Limitations of the Data

 As for Sony's several lines of business, listed on Table 1, we did not attempt to study each in equal depth, but focused on two of the earliest and most important LOBs (televisions and magnetic tape), and on two of the most recent entries (data storage and personal telecommunications). We thereby were able to focus on a subset of LOBs which were representative of the larger set, providing an implicit comparison of four examples of entry embedded within a single case study. 
 
 

FINDINGS

 Sales boomed in the late 1960s, in part due to Sony's advanced Trinitron color television. As Sony's U.S. revenues increased, however, it became concerned about potential vulnerability from trade frictions. In the early 1970s, Japanese television makers were charged with dumping, prompting a Department of Commerce investigation. Sony's chief executive, Akio Morita, fought the dumping charges, but also decided to reduce Sony's potential vulnerability to trade frictions by setting up a manufacturing site in the United States. 

 Initial entry to the United States

 In August 1972, Sony established a small manufacturing plant in San Diego, CA. Thirty Americans were hired, and supervised by a team of Japanese expatriates. The plant built just one product-- televisions --and performed just one function -- final assembly. In fact, in the beginning San Diego merely re-assembled televisions that had already been assembled, adjusted, then disassembled, in Japan. So concerned were Japanese managers about the quality of manufacturing in the United States that at first they wanted only to make sure that Americans, working under the supervision of Japanese managers, could properly re-assemble a working television! 

Once Sony managers were satisfied that American workers could re-assemble televisions, they conducted final assembly without previous assembly in Japan. They also added more capacity, going from one line to eight. Still, only assembly was performed; all parts were shipped from Japan. Of these parts, the heaviest and most brittle were picture tubes. For that reason, in 1974 Sony began to manufacture tubes locally, saving on the cost and breakage of trans-Pacific shipments. Other key parts, including components and the unique Trinitron technology, continued to be sourced from Japan. 

 Entry in additional lines of business Based on its initial success in San Diego, Sony began to set up other U.S. plants in other major lines of business. In 1975, Sony built a factory to manufacture magnetic tape in Delano, PA. Magnetic tape, like televisions, was a line of business where Sony had a distinctive superiority, and where it could transfer product technology and manufacturing process technology to the U.S. and replicate its success. Next, in 1977, Sony built a magnetic tape plant in Dothan, Alabama. 

 In early 1980s, Sony leveraged its strength in CRT manufacturing to expand more broadly in computer display systems. The explosion in personal computer sales created a demand for high quality monitors, and Sony worked with IBM to supply the market. Later, Sony's display system technology expanded further, and by the 1990s included sophisticated wide screen monitors for air traffic control. 

 Functional migration in televisions

 During the late 1970s and early 1980s, Sony's San Diego plant grew in size, producing more and more televisions, yet continued to rely on key parts shipped from Japan. In addition, all key decisions about television strategy, product design, and business planning, were centered in Japan. 

 In the late 1980s, Sony began to emphasize "global localization," where products would be designed and manufactured locally rather than in Japan. In part, this shift was a natural evolution from "global" to "transnational" management (Bartlett and Ghoshal, 1989). It was also spurred by the rising value of the yen following the Plaza Accord of September 1985, as it became more important for Japanese-based firms to incur costs close to the point of sale, rather than in Japan. 

 For televisions, "global localization" meant that additional functions should be performed in the United States. Yet as one manager told us, "We couldn't just say, 'Let's move everything at once to the U.S.'" Rather, an incremental process took place, during which specific local capabilities were developed one after the other. One manager recalled that product design could not be conducted in the United States as long as the parts in those designs were procured in Japan. The first step, therefore, was local procurement. Sony managers in San Diego began to work actively with U.S. vendors to determine the specifications, quality, and delivery of key inputs. Once key inputs could be reliably procured in the U.S., certain design functions were then shifted to San Diego. This shift, managers recalled, was often sensitive, as it involved a net transfer of important responsibilities from Japan to the United States. 

 In time, key managers responsible for business planning were sent to San Diego, effectively bringing with them the entire infrastructure for strategic planning, and signalling a major shift of power. Key in this regard was the seihan process, which matched marketing requirements and factory allocations on a worldwide basis. Several managers told us that management of the seihan process is critical, because it establishes the operating capacities for factories around the world, affected performance evaluation, and therefore was a profound event in the maturation of the U.S. operation. 

 In 1988, Sony built a plant in Tijuana, Mexico, to perform assembly of its smaller televisions, which were relatively labor intensive and price competitive. In 1990, a new television plant was opened in Mt. Pleasant, PA, for the manufacture of large and projection televisions. Also in 1990, Sony's television sector was effectively divided into four geographic units, one of which, Television of America (TVA) assumed business planning responsibility within the United States. In 1994, TVA was responsible for all of the engineering of its models. Of the 25 to 30 new television models introduced each year, half were "productized" in the United States. A few key parts were still purchased from vendors in Japan, more than 70% of the dollar value of inputs was purchased locally.2 Small televisions were built in Mexico, 27" and 32" color televisions were built in San Diego, and large screen televisions were built in Pittsburgh. Functional migration in televisions had reached completion. 

 Further entry in additional lines of business

 At the same time that Sony's U.S. television activities were increasing in responsibility, new lines of business were being added. Sony's first entries to the U.S. had been in lines of business where it enjoyed a superiority over local firms, such as television and magnetic tape. Sony had leveraged that superiority by relying on greenfield entry, so that it could start the operation from scratch, determining the factory layout, bringing in key process equipment, and hiring its own employees. Not only had Sony initially limited the functions performed in the United States, it had also limited its activities to LOBs for which it was confident of success in the local market. 

 By the 1980s, as Sony's confidence about managing in the United States grew, it expanded into new lines of business, some of which were further removed from its core, and some of which involved acquisitions rather than greenfield investment. In 1982, Sony Professional Products was created from the acquisition of Music Center Inc. in Boca Raton, FL. In 1989, Sony acquired a Materials Research Corporation, a producer of thin metal film and etching equipment. In 1990, Sony acquired AMD's semiconductor equipment in San Antonio. In each of these instances, Sony was branching out into new LOBs and unafraid of acquisition, something it would never have contemplated in the mid-1970s. 

 At the San Diego plant, home to Sony's television operations, new lines of business were also being added. In 1991, Sony established the Data Storage Systems Division, which manufactured floppy disk drives, CD-ROM drives, and computer mass storage systems. At first, the division was set up to avoid import duties, and performed final assembly of key parts imported from Japan. Yet as technology changed and Sony's data storage products needed to draw on the latest advances, it was recognized that advances in data storage technology were coming from the United States, not Japan. Rather than perform only final assembly, the Data Storage Division quickly became a worldwide leader, setting up R&D labs in San Jose, CA, and Boulder, CO, where other major computer and information systems firms were located. By 1994, just three years after its establishment, Sony's Data Storage System Division in San Diego was responsible for R&D, product design, and manufacturing, and was expecting soon to become a full business unit, or jigyobu. 

 Most recently, Sony established in 1994 a Personal Telecommunication Division at the San Diego plant, seeking to gain a share of the rapidly growing digital personal communications market. Since 1991, Sony had produced analog cellular phones in Japan and imported them to the United States. By 1994, however, it was clear that digital technology was becoming dominant, and again the most advanced technology was in the United States, not Japan. Accordingly, the Personal Telecommunications Division was established, not as an offshoot of a dominant Japanese business unit, but with the expectation of spearheading Sony's worldwide personal telecommunications business. In an effort to jumpstart its digital cellular phone business, Sony established a joint venture with a local U.S. company, Qualcomm. The expectation was that Sony could provide manufacturing and marketing, to combine with Qualcomm's advanced digital technology. 
 
 

Summary of Findings

 In many ways, Sony's experience conforms with existing theory. The gradual migration of functions, from sales and marketing to manufacturing, and eventually to business planning, is consistent with models of progressive commitment in internationalization theory (Johansen and Vahlne, 1977). The initial tendency to rely on small scale greenfield investment, over time growing in confidence to undertaking acquisitions, is consistent with patterns of foreign market entry described by Root (1994). The increasing role of the United States in Sony's worldwide network is consistent with the typology of Bartlett and Ghoshal (1986, 1989), in which subsidiaries may begin as implementors, eventually become contributors, and may ultimately play a role of strategic leadership. Sony's television activities evolved from implementor to strategic leader over a period of many years; whereas data storage made the shift from implementor to strategic leader in just three years. 

 As for entry by line of business, Sony began in core lines of business such as televisions and magnetic tape, where it had a clear competitive edge over local American firms; over time entered in lines of business where superiority was less clear, such as semiconductors; and eventually entered in lines of business where it sought to tap U.S. technological leadership, such as advanced data storage and digital personal telecommunications. This pattern is consistent with the findings of Chang (1995). 

 Yet our study of Sony in the United States reveals much more than high-level patterns of functional migration or of entry by LOB. Based on our extensive interviews with managers, both Japanese and American, at business units and at SEMA headquarters, we also learned about the managerial process by which evolution along both dimensions took place. In this next section, we draw on our findings from Sony to develop such a model. 
 
 

A PROCESS MODEL OF MNC EVOLUTION

LOB-specific Capability Development

 At first, Sony managers determined the initial level of subsidiary functionality. Because they were mainly concerned with avoiding import barriers, their intention was to perform some limited functions in the U.S., so that televisions could be considered locally built rather than imported. Final assembly only was initially performed. 

 Once the initial level of functionality was determined, knowledge was transferred from Sony's television division in Japan to San Diego by Japanese expatriates, who set up the assembly process to match the process in Japan. These expatriates also trained local American workers. As Akio Morita recalled, expatriates were essential for training because job manuals had not been kept up to date in Japan (Morita, 1986:332). 

 In addition to receiving resources from the MNC parent, the subsidiary secures other resources locally. Plants were built or leased, equipment purchased and installed, employees hired, and local inputs sourced. For Sony, this involved the leasing of factory space at Rancho Bernardo, building the assembly line, and hiring local employees. Although Sony was inexperienced and was concerned about securing resources, it was assured by its neighbors, including Hewlett-Packard, NCR, and Burroughs, that hiring at the Rancho Bernardo Industrial Park was not difficult. Rather than hire people with experience in appliance manufacturing, however, Sony hired capable people with no related experience, so that it could train them from scratch.3 Also important was the ability to source local parts of the desired specifications or quality, and with delivery that was timely and reliable. By combining resources transferred from the parent with resources secured locally, Sony began operation, assembling kits into working televisions. 

 Evaluation and subsequent loops

 Once in operation, the San Diego plant was closely monitored: could American workers, supervised by Japanese expatriates, actually perform final assembly at an acceptable level? Once it was apparent that acceptable level of quality could be achieved, greater functionality was evaluated. In particular, the benefits of adding CRT manufacturing to final assembly were quite clear, given the transportation costs and breakage of CRTs. This triggered a second loop of functional migration, with additional knowledge transferred from Japan, engineers and technicians with expertise in CRT manufacturing sent to San Diego, and additional resources secured locally. By 1974 the San Diego plant had added CRT manufacturing to its assembly operations. 

 Over the next several years, Sony continued to design its televisions in Japan, and to source key inputs in Japan, with CRT manufacturing and final assembly in the United States. A number of factors impeded more rapid migration. Managers in Japan were concerned that local vendors could provide parts of sufficient quality and with sufficient reliability. They also were concerned about giving up key functions, including product design. By the late 1980s, however, the yen-dollar exchange rate pushed Sony toward greater localization of activities. New loops were undertaken, each of which involved the transfer of knowledge, securing local resources, and undertaking greater local activities. Performance was closely monitored, with satisfactory performance a precondition for further functional migration. Expanding local procurement was a discrete loop. Based on success in this activity, local design was carried out, representing another loop. Given success in local design, eventually business planning functions were shifted, including the seihan process. Finally, the creation of a self-sufficient business unit, or jigyobu, represented a final loop. Thus the functional migration of Sony's television activities in the United States can be understood as a series of loops, each one involving a discrete transfer of knowledge, local securing, building capabilities, evaluating capabilities, and deciding whether to pause or to continue. 

 Functional Migration within MNCs

 Based on the specific experience of Sony, we can develop more generally a model of functional migration in MNCs as a series of loops. The first steps in this loop are depicted in Figure 1A. 

As they enter new markets, MNCs seek to exploit their competitive advantages by transferring knowledge to the subsidiary and building its capabilities. At first, they determine the initial level of subsidiary functionality (Step 1). For Sony, the establishment of a plant in the United States was driven not by purely economic considerations, but by a fear of tariff barriers and import restrictions. For other firms, a wide variety of motivations exist, including location advantages (Dunning, 1988), access to low cost inputs of materials or of labor, or distinctive country capabilities (Kogut, 1991). Depending on these motivations, initial functionality may be limited to final assembly, or may involve more extensive and vertically integrated activities, including local sourcing or even product design. Initial functionality may also be influenced by the existence of other subsidiaries which may already perform key functions in that region of the world, making duplication inefficient. 

 Having determined the subsidiary's initial level of functionality, the MNC transfers resources to the subsidiary (Step 2). These may include financial resources as well as knowledge-intensive resources such as patents and formulas, proprietary processes or proprietary equipment, and managerial practices. When knowledge is explicit and codifiable, transfer may be achieved by documents and manuals (Kogut and Zander, 1993). When knowledge is complex and difficult to codify, MNCs often send expatriates to perform key functions and to train local employees, or may send local employees to the MNC parent for training. Effective transfer of resources is not always easy, but may be impeded by the choice of inappropriate means, such as a reliance on documents and manuals when side-by-side training is required. It may also fail if the firm devotes insufficient resources, such as expatriates. 

 In addition to transferred knowledge, resources are secured locally, including human resources and input parts (Step 3). Here, too, impediments abound. One of the most important aspects is the inability to attract talented local managers, perhaps because of a hesitancy to work for foreign-based firms. 

 By combining resources from the MNC parent and resources secured locally, the subsidiary begins operation (Step 4). Once in operation, the subsidiary works with local suppliers, performs proprietary manufacturing processes, and markets its output to local customers. In short, it gains experience competing in the local market. The result is an accumulation of capabilities, defined as skills, routines, and complementary assets, residing in the subsidiary's employees and its relationships with external entities, specific to that line of business (Teece, Pisano, and Shuen, 1990). 

 The process of subsidiary capability development, shown in Figure 1A, involves several steps and confront a variety of impediments, yet is not especially novel. Perhaps more interesting in our model are the subsequent steps of feedback, evaluation, and decision regarding further capability transfer. These are shown in Figure 1B. 

As the subsidiary carries out its initial functions, its performance is monitored by managers of the LOB at the MNC parent (Step 5). Through frequent communication with expatriate managers at the subsidiary, through regular financial reporting, through examination of output, and through discussions with customers, these managers evaluate the subsidiary's performance. 

 Evaluating subsidiary performance is a complex matter, involving multiple criteria. First is a comparison to other firms in the local market: how is the subsidiary doing relative to local firms? Second is a comparison to other subsidiaries elsewhere in the MNC. Even if a subsidiary is doing well compared to local firms, it may not be performing as well as a sister subsidiary in another country, and therefore may not merit greater responsibilities. Third, the subsidiary may contribute to the MNC in other than financial ways, such as by serving as a source of worldwide innovation, or as a training ground for new managers, or simply as a way to help absorb fixed costs. In some instances, the subsidiary may even serve a purpose by pinning a competitor in its home market, allowing the firm's subsidiaries elsewhere in the world to enjoy greater success. Evaluating subsidiary performance is therefore a multifaceted matter, and cannot be measured by a simple financial criteria. Even so, subsidiary performance is evaluated, with one of a few outcomes. If the MNC parent is not satisfied with the subsidiary's performance, it may shut down the subsidiary, or may invest additional resources to remedy any deficiencies. However, if the subsidiary is performing at or above some desired level, a first loop has been completed successfully. 

 MNC managers not only evaluate the subsidiary's present performance, but also assess the benefits of further expanding the subsidiary's activities (Step 6). This evaluation involves two components. First, the MNC must identify potential gains from greater functionality, such as cost advantages to be gained from expanding subsidiary manufacturing, location advantages of sourcing inputs locally, and benefits of performing local design. Second, the MNC must assess whether the subsidiary will be able successfully to carry out additional functions. It might conclude, for instance, that economic gains would accrue from sourcing more inputs in the local market, but that local vendors do not yet offer sufficient quality to justify additional functional migration. Alternately, the MNC might decide that it would be cost effective to perform a wider range of functions at the subsidiary, such as local design, but that subsidiary employees were not yet sufficiently skilled to carry out these additional functions. Several impediments may complicate this step. If the MNC does not possess sufficient information about the benefits of expanded functionality, it may overlook opportunities for further migration. Similarly, if information about subsidiary performance, both current and potential, is not accurately conveyed or well-understood, the MNC parent may fail to recognize opportunities for further functional migration. 

 Even if the benefits of subsidiary functionality are recognized, there must be a willingness to undertake further migration (Step 7). Impediments at Step 7 can be very strong. Managers at the MNC may resist additional migration out of a desire to retain key functions for themselves, such as basic R&D or strategic decision making. At some firms, requests by the subsidiary to adapt product specifications for local customer needs may be only slowly granted by the MNC parent, which is used to having exclusive control of product design. At other firms, the MNC parent might not allow any basic research to be performed at the subsidiary, arguing that a large investment had already been made at the parent site and that duplication at the subsidiary would be wasteful. 

 If Steps 5, 6, and 7 are completed, and the MNC decides to increase the functions performed by the subsidiary, a second loop of development begins, repeating the steps in Figure 1A but at a higher level. Additional resources will be invested, additional technology may be transferred, and additional resources will be secured locally, so that the subsidiary can take on greater functionality, perhaps in a more complex manufacturing process or local product development. Once again subsidiary performance is monitored by the MNC parent according to a variety of criteria, repeating the steps in Figure 1B. If further migration is desired a third loop may begin, leading to the development of still greater subsidiary capabilities. 

 The complete process of resource transfer, capability building, monitoring and evaluation, is shown as a loop in Figure 1C. Subsidiary evolution can undergo any number of loops. The MNC may cease the process after a few loops, perhaps determining that the subsidiary should migrate to a certain level of functionality but no further; or it may undertake many loops, continuing to transfer resources and migrate functions until the subsidiary becomes a self-sufficient business unit, performing all key business planning functions, or even until the subsidiary takes worldwide leadership for the LOB. 

Capability maintenance, erosion, and extension

 Figure 1C suggests that subsidiary capability development is an additive process, with knowledge-intensive resources from the MNC parent combined with locally secured resources to produce higher levels of subsidiary capability. In reality, of course, subsidiary capabilities are dynamic and cannot simply be added up. They may erode with disuse, or may be lost when key employees leave the firm, meaning that the subsidiary must continually use them to maintain and preserve them. On the other hand, the subsidiary may combine knowledge-intensive resources and locally secured resources to not only match the parent's level of performance, but surpass the capabilities of the parent. For example, local customers may require specific product adaptations, and subsidiary employees may find ways to improve existing processes. For these reasons, subsidiary capability development is a more complex process, as it involves not just recurring transfer from the parent, but continuing efforts to maintain capabilities (and avoid erosion) as well as possibilities for local extension. 
 
 

Country-specific capability development

 Over time, as its experience in the United States progressed, Sony developed a strong reputation as a good employer and as a good customer for local suppliers, learned about local regulations, and generally developed capabilities that were not specific only to televisions and magnetic tape, but could serve other lines of business as well. It also developed a strong country headquarters at Sony Engineering and Manufacturing America, where technical experts could assist all manufacturing units, and leverage the company's strengths across lines of business. Some additional lines of business, such as display systems, were related to the CRT business, and drew on Sony's existing capabilities. Other entries, however, were in LOBs where Sony did not have a strong advantage over U.S. firms, but sought to tap local expertise, as with the Data Storage Division and Personal Telecommunications Division. These recently founded divisions would never have been attempted in the 1970s, when Sony concentrated on exploiting its competitive advantages; rather, they were possibly only because of the existence of a strong country organization and the abundance of Sony capabilities--including management support, financial infrastructure, and engineering talent--that could be shared across units. 

 Sequential LOB entry as a series of loops

 The process of sequential LOB entry can also be modelled as a series of loops, as shown in Figure 2. As it develops capabilities in a given LOB, the subsidiary also develops capabilities that are associated more generally with the host country. These country-specific capabilities can be divided into two categories: operating synergies and administrative synergies. 

 Business synergies include physical sites, such as industrial parks or factories, that can accommodate additional lines of business; technical expertise that can be shared across LOBs; existing supplier relations, which can be used by new LOBs; and existing channels of distribution, which can accommodate products from new LOBs. The presence of these may allow a new LOB to more quickly overcome a liability of foreignness, and become competitive with local firms. Administrative synergies include such things as finance functions, such as treasury, tax, legal; health and safety departments, which can help support new LOBs at declining unit costs; contacts with governmental agencies and regulatory bodies; a human resource function, which provides not only central records and benefits administration, but also an improved ability to attract talented employees and to provide greater opportunities for their development. 

 As the subsidiary develops these country-based capabilities, other LOBs in the home country may perceive that they are now able to compete effectively in the host country, and may enter the country. With the addition of a new LOB, the subsidiary expands further in size and scope, gains additional experience, and expands its country-specific capabilities further. With a higher level of country-specific capabilities, the process can recur again, leading to still more LOB entries. As shown in Figure 2, this process can recur, with further entries building additional capabilities, leading to additional LOBs. 

The process of sequential LOB entry can continue until LOBs are added for which there is relatively little competitive advantage over local firms. Entry in these LOBs might have been impossible when the subsidiary was first established, but are now possible thanks to the increase in country-specific capabilities. 

 Once added, each new LOB undertakes its own process of LOB-specific capability development and functional migration, as was described above. As long as the parent MNC has a competitive advantage in the LOB, the MNC exploits its advantages in the host country, with the direction of knowledge-intensive resource transfer from the MNC to the subsidiary, as depicted in Figure 1A. However, once it has developed sufficient country-specific capabilities it may enter in LOBs for which it has no advantage over local firms. In such cases, its intention is not to exploit MNC capabilities but to tap host country capabilities. The tapping process involves the transfer of knowledge-intensive resource, but the direction is the reverse of Figure 1A; that is, the flow is from subsidiary to elsewhere in the MNC. 
 
 

Combining Functional Migration and Sequential LOB Entry

A Third Dimension: Inter-Regional Evolution

 Although Sony's activities elsewhere in the world are beyond the scope of this study, we can at least note that there appears also to be an evolutionary pattern across countries based on the accumulation of capabilities. The milestones of Sony's history, shown on Table 2, makes clear that the establishment of a television plant in the U.S. in 1972 was followed by establishment of a television plant in Wales in 1974; that the U.S. CRT operation in 1974 was followed up by a U.K. CRT operation in 1982; and that the U.S. magnetic tape plant, founded in 1977, was followed up by a similar plant in France in 1980. In each instance, the first plant was set up in the United States, with a European plant following just a few years later. Although we have not studied Sony's European activities in depth, we can suggest that the accumulation of capabilities in the United States led not only to additional functional migration and sequential LOB entry in the United States, but also to a replication of activities in Europe. 

Table 2 
Selected dates in Sony Corporation history

1946 May  Founding of Tokyo Tsushin Kogyo (Tokyo Telecommunications Engineering
          Company)
1950 Jul  Marketing of first magnetic tape recorder
1958 Jan  Name changed to Sony Corporation
1960 Feb  Founding of Sony Corporation of America
1972 Aug  Opening of color television assembly plant in San Diego, CA
1974 Jun  Opening of color television assembly plant in Bridgend, Wales, U.K.
1974 Aug  Opening of CRT plant at San Diego, creating the first integrated 
          color television production facility established by a Japanese firm.
1977 Feb  Opening of magnetic tape plant, Dothan, Alabama
1980 Dec  Opening of magnetic tape plant in Bayonne, Aquitaine, France
1982 Apr  Opening of CRT plant at Bridgent, Wales, U.K.
1982 Jun  Acquisition of Music Center Inc., Lt. Lauderdale, FL
1984 Sep  Opening of CD software plant in Terre Haute, Indiana
1987 Apr  Opening of plant for production of key components for CD players and
          8mm video equipment in Colmar, France
1987 Jul  Opening of CD software plant in Salzburg, Austria
1988 Jan  Acquisition of CBS Records, Inc.
1988 Apr  Opening of audio equipment plant, Penang, Malaysia
1988 Sep  Opening of audiocassette plant in Rovereto, Italy
1988 Sep  Opening of videotape plant, Bangkok, Thailand
1989 Sep  Acquisition of Material Research Corporation, U.S.
1989 Nov  Acquisition of Columbia Pictures Entertainment Inc.
1990 Feb  Acquisition of semiconductor manufacturing facility from Advanced 
          Micro Devices, San Antonio, TX

Source: Sony Corporation Annual Report, 1993, p.29

DISCUSSION AND QUESTIONS FOR FURTHER RESEARCH

 The case of Sony Corporation has been most useful in that its evolution in the United States was protracted and incremental, allowing us to observe a gradual process of change over many years. Going forward, a number of questions merit further study. Among the most promising questions for study are these:

 How do impediments change in successive loops?

 We have noted some of the impediments to MNC evolution at several different stages in the model. Yet impediments may differ at each level of functional migration. At initial stages, the MNC parent may simply want to be sure that the subsidiary can handle basic activities. At later stages, functional migration may encounter a different sort of resistance, as continued migration involves a shift of power to the subsidiary. Resistance may be strongest when functional transfer involves the transfer of LOB worldwide headquarters to the subsidiary, since such a step means the loss of headquarters and its associated power. More clearly identifying the nature of impediments and their relation to successive loops would be an important follow-up study. 

 What actions are most effective in facilitating evolution?

 Overcoming the impediments to MNC evolution is a topic of high managerial importance, and worthy of considerable study. A few major facilitators can be suggested here. First is the development of effective and compatible information systems, which allow subsidiary performance to be accurately assessed and its possibilities for further development realistically appraised. Such information is critical for Steps 5 and 6, when MNC managers must understand present and future subsidiary capabilities, prior to deciding whether to continue with migration. Second is trust among parent and subsidiary employees. Helping to offset misperceptions of subsidiary capabilities, as well as unwillingness on the part of the MNC parent to transfer further capabilities, is a climate of trust among employees. Trust has at least two elements: technical and personal. Technical trust involves confidence that the subsidiary has the skill to take on increased functionality. Managers can help create this trust by providing forums to demonstrate technical skill, such as joint development projects, worldwide technical exchanges, and worldwide LOB councils which allow managers to learn about each other's technical competencies. A by-product of these face-to-face exchanges is personal trust, allowing managers to get to know each other, and perhaps making them more amenable to the migration of additional functions. Third, and related to the notion of trust, is the development of a global mindset, wherein a broader perspective replaces narrow parochial attitudes (Bartlett and Ghoshal, 1987). Fourth, and perhaps most important, is the presence of a strong champion. Very often, a critical factor in MNC evolution is a strong champion who not only understands the benefits of functional migration, but who has the power to effect a shift, overcoming resistance from within the organization. At Sony, one strong champion--Akio Morita--was key in recognizing the importance of the United States and in making the initial decision to establish manufacturing there; and another--Kunitake Ando--was critical in pushing for increased responsibilities in the U.S. in the 1990s. Even with compatible information systems and mechanisms to build trust among countries, it is hard to imagine effective MNC evolution at Sony without key individuals pushing actively for change, and overcoming natural tendencies for resistance and inertia. 

 Are many small loops better than a few large loops?

 MNCs may also differ in the extent of capability development in any given loop. Sony undertook many loops, each of which involved a small increase in subsidiary functionality, and monitored the results of each loop carefully before beginning a next loop. Other MNCs may be relatively more ambitious, expanding the activities of their subsidiaries by taking a fewer number of large steps. Although either approach can be understood in terms of Figure 1C, with a transfer of resources, subsidiary capability development, feedback and possible further transfer, the managerial challenges of many gradual loops versus a few large loops are likely to be very different. In the former instance, many loops may reflect a careful and deliberate approach, as we saw with Sony, or could suggest an unnecessarily lengthy and protracted process, and providing many invitations for impediments. In the latter instance, attempting to compress evolution into just a few large steps runs the risk of asking the subsidiary to undertake a role for which it is not fully ready, or of triggering major resistance from within the MNC. Tentatively, then, we may hypothesize that optimal performance may involve an avoidance of either extreme. 

 Are there national patterns in MNC evolution?

 It is not surprising that a Japanese-based MNC has provided a clear example of MNC evolution. As noted by Kagono and colleagues, Japanese firms are known for taking an "inductive, gradual adjustment approach" to their strategic development (Kagono, Nonaka, Sakakibara, and Okumura, 1985:83). Driving this evolutionary process is the "continuous accumulation and development of in-house resources" through "learning and accumulated expertise" (1985:151), again very much in keeping with the model set forth here. 

 By contrast, it has been suggested that European firms are more likely to take a few bold steps, rather than a gradual approach (Bouchikhi and Kimberly, 1991). Whether there are important national patterns in the process of MNC evolution is a topic worthy of study, and a natural extension to this single case study of a Japanese-based firm. 

 Additional topics for study may be suggested, as well, including differences between global industries and multidomestic industries (Porter, 1986), and differences among host countries. Many extensions are possible once a basic model has been articulated. The major contribution of this process model has been to distinguish between different dimensions of MNC evolution, to conceive of evolution along both dimensions as a series of loops, to identify specific steps in each loop, and to suggest impediments at various steps. Understanding how the basic model is manifested in firms other than Sony, and in other industries and other host countries, awaits our further study. 

NOTES

2. "Global localization" did not suggest that eventually 100% of inputs would be sourced locally. Some television inputs were most efficiently produced in low wage countries of South East Asia, and were expected to be imported to the United States for the forseeable future. 

3. Many other Japanese firms later did the same thing, including Honda Motor Company at its Marysville, Ohio, plant. These firms were keen to train employees from scratch, and not struggle to change existing habits. 

REFERENCES

 Bartlett, Christopher A. and Sumantra Ghoshal, Managing Across Borders: The Transnational Solution, Boston: Harvard Business School Press, 1989 

 Bartlett, Christopher A. and Sumantra Ghoshal, "Tap your subsidiaries for global reach," Harvard Business Review, November-December 1986, 87-94 

 Birkenshaw, Julian, "Approaching heterarchy: a review of the literature on multinational strategy and structure," Advances in International Comparative Management, 9:111-144, 1994 

 Bouchikhi, Hamid, and John R. Kimberly, "Clausewitz contre Sun Tzu: La conquete du marche Americain," Gerer et Comprendre, June 1991 

 Caves, Richard E. and Sanjeev K. Mehra. "Entry of foreign multinationals into U.S. manufacturing industries." In Porter, Michael E. (ed.), Competition in Global Industries, Harvard University Press, Boston, 1986, pp. 449-481. 

 Chang, Sea-Jin "International expansion strategy of Japanese firms: capability building through sequential entry," Academy of Management Journal, 383-407, 1995 

 Doz, Yvez L., and Prahalad, C.K., "Managing DMNCs: a search for a new paradigm," Strategic Management Journal, 12:145-164, 1991 

 Dunning, John H. Explaining International Production, London and Boston: Unwin Hyman, 1988 

 Geringer, J. Michael, Paul W. Beamish, and Richard C. daCosta. "Diversification strategy and internationalization: implications for MNE performance," Strategic Management Journal, 10, 1989, pp. 109-119. 

 Ghoshal, Sumantra and Christopher A. Bartlett, "Creation, adoption, and diffusion of innovations by subsidiaries of multinational corporations," Journal of International Business Studies, Fall 1988, pp. 365-388. 

 Ghoshal, Sumantra, Harry Korine, and Gabriel Szulanski. "Interunit communication in multinational corporations," Management Science, 40(1), January 1994, pp. 96-110. 

 Ghoshal, Sumantra and Nitin Nohria. "Internal differentiation within multinational corporations," Strategic Management Journal, 10:323-337, 1989 

 Ghoshal, Sumantra, and Christopher A. Bartlett, "The Multinational Corporation as an Interorganizational Network," Academy of Management Review, 15, 4:603-625, 1990 

 Glaser, Barney G., and Anselm L. Strauss, The Discovery of Grounded Theory, Chicago: Aldine, 1973 

 Gupta, Anil K. and Vijay Govindarajan. "Knowledge flows and the structure of control within multinational corporations," Academy of Management Review, 16(4), October 1991, pp. 768-792. 

 Hedlund, Gunnar. "The hypermodern MNC: A heterarchy?" Human Resource Management, 25:9-35, 1986 

 Hill, Charles W.L., Peter Hwang, and W. Chan Kim. "An eclectic theory of the choice of international entry mode," Strategic Management Journal, 11, 117-128, 1990 

 Hymer, Stephen H. The International Operations of National Firms: A Study of Direct Foreign Investment, Cambridge, MA: MIT Press, 1976 

 Jarillo, J. Carlos, and Jon I. Martinez, "Different roles for subsidiaries: the case of multinational corporations in Spain," Strategic Management Journal, 11:501-512, 1990 

 Johanson, Jan, and Jan-Erik Vahlne, "The internationalization process of the firm: a model of knowledge development and increasing foreign market commitments," Journal of International Business Studies, 8:23-32, 1977 

 Kagono, Tadao, Ikujiro Nonaka, Kiyonori Sakakibara, and Akihiro Okumura, Strategic versus Evolutionary Management: a U.S.-Japan Comparison of Strategy and Organization, Amsterdam: North Holland, 1985 

 Kim, W. Chan and Peter Hwang. "Global strategy and multinationals' entry mode choice," Journal of International Business Studies, 29-53, 1992 

 Kogut, Bruce "Foreign direct investment as a sequential process," in C. Kindelberger and D. Audretsch (Eds.), The Multinational Corporation in the 1980s, Cambridge, MA: MIT Press, 35-56, 1983 

 Kogut, Bruce "Designing global strategies: comparative and competitive value-added chains." Sloan Management Review, Summer:15-28, 1985 

 Kogut, Bruce, and Sea-Jin Chang, "Technological capabilities and Japanese foreign direct investment in the United States," Review of Economics and Statistics, 73:401-413, 1991 

 Kogut, Bruce and Udo Zander. "Knowledge of the firm and the evolutionary theory of the multinational corporation," Journal of International Business Studies, 24, 4:625-645, 1993 

 Magee, Stephen P. "Information and the multinational corporation: an appropriability theory of direct foreign investment." In Bhagwati, Jagdish N. (ed.), The New International Economic Order, MIT Press, Cambridge, MA, 1977, pp. 317-340. 

 Malnight, Thomas W. "Globalization of an ethnocentric firm: an evolutionary perspective," Strategic Management Journal, 16:119-141, 1995 

 Martinez, Jon I., and Jarillo, J.Carlos "The evolution of coordination mechanisms in multinational corporations," Journal of International Business Studies, 52, 489-514, 1989 

 Morita, Akio, Made in Japan: Akio Morita and Sony, New York: E.P. Dutton, 1986 

 Nohria, Nitin, and Sumantra Ghoshal, "Differentiated fit and shared values: alternatives for managing headquarters-subsidiary relations," Strategic Management Journal, 15:491-502, 1994 

 Perlmutter, Howard. "The tortuous evolution of the multinational corporation," Columbia Journal of World Business, 4:9-18, 1969 

 Porter, Michael E. "Changing patterns of international competition," California Management Review, 28:2, 9-40, 1986 

 Prahalad, C.K., and Doz, Yvez L. The Multinational Mission: Balancing Local Demands and Global Vision. New York: The Free Press, 1987 

 Pugel, Thomas A. "The United States." In Dunning, J.H. (ed.), Multinational Enterprises, Economic Structure, and International Competitiveness, John Wiley & Sons, Chichester, 1985, pp. 57-90. 

 Pugel, Thomas A. "The determinants of foreign direct investment: an analysis of U.S. manufacturing industries," Managerial and Decision Economics, 2, 4:220-228, 1981 

 Teece, David, Gary Pisano, and Amy Shuen, "Firm capabilities, resources, and the concept of strategy," Working Paper, University of California, 1990 

 White, Roderick E., and Thomas A. Poynter, "Organizing for world-wide advantage," in Managing the Global Firm, Christopher A. Bartlett, Yves Doz, and Gunnar Hedlund, eds., London: Routledge, 1990 

 Yin, Robert K. Case Study Research: Design and Methods, Newbury Park: Sage, 1989 

 Zaheer, Srilata, "Overcoming the liability of foreignness," Academy of Management Journal, 38, 2:341-363, 1995