The Process of Chinese Manufacturing

Home , Transitioning (Glee)

Global Strategy Journal Global Strat. J., 3: 109–125 (2013) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1111/j.2042-5805.2012.01046.x

THE PROCESS OF CHINESE MANUFACTURING UPGRADING: TRANSITIONING FROM UNILATERAL TO RECURSIVE MUTUAL LEARNING RELATIONS GARY HERRIGEL1*, VOLKER WITTKE2, and ULRICH VOSKAMP3 1Department of Political Science, University of Chicago, Chicago, Illinois, U.S.A. 2Soziologisches Seminar, Georg August Universität, Göttingen, Germany 3SOFI, Göttingen, Germany

This article argues that upgrading in China has been a historical success, that upgrading must be seen as a learning process, and that current Chinese upgrading involves a transformation in industrial learning dynamics. During China’s initial export-oriented industrialization strategy, indigenous producers successfully upgraded by apprenticing themselves to their foreign customers, and they learned through integration in transnational communities of practice. The success of those initial unilateral learning relations enhanced the sophistication of the Chinese market, both as a community of producers and as a market for manufactured goods. This has generated a new phase of learning-driven upgrading in which Chinese producers and MNC manufacturers both seek to make their Chinese operations more sophisticated. In this new context, apprenticeship disappears and Chinese and foreign players learn from one another. A core claim about the new mutual learning is that it is facilitated by the globalization of formal learning systems, such as corporate production systems (CPSs). Copyright © 2013 Strategic Management Society.

INTRODUCTION are leveraging their volume production expertise (which involves remarkable manufacturing flexibil- In its initial post-1992 industrialization push, ity) to move up the value chain into designing and Chinese manufacturing excelled in global markets developing their own (ever more sophisticated) prod- as a platform for high volume, low cost, export- ucts (Brandt and Thun, 2010; Shafaeddin and oriented production. (Steinfeld 2010; Lemoine, Pizzarro, 2010). 2000). Since China’s accession to the World Trade Developed country multinational corporations Organization in 2001, however, successful industri- (MNCs) from Asia, Europe, and the United States alization has created the home market conditions for are responding to the same emerging market possi- very different manufacturing strategies. In particular, bilities by deepening and enhancing their commit- more sophisticated domestic demand for manufac- ments in China (Brandt and Thun, 2010; Marsh, tured goods is shifting Chinese manufacturers’ stra- 2004, 2005a, 2005b, 2005c, 2007, 2008, 2009, tegic focus toward more advanced production. They 2010a, 2010b, 2011a, 2011b). Most significantly, they are upgrading local Chinese production, engineering, and design capabilities, as well as training Keywords: China; upgrading; organizational learning; gover- and supply chain infrastructures, to adapt their prod- nance; corporate production systems ucts to the specific technical, regulatory, and cultural *Correspondence to: Gary Herrigel, Department of Political Science, University of Chicago, 5828 S. University Avenue, characteristics of the Chinese market. Crucially, Chicago, IL 60637, U.S.A. E-mail: [email protected] these MNC projects do not involve unidirectional

flows: adapting/developing products for competitive mutual learning among interlocutors in MNCs, production in China involves at least as much learn- markets, and supply chains. In the new context, for- ing by MNCs from their local Chinese interlocutors mally organized joint inquiry and experimentation as the other way around. processes embedded in, among other things, CPSs, This article argues that upgrading in China has generate recursive learning by continually disrupting been a historical success, that upgrading must be and fostering creative recomposition of practice seen as a learning process, and that current Chinese communities. upgrading involves a transformation in industrial Our argument proceeds in five steps. First, we learning dynamics.1 During the initial export- briefly discuss the empirical basis upon which our oriented industrialization strategy, Chinese pro- argument rests, and we outline, broadly, the research ducers successfully upgraded by apprenticing program out of which the current material has been themselves to their foreign customers and learned taken. Second, we describe the pragmatism- through integration in transnational communities of informed theoretical perspective that drives our ren- practice (Brown and Duguid, 1991). The success of dering of the Chinese upgrading dynamic. Step three those initial unilateral learning relations enhanced outlines the initial industrial upgrading period in the sophistication of the Chinese market, both as a which learning was unilateral and shows that, ulti- community of producers and as a market for manu- mately, this apprenticeship learning encountered factured goods. This has generated a new phase of limits. Chinese producers developed capabilities learning-driven upgrading in which indigenous within transnational practice communities but Chinese producers and MNC manufacturers both foreign customers were either unwilling or unable to seek to make their Chinese operations more sophis- take advantage of them. This gave rise to a turn ticated. In this new context, apprenticeship disap- toward the domestic Chinese market. Step four then pears and Chinese and foreign players learn from describes the emergent Chinese market contours and one another. the strategies that both indigenous and foreign MNC A core claim we make about the new mutual learn- producers are pursuing to gain positions within it. In ing is that it is facilitated by the globalization of the final step, we show how these strategic efforts are formal learning systems, such as corporate produc- transforming the learning dynamics within the com- tion systems (CPSs). Originally developed to facili- munity of competitors. tate innovation and continuous optimization in MNC operations within competitive developed markets, Empirical basis of our upgrading argument CPS principles have diffused into China via transnational supply chains and intra-MNC governance pro- Our argument is about manufacturing upgrading as a cesses (Sabel, 2005; Spear, 2009). Global CPSs are learning process. We draw broadly from the second- governance architectures that generate joint reflec- ary empirical literature on manufacturing experience tion, deliberation, and experimentation. When suc- in China, especially in the electronics, automobile, cessful, this produces multidirectional (recursive) and complex machinery sectors (including their learning among all participants. supply chains). More fundamentally, our argument Ultimately, we make three major points: (1) draws on nearly a decade’s worth of primary manufacturing upgrading has taken place in China research in Germany, the United States, Central over the last two decades; (2) initially, upgrading in Eastern Europe, and China on automobile and China occurred through predominantly unidirec- complex machinery industry production strategy and tional learning relations. Chinese producers were supply chain recomposition. In these projects, our integrated into transnational communities of manu- focus has been on understanding how changing cus- facturing practice in which they learned from foreign tomer firm product and production strategies involve interlocutors; (3) in the last decade, as domestic learning in cooperative and competitive supply chain manufacturing demand sophistication grew, Chinese relations. Under this rubric, in China alone we have upgrading processes transitioned from unilateral conducted nearly 100 face-to-face interviews with learning within stable communities of practice to managers in indigenous Chinese, foreign MNC foreign direct investment (FDI) and joint venture (JV) facilities. These China interviews are comple- 1 For other perspectives on upgrading as learning, see mented by extensive interviews with German and Mudambi, 2008; Kumaraswamy et al., 2012. U.S. managers in home country locations across the

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 111 supply chains in both the automobile and machinery This sociological literature is limited, however, sectors.2 Although we make an effort to contextual- because it reduces learning to the transfer of techno- ize our observations in the relevant findings available logical know-how or knowledge about technologies in the secondary literature, most of the empirical or even specific products. Moreover, even though assertions and specific illustrations we adduce below they emphasize the role MNCs play in upgrading are drawn from this (ongoing) interview-based processes, the literature focuses primarily on one- research. way flows from MNCs to the upgrading producers in the emerging economy. As a consequence, they leave the possibility that emerging economies have some- Theoretical framework: sociological approaches thing to teach developed country players relatively to upgrading and pragmatic approaches to undertheorized. organizational learning We believe the practice-based organizational Our approach in this article builds on the advances learning literature (e.g., Brown and Duguid, 1991; that sociological approaches to upgrading, especially Blackler and McDonald, 2000) better addresses the relational- and networked-based work, have made to learning processes observed in China by shifting our understanding of processes of industrial upgrad- attention from the discrete nuggets of knowledge ing in developing countries. This literature moves or technology that firms ‘learn about’ to the formal beyond alternative efforts to root upgrading pro- and tacit interactive practices—communities of cesses in either traditional incentive alignment and practice—that firms become involved in. In the next market forces-based arguments (e.g., Moran et al., step, we show that gradual Chinese producer inte- 2005) or technologically determinist arguments gration into transnational communities of practice about how the modular characteristics of contempo- with foreign customers taught them how to be rary manufacturing technologies limit possibilities reliable and competent manufacturing exporters. for supplier and emergent producer firm upgrading Foreign customers apprenticed their Chinese suppli- (Sturgeon, 2002; Lüthje, Schumm, and Sproll, 2002; ers by showing them how to meet ever more exacting Gereffi, Humphrey, and Sturgeon, 2005; Sturgeon manufacturing and commercial standards. This was and Florida, 2005; Giuliani, Pietrobelli, and Rabel- done iteratively over time, through myriad contracts, lotti, 2005; Humphrey and Memedovic, 2003; audits, supplier quality assurance encounters, com- Breznitz and Murphree, 2011). The virtue of the parisons with competitors, and conferences over sociological accounts is that they identify specific detected errors and possible process improvements relations (especially interactive, non-arm’s-length (Lui, forthcoming; Ivarsson and Alvstam, 2010). relations in supply chains) and environmental condi- The main point of these interactions was not so tions (interconnected clusters of normally region much the specific substance or content of any given based producers and supporting educational and technological or knowledge exchange as it was the policy institutions and associations) as preconditions development of the Chinese firm’s capacity to learn for successful upgrading that the nonsociological how to identify the changing production quality and perspectives tend to ignore or underestimate (Gulati, cost needs of their foreign customer. Mistakes could Nohria, and Zaheer, 2000; McEvily and Marcus, be tolerated if prompt moves were made to correct 2005; Uzzi and Lancaster, 2003; Lorenzen and their source. For customers, goodwill and demon- Mudambi, forthcoming; Blalock and Gertler, 2005; strable learning trumped technological backward- Lin, 2001; see McDermott and Corredoira, 2010, for ness or a lack of specific know-how. Competency an overview). and learning, however, could not be imposed. They had to develop through the practice of the relationship itself (cf. Dyer and Chu, 2011; Lui, forthcom- 2 This research has been funded by the Sloan Foundation and the Hans Boeckler Stiftung. The former funded two interna- ing; Ivarsson and Alvstam, 2010; Brown and tional research consortia—the Advanced Manufacturing Project Duguid, 1991). (AMP), 2000–04, and the Global Components Consortia, Although these practice-based organizational 2004–09; the latter funded the exclusively Germany focused Globale Komponenten project, 2010–12. AMP and Global learning theories are especially helpful in explaining Components involved the participation of Joel Rogers, Josh the first phase of Chinese manufacturing upgrading Whitford, Dan Luria, Susan Helper, Charles Sabel, Ed Stein- (described in step three), they are less helpful in feld, Jonathan Zeitlin, Peer Hull Kristensen, Gary Herrigel, and Volker Wittke. The latter two researchers and Ulrich Voskamp accounting for the upgrading dynamics we describe are the principle investigators in Globale Komponenten. in steps four and five. Community of practice

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x 112 G. Herrigel et al. arguments tend, on the whole, to focus on routine competence and discretion to adapt and develop suc- practices and the way in which apprenticeship rela- cessful products in China. Moreover, since the situ- tions (what Lave and Wenger, 1991, call legitimate ation is so dynamic, firms rapidly and systematically peripheral participation) unfold within stable sets of induce local reflection on the limits of central practices—such as the asymmetric ties between knowledge to facilitate competitive adaptation of Chinese exporters and their foreign customers (see home country products and practices to Chinese also Blackler and McDonald, 2000). conditions. We will show, however, that the gradual routinized We argue that MNCs accomplish this by deploy- apprenticed learning process ran up against limits in ing formal systems designed to combine joint China, in large part due to its own success. Chinese general goal setting with local discretion. Although producers developed competences that exceeded there are many variants, we will highlight the way in what customers were demanding. Awareness of this which CPSs accomplish this. As multidimensional mismatch between capability and demand in tradi- and recursive ‘constitutional processes’ (Ansell, tional export supply relations, we show later, ulti- 2010), CPSs create cycles of learning and recompo- mately generated a strategic shift on the part of both sition within MNCs and between MNCs and their indigenous producers and MNC manufacturers suppliers. The cycle starts when relevant stakeholder within the Chinese market. Both want to gain market teams (involving central and local players) establish share in emerging Chinese markets for dynamic provisional and revisable central metrics, standards. ‘middle range’ products. In order to achieve this, and product designs with the expectation that local however, firms need to disrupt their routines and players will use discretion in implementing and reflect upon both the formal and the tacit ways in adapting them to their circumstances. Local experi- which their practices are organized and their know- mental efforts interrogate and unpack home country how is deployed. Strategic uncertainty, unfamiliarity products and production practices to make them with new forms of production, and lack of experi- work in the Chinese context. Recursivity from the ence with product development and design, more- local back to the center is generated as home country over, leads producers to both tolerate and even teams monitoring and assisting local Chinese experi- encourage experimentation. mentation processes are induced through the moni- The pragmatism-based organizational learning lit- toring itself to reflect on the adequacy and optimality erature is centrally concerned with these dynamics. of their own routines both at home and in other It focuses on collective reflection (inquiry) and global locations. As Sabel (1994, 2005) points out, experimentation processes within organizations that these serially deconstructive and reflective actions occur when routine practices are disrupted (cf. make the tacit routines within communities of Elkjaer, 2003, 2004,; Elkjaer and Simpson, 2006; practice explicit to all the players and, thereby, Simpson, 2010; Sabel, 2005). Reflection and experi- systematically encourage stakeholders to alter, opti- mentation can be (and surely were) randomly pro- mize, and recompose their practices through voked during the first phase as relations between experimentation—all of which can stimulate creativ- Chinese and foreign customers experienced tempo- ity and generate new practice and product ideas rary crises or interruptions. This is how the commu- (Elkjaer, 2003, 2004; Elkjaer and Simpson, 2006; nity of practice literature understands how upgrading Simpson, 2010; see also Romme, 1996, 1997). We occurs through apprenticeship relations. Our claim, will later outline these recursive learning dynamics however, is that in the new situation, there has been and the various constitutional organization and gov- a concerted shift toward the systematic provocation ernance mechanisms they entail. of reflection and experimentation in order to gener- There are similarities in these practices to the ate recursive learning processes. The mechanisms dynamics described in traditional organization theo- for the inducement of reflection and experimentation ry’s understanding of ‘exploration’ (March, 1991) in learning have been a central preoccupation, in and to what strategy theorists try to capture with the particular of Sabel’s (1994, 2005) work. notion of ‘replication’ (Winter and Szulanski, 2001). In this second upgrading phase, reflection and The China dynamics differ from exploration and rep- experimentation-based learning processes grow out lication, however, in two ways. First, both of the of foreign MNC awareness of the limits of their latter approaches are equilibrium approaches where centralized, home country-developed knowledge, disruptive periods of exploration are replaced by new products, and practices. They must rely on local and stable periods of routinized exploitation. The

Chinese MNC learning processes, however, involve argue that the export platform strategy is inherently ongoing mixtures of exploration and exploitation. self-limiting. Although it produces remarkable Second, both March’s (1991) notion of exploration industrial growth and jobs, skeptics maintain that the and exploitation and Winter and Szulanski’s (2001) strategy fails to give Chinese producers access to the closely related understanding of replication are tools they need to become truly sovereign players in focused on asymmetric learning dynamics: firms the global economy because it focuses on a limited transfer models and seek to exploit gains from the area of manufacturing practice at the low end of new location, but they do not seek systematically to global value chains and the bottom of technological learn from those locations in ways that will recur- hierarchies. Specifically, none of the most techno- sively induce experimentation and change in the logically advanced and highest value functions sending location. The latter, however, is precisely (engineering and design capability, distribution what we will later show to be occurring in the new know-how, branding) are performed in China, and phase of mutual learning in China. the functions that are performed in China do not enable the Chinese to bootstrap themselves upward into those more sophisticated and valuable areas of Learning from foreign customers: the fruits value and technology (summary in Steinfeld, 2010; of export processing are not more Sturgeon and Lester, 2003; Nam, 2011). To the skep- export processing tical eye, China appears to have pursued a strategy China’s rapid industrialization in the last two that led it to become trapped at the bottom of the decades was driven by its distinctive integration into U-shaped technology development curve (Steinfeld, the global economy. Unlike the other Asian industri- 2010; Nolan, Zhang, and Liu, 2007).3 alizers (Japan, South Korea, and Taiwan) and fellow There is no denying that cases of this self-limiting BRICs (Brazil, India, and Russia), China relied dynamic exist and, moreover, that many Chinese heavily on FDI to industrialize (Steinfeld, 2010; producers follow exploitative low cost strategies Huang, 2003, 2008; Lemoine, 2000; Qian, 2010; with little hope for significant upgrading. Our claim, Ljungwall and Tingvall, 2010). By exploiting its however, is that by reducing the Chinese industrial- labor cost advantages, China lured FDI in a broad ization experience to those relations and outcomes, array of low-end, high volume manufacturing indus- the skeptics underestimate the dynamic learning tries. Foreign producers either constructed pure benefits associated with the Chinese strategy. Power manufacturing facilities in China or they helped and learning have a dynamic relationship, not an indigenous Chinese do so. The newly created facili- antithetical one. ties produced components and assembled final prod- In any case, initially the Chinese industrialization ucts, many of which were ultimately exported out of strategy involved no practical exposure to engineer- China for sale in other markets (mainly in the United ing and design know-how. Instead, it emphasized the States, Europe, and Japan). Especially after 1992, skill and know-how development needed to manu- this strategy gradually led to the emergence of highly facture other people’s products competitively and in competitive low cost manufacturers across a wide very high volumes. This was a learning strategy. array of industries: automobile and machinery com- With their customers or foreign employers as teach- ponents, product-level electronics, textile and ers, apprenticing Chinese producers became better apparel, footwear, and countless other commodity manufacturers, more and more able to achieve manufactured goods. This business was notoriously the quality standards their customers demanded low margin, often exceedingly exploitative of (Shafaeddin and Pizzarro, 2010). In particular, workers, and it involved very little knowledge of through their relations with foreign customers, how to develop or design products (Sturgeon and Chinese producers grew proficient in lean manufac- Lester, 2003; Lüthje et al., 2002; Dedrick, Kramer, turing principles and the new logic of formal CPSs.4 and Linden, 2010; Breznitz and Murphree, 2011). But it attracted manufacturing investment and, we 3 Optimists within this framework suggest that China can make claim, gradually integrated many Chinese produc- a virtue of its backwardness by trying to dominate mid-and tion facilities into transnational manufacturing com- low-level technological markets (Breznitz and Murphree, munities of practice. 2011). 4 For information on new style corporate production systems, Industrialization and development scholars have see Spear (2009) and Sabel (2005). For case studies of Chinese been skeptical of this strategy. In particular, many manufacturing upgrading, see Hawes and Chew (2011) and

This learning, in turn, impacted the development was concentrated outside of China, while focus in of Chinese human capital. Skilled worker demand China was on low cost.6 increased, especially in setup and maintenance, but Even in this kind of business, however, low labor even on the production lines themselves. Manufac- cost was not the only element of the story. The turing engineers were needed to implement more ability to achieve lean production flow, use automa- complex product designs in Chinese factories. Good tion to insure quality, and reduce change over time Chinese manufacturers came to be admired for their also played an important role. In this event, electron- ability to produce high volumes with good quality ics industry contract manufacturers educated and significant variety. By the beginning of the Chinese managers and workers in the principles of twenty-first century, rapid changeover and optimiza- modern lean manufacturing practice (Ernst, 2005, tion practices at high volume had become a Chinese 2007; Chesbrough, 2002; Lui, forthcoming; Dedrick comparative advantage. et al., 2010; Brown and Linden, 2011; Lüthje et al., This upgrading through apprenticeship within 2002; Sturgeon, 2002; Breznitz and Murphree, transnational communities of practice occurred 2011). through three distinctive kinds of relations between The other two avenues into transnational commu- foreign customers/investors and Chinese producers. nities of modern manufacturing practice were in The three relations varied by the intensity of integra- complex engineering sectors (automobiles, machin- tion in foreign communities of manufacturing prac- ery), where product architectures were more inte- tices. The stingiest and least integrated relation was grated and the clean separation between value-laden the contract-manufacturing model used extensively operations and cost-intensive operations was not as in many areas of the electronics industry.5 This clear as it was (alleged to be) in the electronics model presupposed a modular product architecture sectors.7 As a result, the intensity of learning induc- that enabled a clear separation between value-laden ing (apprentice-like) interaction between Chinese product design and development functions and cost- producers and foreign designer-customers was intensive manufacturing functions. Computer and greater. mobile device companies in the U.S. and Europe The second upgrading path grew out of offshore were the main drivers of this kind of arrangement. outsourcing relations. During the 1990s and early They designed notebook computers, mobile phones, 2000s, foreign automobile and machinery MNCs— PDAs, GPSs, iPads, iPhones, iPods, flat screen TVs, overwhelmingly from North America—targeted and countless other devices at home and had the indigenous Chinese producers as capacity suppliers hardware produced and (sub-) assembled in China and producers of low cost standardized or simple (e.g., Dedrick et al., 2010; Lüthje et al., 2002). components. The foreign players were typically Chinese manufacturing in this area was, to a large looking offshore for respite from the high manufac- extent, organized by foreign subcontractors, in turing costs (in particular high labor costs) in their particular by very sophisticated contract manufactur- home markets. In contrast to the contract manufac- ing multinationals (e.g., Flextronics, Selectron, turers in electronics, foreigners in these complex Foxconn, Jabil, etc.) with considerable know-how engineering sectors initially did not directly control regarding organizing production for exceedingly low production in China. Instead, companies such as margin products in a profitable way. Among such John Deere, Caterpillar, and CNH in the agricultural producers, product design and production know-how equipment and construction machinery sectors contracted with indigenous Chinese producers and

6 There were also less dominant examples of captive electronics Lockström et al. (2010). This learning seems especially pro- production in China; i.e., integrated operations by end produc- nounced in the sectors we have studied—automobiles and ers such as Siemens or Nokia. Chinese subsidiaries, in these machinery, see illustratively, Thun (2006). But there is evidence cases, functioned as in-house contract manufacturers, with of this kind of upgrading in other sectors as well, cf. Bellandi many of the same divisions between design and production, and Di Tommasso (2005) and, more broadly, Shafaeddin and though with slightly greater community of practice integration Pizzaro (2010). The skeptics have concentrated unduly on the dynamics. Such captive plants also often shared business from electronics sector, where the evidence for upgrading is more the parent with contract manufacturers. ambivalent. 7 For critiques of the ‘clean hand off’ view of modularity, see 5 The model could also be found in other sectors, such as tex- Chesbrough (2002); Ernst (2005); Fixon and Park (2007); tiles and clothing, where the stinginess toward the upgrading we MacDufﬁe (2007); Sabel and Zeitlin (2004); and Herrigel describe was perhaps even more entrenched. (2010).

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 115 brought them up to speed through their own supplier regarding component engineering when problems in quality assurance departments. Others, often smaller the flow of the supply chain emerged (Thun, 2006; companies active as suppliers to those large OEMs, Chu, 2011; Brandt and Thun, 2010; Shafaeddin and enlisted engineering and manufacturing quality con- Pizzarro, 2010; Lockström et al., 2010; Marsh, sultancies that sprang up in China to teach indig- 2008; Chang and Xu, 2008). enous manufacturers to produce to foreign customer Each of these avenues into a transnational com- quality standards. munity of manufacturing practice —contract manu- For their part, the Chinese producers frequently facturing, offshore outsourcing, joint ventures— attracted foreign business by hiring experienced generated more and less intensive positive learning Chinese managers, fluent in Western market and pro- and upgrading cycles. Interacting actor ambitions duction practices, away from foreign contract pro- expanded as their competences increased. Recogni- ducers. Many such managers also established their tion by the customer of the supplier’s ability to learn own businesses and used their proficiency with favorably disposed them to support the supplier’s foreign manufacturing standards as a selling point to ambitions to learn more. By the same token, an garner foreign business. Initial rounds of contracting inability to effectively collaborate and learn resulted and offshoring were plagued by the inconsistencies in exit from the customer’s stable of suppliers (Lui, and quality problems associated with novice work. forthcoming; Herrigel, 2010). Learning and success But with time and foreign investment in teaching were linked. and Chinese learning, the indigenous producers All these Chinese upgrading trajectories, enhanced their capabilities and improved their however, gradually bumped up against limits. Inter- production expertise and quality. A genuinely trans- estingly, these limits are different than those typi- national community of continuously improving cally pointed to by the skeptical development manufacturing practice emerged. literature. There the limit follows from the hierar- The third avenue for learning between foreign chical organization of global manufacturing know- MNCs and Chinese suppliers was in the early joint how, with high value design and engineering ventures in the automobile and machinery industries competence located outside China and production (including players in supply chains). Here Chinese know-how inside China. Hierarchy blocks Chinese governments enjoined foreign MNCs both to work producers from developing the crucial know-how with joint venture partners to create assembly opera- needed to develop their own products.8 tions on Chinese soil and to develop an indigenous Without denying that such hierarchies exist in supply base for their components (Chu, 2011; Thun, some product areas or that they are striven for in 2006). Once again, these large MNCs, both final many others, we claim that the Chinese upgrading assemblers like Volkswagen and component suppli- strategy actually encountered quite different ers like Bosch, directed purchasing and manufactur- limitations—and these alternative limitations proved ing engineers and supplier quality assurance experts not to be insuperable. The continuous apprenticeship to create an indigenous supply base for their prod- learning-based upgrading process reached limits ucts. The MNCs also worked closely with local gov- because while the Chinese developed the perpetual ernments to upgrade training infrastructures for ability to learn and participate in collaborative devel- themselves and suppliers. The products made in the opment projects, their foreign customers, for various initial automobile and machinery joint ventures were reasons, could not utilize or keep up with developing older models (e.g., the VW Santana) that incorpo- Chinese supplier sophistication. rated less competitive technologies and components Take the contract manufacturing and offshore out- than were used in the modern products the MNCs sourcing arrangements first. The basic limit in these were selling in richer developed markets. But the relations is that the more intensively collaboration learning that took place in the supply base was real became a part of the relationship between customer and increasingly sophisticated: Chinese suppliers and supplier, the more distance became a prohibitive had to develop the ability to machine to the more stringent MNC tolerances, to produce and deliver their goods in a more unforgiving MNC time frame, to apply themselves to continuous cost reduction, to 8 On China, see Breznitz and Murphree (2011). More generally, see also the skeptical argument for upgrading in supply chains develop the ability to cooperate with customers, to presented in Humphrey and Memedovic (2003) and Giuliani respond to special requests, and to make suggestions et al. (2005).

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x 116 G. Herrigel et al. cost. The iterated experimentation that is essential to ing capability was like a child who had outgrown collaborative product development was cumbersome the clothes its parent had provided. Fortunately, suc- to organize across long distances; apart from the cessful industrialization produced conditions that simple fact that transport costs and time were expen- facilitated new strategic and learning practices. It sive, there were also many unpredictable hazards— gradually became clear, for example, that increas- accidents, strikes at ports, weather on the oceans, ingly sophisticated Chinese manufacturers needed natural disasters—that could also generate cost. Cus- ever-more sophisticated capital goods and more tomers in developed markets, who originally encour- elaborate infrastructure. Successful industrialization aged Chinese efforts to become more sophisticated, produced higher income and education levels became more reluctant to welcome even highly com- throughout the most developed Chinese regions. By petent Chinese suppliers into the most sensitive the beginning of the new century, private consumers, and high valued collaborative development projects. industrial producers, and the Chinese state (through As a result, Chinese suppliers (and foreign-owned ambitious infrastructure projects) were ratcheting up subsidiaries in China) gradually found that they demand in the domestic market. At the same time, were participating in supply chains with producers foreign manufacturing MNCs grew eager to expand abroad in ways that failed to utilize all of their own and upgrade their existing Chinese operations to capabilities. capture part of this new growth potential. China There was a comparable limiting dynamic in strategy change became even more urgent for these manufacturing JV supply relations. There the initial MNCs as indigenous Chinese end product producers MNC strategy was to transfer relatively simple, often challenged them in the middle-value product ranges frankly outdated, products and technologies into in a wide array of sectors, such as automobiles, their Chinese JVs (cf. the VW Santana referenced omnibuses, power drives, turbines, construction earlier). The terms of the JV agreements, however, machinery, and sophisticated industrial machinery required MNC JV partners not only to have local and tooling (Brandt and Thun, 2010). content, but to use and develop local suppliers. As a Ultimately, the dam burst. Old self-limitation result, companies like VW encouraged its most dynamics dissolved into new competition and inno- trusted European suppliers to invest in China to vation dynamics in China. The current situation pits supply the JV and help develop the local supply base. ambitious and strikingly capable Chinese producers Together with these suppliers (and with the coopera- looking to move up and capture more advanced tech- tion of Chinese local governments), foreign MNCs nological business against foreign MNCs seeking to then cultivated indigenous Chinese suppliers able to adapt their technological know-how and skills to the reliably produce needed parts and components that local Chinese conditions. Both have advantages and either could not be supplied from Germany or the yet both are confronted with obstacles that require U.S., or that were required under the JV agreement cooperation with the other side. We will next (Chu, 2011; Thun, 2006; Brandt and Thun, 2010). describe these new dynamics in greater detail and To a point, this proved to be a successful strategy show how, ultimately, producers have begun to move for all the parties involved. Suppliers did develop and away from unidirectional learning from relations to they soon became eager to upgrade, expand their dynamic learning with relations. customer base, and produce more sophisticated products. For various reasons (many of them having to do with the political embeddedness of the customer JVs Competition between MNCs and indigenous in regional Chinese government procurement struc- Chinese producers for the ‘middle’ of the market tures), however, large JVs like VW Shanghai had difﬁculty altering their technologies and creating the New foreign MNC Chinese market engagement is possibility for further supplier upgrading (Sun, Thun, shaped not only by the quantity, but also by the and Mellahi, 2010). In this case, both indigenous changing character, of Chinese demand. Contempo- suppliers and FDI suppliers began to push against the rary Chinese customers demand products with dis- limits of the original FDI customer strategy for the tinctive characteristics. They want either product Chinese market. Producers wanted to be able to do versions that are notably less technologically sophis- more than the existing market structure permitted. ticated than the variants MNCs sell in richer more All these developments make it clear that by the developed markets; or they want high-end products beginning of the new century, Chinese manufactur- like those sold in developed markets, but with

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 117 specific modifications to accommodate Chinese level controllers in China. Knockoffs of older preferences, regulations, standards, and resource and designs already existed in the market at price points material input differences. In the former case, even the foreign producers could not match. Instead, when products are less sophisticated or involve they developed entirely new ‘simple but sophisti- materials or components not used in home country cated’ controllers, designed specifically for the versions, foreign MNCs can’t compete against needs of Chinese customers, that integrated new native Chinese producers by modifying and selling electronics into a simpler delivery unit. Indigenous older or outdated versions of their current equip- Chinese producers did not have the electronics ment. Native Chinese producers quickly copy such know-how to be able to compete with the new technologies and underprice the foreign producers in products. Crucially, in order to more quickly and the domestic market (Xi, Lei, and Guisheng, 2009; efficiently supply demand at a competitive cost Li, 2009; Ivarsson and Alvstam, 2010; Breznitz and level, the MNCs also shifted new controller pro- Murphree, 2011). duction to China, using indigenous Chinese design- In order to take advantage of the emerging Chinese ers (who cooperated with designers back in the customer sophistication, then, foreign MNC players home location), less complex production technolo- in the automobile, automobile component, and gies, and Chinese suppliers. complex machinery sectors significantly modify and Naturally, one could ask why the Japanese or reconceive their current forms of technology or even German companies would even bother to invest in develop entirely new products adapted to Chinese this kind of less than cutting-edge business. The idiosyncrasies. The aim is to reduce sophistication answer one gets in interviews with firm representa- and narrow product functionality. For German firms, tives is that the companies need to establish their in particular, this means learning how to design less brand positions in the Chinese market so that it will durable products that can be easily maintained and be possible to grow in the market. Foreign producers manufacture them with simpler production technolo- regard it as inevitable that the current midmarket gies and less skilled labor. These modifications niche will be overrun by rapidly improving indig- directly address the needs and usage norms of enous producers. But the more sophisticated the Chinese consumers and make the product more technology becomes, they believe, the slower the affordable than the versions sold in MNC home process of indigenous learning will become. If markets. The sweet spot for foreign MNCs is a the MNCs can establish their brand position in this midrange market above the highest volume, lowest emerging market, they’re betting that the market will quality commodity segment and yet below the grow toward their strengths. highest quality, most sophisticated technologies typi- Rather than designing a wholly new machine or cally sold in developed markets. Chinese customer component for the Chinese market, many firms rede- desires for such products exceed native Chinese pro- sign existing offerings to make them more affordable ducers’ abilities to produce them, so there is a legiti- and appropriate for Chinese needs. Peter Marsh of the mate market opportunity for foreign firms. Financial Times has been following this phenomenon There are countless examples of this sort of mid- and he quotes a manager at Mindray, a producer of market product. Take the case of computer numerical medical devices and patient monitoring systems: control technology (CNC) for machine tools and other steerable capital goods. Due to insufficient ‘We look at what parts we can standardise, where we demand, Japanese and German CNC producers can reduce the level of technical sophistication cannot widely sell their highly complex highest-end without comprising quality, and in what instances we controllers to machine tool and other capital equip- can substitute software for electronic components,’ ment producers in China. Instead, the foreign firms says Joyce Hsu, Mindray’s chief financial officer. succeed by selling specially designed ‘simple’ CNC The result, she says, is often a low cost product that may not have so many features as an equivalent piece controllers. As a manager at one German firm told of equipment made in Western Europe or the U.S. but us, ‘There are millions of conventional machine which satisfies requirements in hospitals—in China tools in China that can be easily converted to simple and elsewhere—that are trying to cut back on costs’ CNC machines if they are provided with the right (Marsh, 2008: 4). kind of controller.’ Interestingly, neither the Japanese nor German A German automobile supplier who has been part producers could sell older versions of their entry- of our Globale Komponenten study has followed

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x 118 G. Herrigel et al. a similar strategy for the design and delivery of customers could work with suppliers to improve internal frames (front ends, engine cradles, cross car their production quality. beams) in China. Even when the firm wins a bid on The VW experience in China illustrates one way a global component that will be built into the same in which this interconnected upgrading process takes automobile model in Europe, North America, and place.9 As we noted earlier, for decades VW pro- China, materials, component designs, and manufac- duced the modest and outdated Santana only in turing procedures still differ in each market. In China, the company does not use the same quality China, in organizationally and technologically mini- steel or the same sophisticated welding processes malist production locations. But toward the end of that it uses to construct the customer’s front ends in the twentieth century, when the company recognized Europe. The China product is manufactured in a that a market for its luxury brand Audi was emerging comparatively primitive fashion and is less complex, with the growth of a wealthy class in China and that less durable, and incapable of the same level of per- demand growth was too robust to service with formance as its European counterpart. German exports, the company erected production We found producers in many different machinery and assembly facilities in China that mirrored those and automobile component sectors undertaking in Germany. Their aim was to produce a Chinese these sorts of quite substantial product and produc- Audi A6 that was identical in quality to the A6 tion modifications (reconceptualizations), including manufactured in Germany. firms making power drives, turbines, gear units, While working to achieve this goal, Audi recog- transmissions, and woodworking machinery. Cru- nized that Chinese customers actually wanted an A6 cially, modifications like these are most conveniently with particular characteristics unwanted by German done in China, even when there is substantial coop- (and other European and North American) consum- eration with home country designers. Chinese engi- ers. For example, Chinese consumers wanted neers understand the customer requirements, local limousine-like sedans with significantly more regulations and standards, and the quirks and quality legroom (30 centimeters) than existed in German of local materials. And local production managers versions. Incorporating such design changes within have an active sense of labor capabilities and the an integral architecture like the A6’s entailed corre- realistic costs of running complex Western produc- sponding changes in materials quality, component tion machinery in the Chinese context. machining, in-house system assembly, and the char- Such moves on the part of MNC firms, of course, acter of local supplier relations. Audi engineers, create opportunities for local Chinese suppliers to planners, and purchasing teams could not manage all become integrated into the newly recast production of these changes from Germany. Therefore, local strategies. MNCs need local suppliers because the Audi engineering, production worker capability, and volumes for castings or stamped metal frames for a supplier quality assurance had to be developed and controller (or the front end of an automobile or the maintained to work together with the home country cab of a construction machine made in China) actors to manage the changes. simply overwhelm the operations of suppliers back New and old orientations to the Chinese market in the home market. It is not simply that such distant can often be seen together, like geological sedimen- suppliers are already busy with production slated for tations, in German MNC factories in China. At our home region customers, but also that transportation visit to the VW/Audi/FAW JV in Changchun, for costs make home country components too difficult example, alongside the old equipment still in use for to justify. This encourages component-producing the VW Jetta, were brand new assembly lines with MNCs to invest in China to offer their customers a flexible automation and materials handling work sta- ‘global footprint.’ But it also provides an opportunity tions that had been specifically designed for Audi’s for indigenous Chinese producers to enter into newly local Chinese assembly needs. The equipment was emerging and more sophisticated foreign MNC highly automated, but distinctive in its ability to supply chains. In any case, as indicated earlier, many handle radical amounts of variety: different versions capable Chinese suppliers have emerged who are of the A6, Golf, Jetta, and Bora were assembled in chafing at the limits of their old export processing routines and are in a position to take advantage of such business (Chu, 2011; Thun, 2006; Xi et al., 9 We saw analogous movements toward this kind of localization 2009; Li, 2009). In cases where capable suppliers upgrading in visits to auto suppliers, woodworking machinery, and drive train and braking systems producers in the Shanghai were difficult to identify immediately, MNC and Tianjin regions.

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 119 the facility. Although German engineers codesigned competence transfer and indigenous competence the equipment and supervised its implementation, development. We observe that much of this process is local Chinese engineers, maintenance, and setup not at all random or an expression of some sort of specialists were deployed to implement, maintain, ‘natural’ development path. Rather MNCs and the and operate this quite sophisticated equipment. indigenous Chinese producers they interact with are Audi’s traditional German suppliers—e.g., Bosch, systematically inducing and optimizing learning ZF, Hella—were also forced to adjust to the design through the procedures they deploy to combine changes introduced into the A6 (and other VW (global) products, standards, and metrics with disci- models) in China. Since the changes were China plined local discretion. Continuous adaptability and specific, it made sense for these suppliers to imple- innovation, driven by experimentation and learning is ment those changes locally, in their Chinese opera- essential for competitiveness in the Chinese market. tions, as well. These changes then forced all of those Moreover, MNCs view the learning and innovation producers to alter their external sourcing strategies to taking place in their Chinese operations as a source of incorporate more local Chinese suppliers. This global advantage that can benefit operations else- process, in turn, led Audi suppliers to implement the where. As a result, the formerly one-way learning kind of modifications in material usage and manu- relations characteristic of the communities of manufacturing engineering and technology described facturing practice described earlier are giving way to earlier in the case of our German front end supplier. recursive, multidirectional, mutual learning relations Characteristically, once the expanded Audi China based in joint reflection and experimentation. production complex proved successful, it made The paradigmatic mechanism used to generate further innovation possible. VW used its Chinese these new mutual learning dynamics is the formal competence to transform the Jetta into the Bora, a CPS that manufacturing firms increasingly deploy simple yet sophisticated hybrid model that mimics throughout their global operations. Examples include characteristics of the Jetta (Golf platform) and the the ACE system at United Technologies, Formel ZF company’s simpler Polo model. at ZF, the Siemens Production System, the VW production system (known as ‘The Volkswagen Way’) and the Caterpillar Production System, among count- The organization of mutual learning: less others (cf. Sabel, 2005; Spear, 2009; Fujimoto, systematically disrupting routine and inducing 1999). Some, mostly smaller MNCs like the wood- reflection, experimentation, and creativity working machinery producer in our Globale Kompo- within and among firms nenten case studies, stop short of branding their We have shown that the foreign MNCs and indig- corporate systems, but nonetheless self-consciously enous manufacturers are altering their strategies and deploy extensive formal procedures that mimic many commitments in the Chinese market. Rather than aspects of the CPSs in larger MNCs. simply purchasing component inputs from Chinese Many of the companies we interviewed not only suppliers or running sleepy low-tech operations infused all of their operating practices with the producing anachronistic technologies, MNCs are formal procedures of a CPS, but also maintained developing serious Chinese production and design elaborate continuous improvement teams (CITs) that operations to compete in a dynamic and rapidly were charged with the responsibility of spreading the changing market. Indigenous producers, for their CPS gospel throughout the global organization. CITs part, are eagerly casting off their apprentice rela- teach employees lean production, team collaborations and engaging foreign customers in more col- tion, and realistic target setting, while providing con- laborative, design-intensive, and high value-added sulting services and reengineering input to teams, business. departments, and production cells to help them We argue that these changes have given rise to implement new forms of organization and practice. a new multidirectional learning dynamic both In the German automobile components and complex within MNCs and between MNCs and their Chinese machinery sectors, CITs were among the most glo- suppliers. What is remarkable about the stories bally active players within the firm (e.g., Soder, we’ve related is that the upgrading processes have a 2006). snowball quality: the transfer of capability fosters CPSs establish group-based goal setting and indigenous competence development that, in turn, monitoring procedures that systematically induce creates additional possibilities that require still more collective self-observation, problem diagnosis, and

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x 120 G. Herrigel et al. problem solving experimentation among all players ated by this process is recursive, in the sense that throughout a firm’s value chain. In such systems, the output from one application of a procedure or internationally composed team negotiations typi- sequence of operations becomes the input for the cally establish common MNC-wide product designs, next, so that iteration of the same process produces quality standards, cost targets, and even manufactur- changing results (Sabel and Zeitlin, 2012). Local ing procedures. Crucial in these arrangements is that deviations from central designs and practice made by the targets or standards (or, in cases of simultaneous local teams must be justified to their central coun- engineering, the designs) are sufficiently general to terparts. When the changes are accepted they are allow for considerable local discretion in implemen- then themselves formalized and turned into standard tation. Actors in specific markets are encouraged to local practice. The new local ‘standards’ are, in turn, experiment with adapting the standards, targets, and benchmarked by higher order teams within the orga- procedures to local conditions. Local players, nization against similar practices in other areas. however, are not given carte blanche to deviate from Where appropriate or possible, adjustments are common targets. Rather, they are required to justify made elsewhere, which then results in the creation of their decisions to the central teams and provide new higher order standards and targets. The dialec- elaborate quantitative and organizational evidence tical logic of: for the local superiority of their modifications. The possibility of discretion gives local players incentive jointly negotiated central standards→ local discre- to experiment and be innovative, while the require- tion with public justification (peer review)→ recur- ment of justification (and continuous monitoring and sive adjustment of central standards dialogue with skilled and interested teams in other locations) wraps processes of local experimentation creates a continuous process of experimentation and with discipline (for a theoretical description of this optimization within the firm that globalizes learning process, see Sabel, 2005). across the entire MNC. These formal systems of joint goal setting, local Interestingly, although these systems rely heavily discretion, and mutual monitoring aim to generate on formal procedure (metrics, standards, writing positive learning spirals of local adaptation and things down), technology plays only a secondary global improvement, product optimization, and role in the arrangement of the governing relations innovation. Product norms, standards, and metrics (facilitating data collection and monitoring, for for local targets are explicitly, continuously (re)con- example). Indeed, in practice, such systems can be structed by relevant stakeholders in production— surprisingly low tech. For example, at our Globale manufacturing teams, design teams, customer- Komponenten project’s woodworking machinery focused teams, purchasing teams, often with the producer, the firm-managed machinery production input of CIT players. Teams are constituted at the transfer to its Chinese (and Central European) pro- local and central MNC levels and are interdepen- duction locations with what it called a ‘cookbook’ dent: the success of the local players relies on central system. German production teams took pictures of actor input while, at the same time, the success of the each discrete step in the home plant machining pro- central actors depends on local player success. At cesses to be transferred abroad—including machine regular, formally proscribed intervals (and more fre- layout (tools and fixtures), individual setups, tool quently on an informal basis in between), central and positioning, transfer procedures, and work organi- local actor deliberation generates continuous, mutual zation. The pictures were then annotated with self-analysis among production stakeholders (see instructions for setup procedures, machining Romme, 1996, 1997, for an interesting theoretical speeds, tact times, and expected output for each rendering of this dynamic). stage in the production and assembly flow. These In effect, team actors in different locations jointly ‘cookbooks’ were then sent to the company’s reflecting on their mutual activities makes practices Chinese operations, along with the blueprint in each location transparent to all players—the tacit designs for the machines, to guide the construction features of local actions are made explicit. Local of Chinese manufacturing operations for all wood- discretion combined with joint scrutiny induces dis- working machine models. ruption of routine and encourages plant teams to The cookbook functioned as a set of very specific experiment with designs, materials, and production guidelines, but local players were empowered organization. The multidirectional learning gener- to use discretion while implementing them. The

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 121 implementation process allowed for any number of housings. In this case, the Chinese supplier did both local design and procedural modifications in manu- casting and machining for the German firm (the facture: the reduction or enhancement of product latter, in part, on machinery that had been transferred functionality, the excision or addition of steps in the from the German company’s home plant in production process, and the substitution of materials Germany). In both cases, the customer specified were all fair game as the local players sought to broad production quality and cost targets. Formal adapt the home firm’s designs to the Chinese regu- audits and regular joint procedure reviews with the lations, consumer interests, and competitive cost customer took place: engineers from the German conditions. But local modifications had to be justi- company’s Chinese plant in Suzhou regularly visited fied to the exporting multifunctional teams and man- the Chinese supplier, as did skilled workers from the agers that originated the cookbook. Acceptable gear producer’s German plants. The aim was to help changes were noted, new pictures were taken, and a their Chinese counterparts set up the new machinery local cookbook was constructed. If the changes and understand how to solve problems generated in implemented locally improved the way the machines aluminum housing machining. were produced, the local innovation was embraced These interactions yielded a number of jointly by German teams and the new procedures were pho- agreed upon deviations from original customer tographed and integrated into the home cookbook. designs and practices, in particular regarding mate- Because the cookbook is a detailed deconstruction rials used in molds, casting techniques (more skilled of a machine into literally hundreds of discreet laborer input, less automation), and even the main- manufacturing operations, it was possible for home tenance procedures of the German company’s country players to adopt small alterations for their former machinery. In order to optimize the changes, own use at home, even as the overall character of the the Chinese supplier upgraded its manufacturing specific machinery model diverged from the one that engineering and design capabilities (an expense the company was manufacturing back in Germany. which it was encouraged to incur, not only from its In this way, the formal process of systematic stake- German customer, but also from the regional holder monitoring of the cookbook created a recur- Chinese government, which gave tax breaks to firms sive multidirectional process of organizational increasing design capacity). The resulting supplier learning that combined disciplined local innovation improvements helped the German company main- and product adaptation with openness to global tain its quality and cost targets, while simultaneously design and manufacturing process innovation. stimulating ideas to adapt the overall gear unit Crucially, these formal translocational learning designs to facilitate entry into a new user market— practices are also deployed, in modified form, to Chinese omnibus manufacturers. Perhaps most sig- govern relations between customers and suppliers. nificantly, the company took machining process Instead of a formal cookbook of pictures and modification ideas introduced in China and used instructions for how to implement and adapt propri- them to experiment with the setup of similar pro- etary machinery and operations, MNC customers cesses in its Russian and Indian subsidiaries. This provide Chinese suppliers with clear targets for successful mutual learning, moreover, created the cost, quality, and delivery time for a part or com- possibility for additional more challenging design ponent that the customer and supplier design intensive collaborations between the German MNC together. Both parties observe the progress of the and the Chinese supplier in the next contract round. supplier relative to the target and, in cases where targets are missed, both immediately seek to identify the reason and work toward a resolution. When Discussion: is mutual learning really beneficial to MNCs? successful, such formally self-analyzing relations produce learning and continuous upgrading, for There are obvious benefits to Chinese manufacturing both the customer and the supplier, not only across suppliers (and, in reverse, to Chinese OEMs working Chinese supply chains, but across entire global pro- with foreign MNC component suppliers or capital duction networks (Ivarsson and Alvstam, 2010; Lui, good producers) from mutual learning activity. It forthcoming). allows Chinese producers to break through the We observed such relations in a variety of our self-limiting structures of the older unidirectional firms. Take the relations between a German gear export oriented community of practice. The benefits unit MNC and its Chinese supplier of aluminum to foreign MNCs, however, may appear to be more

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x 122 G. Herrigel et al. ambiguous. On the one hand, the new relations customers. Global MNC production arrangements produce mutual learning and, consequently, MNCs and relations with suppliers are, in this sense, stra- gain the ability to improve not only local Chinese tegic formally governed systems focused not on operations, but also similar operations in other loca- making specific products but on constructing tions. On the other hand, if their Chinese collabora- collaborative continuous learning processes that tors are becoming such successful and sophisticated drive competence expansion, innovation, and learners that they provoke MNC learning processes self-transformation. as well, isn’t there a danger that the Chinese will The Japanese and German CNC controller manu- abandon the collaboration with the foreigner and facturers we discussed earlier are exemplary of the manufacture the product on their own? That is, aren’t larger trend. Their long-term market strategy is not to MNCs worried about losing intellectual property produce simple but sophisticated technologies (such and, hence, market presence to their increasingly as CNC units), but to be recognized by consumers as sophisticated Chinese collaborators? reliable and quality manufacturers in their industry Our interviews uniformly reveal that foreign (e.g., computer automation equipment). Such firms manufacturing MNCs in China engage in these believe that if they can create dynamic learning rela- recursive collaboration, learning, and upgrading tionships in China, they will be able to leverage the relations despite the fact that intellectual property global technological know-how they have in auto- cannot be guaranteed in the relationship, especially mation equipment (or gear units or woodworking not in any long-term sense. Moreover, foreign auto- technologies) to make those Chinese relations con- mobile and complex machinery managers broadly tinuously and reliably competitive in the dynamic acknowledged that their indigenous Chinese and expanding Chinese market. The idea is to estab- employees (managers and skilled workers) were lish dynamic learning capabilities in China that can, gaining knowledge of proprietary products, tech- in turn, participate in emergent global learning and nologies, and procedures and at least some of them innovation operations. This strategy promises more were taking what they were learning and deploying return in the long term than one that seeks to protect that knowledge for their own ends. market share on any particular product model or The same was true of suppliers. Successful col- generation in a specific national market.10 laborators were looking to establish ties with other potential customers that explicitly leveraged what they learned in their relationship with the MNC. In CONCLUSION most interviews, foreign managers viewed these dynamics as inevitable and, after a certain point, This article describes contemporary Chinese manu- nonpreventable. All players acknowledge, moreover, facturing upgrading as a multidirectional, interac- that this is true despite what most observe to be tive, recursive, and learning-driven process. Far from continuous improvement in the capacity of Chinese a technological cul de sac, the experience of export authorities to protect property rights. processing and participation in transnational supply Why do foreign MNCs nonetheless proceed with chains in a broad array of cases helped Chinese the creation of these types of mutual learning rela- producers learn international manufacturing best tions? They do it because the emerging Chinese practices. It apprenticed them in cooperative, competitive and strategic conditions are becoming learning-driven supply chain relations and it similar to those in the MNC home regions. Competi- schooled them in the global disciplines of lean tive pressure for continuous innovation, cost reduc- production and supply chain collaboration. Such tion, and change drive most relations in the global learning prepared the ground for them to be able to market place. Regardless of property right quality, in produce their own designs and participate substan- relatively mature, integrated product architecture- tively and creatively in CPSs. Moreover, this manu- based manufacturing sectors, there are few guaran- facturing upgrading process created increasingly tees that a given product or technology will maintain an advantageous position in any market for very 10 This orientation may be specific to the sectors we studied and long. Rather than orienting their strategies around less characteristic of the views of players in industries, such as protecting technology and product designs, then, the electronics and software industries, where first-mover monopolies account for an enormous share of the value produc- firms continuously improve and transform designs to ers are able to capture. We note this possibility and leave its match changing customer needs and identify new exploration to a future article.

Copyright © 2013 Strategic Management Society Global Strat. J., 3: 109–125 (2013) DOI: 10.1111/j.2042-5805.2012.01046.x The Process of Chinese Manufacturing Upgrading 123 sophisticated demand for manufactured products ACKNOWLEDGEMENTS within China itself, and this new demand, in turn, gave rise to new forms of competitive and coopera- Many thanks to Gerry McDermott and three anony- tive market dynamics among both indigenous mous reviewers, each of whom made helpful sugges- Chinese and foreign MNCs. tions on previous drafts. Thanks also to Jonathan We argue that this new market situation is associ- Zeitlin, Gerry Berk, Dennis Galvan, Andreas ated with a shift to a new, more dynamic multidirec- Glaeser, Andy Abbott, Dan Slater, Dieter Ernst, Ed tional and recursive form of mutual learning among Steinfeld, Eric Thun, Marc Blecher, Dali Yang, and Chinese producers and foreign MNCs. Chinese pro- Maja Lotz for comments. ducers, engineers, and skilled workers—suppliers, customers, and local personnel in upgrading MNC Chinese locations—are helping foreign MNCs REFERENCES understand how to adapt their products to the speci- Ansell C. 2010. Pragmatist Democracy: Evolutionary ficities of the Chinese market. Foreign MNCs are Learning as Public Philosophy. Oxford University Press: radically altering their commitments to the Chinese New York. market by developing more capacious engineering, Bellandi M, Di Tommasso M. 2005. The case of specialized design, and organizational capabilities and practices towns in Guangdong, China. European Planning Studies in their subsidiary operations. Key mechanisms for 13(5): 707–729. this transformation are the recursive, team-based, Blackler F, McDonald S. 2000. Power, mastery, and orga- mutual learning processes that are generated by the nizational learning. Journal of Management Studies formal procedures in CPSs. Such systems create 37(6): 833–852. global learning spirals by imposing systematic inter- Blalock G, Gertler PJ. 2005. Foreign direct investment and action between local discretion and global standards. externalities: the case for public intervention. In Does Though highly competitive (and, hence, like any Foreign Direct Investment Promote Development?, Moran TH, Graham EM, Blomstrom M (eds). Institute market process capable of producing losers), our for International Economics: Washington, D.C.; 73–106. story suggests that the larger interactive dynamic Brandt L, Thun E. 2010. The fight for the middle: upgrad- between Chinese players and MNCs has many sur- ing, competition, and industrial development in China. prisingly mutually beneficial dimensions. World Development 38(11): 1555–1574. Finally, we can’t help pointing out that our story Breznitz D, Murphree M. 2011. Run of the Red Queen: is, at least from one point of view, ironic. Initially, Government, Innovation, Globalization, and Economic the worry with developed country engagement with Growth in China. Yale University Press: New Haven, CT. China was that offshoring and outsourcing relations Bronfenbrenner K, Luce S. 2004. The Changing Nature of would involve the loss of competence to China Corporate Global Restructuring: The Impact of Produc- and/or a shift of home country competence away tion Shifts on Jobs in the U.S., China, and Around the from manufacturing (e.g., Bronfenbrenner and Luce, Globe. U.S.-China Economic and Security Review Com- mission: Washington, D.C. 2004). Our research shows, however, that these Brown C, Linden G. 2011. Chips and Change. The MIT worries are misplaced (or perhaps overtaken by Press: Cambridge, MA. events) in the automobile and complex machinery Brown JS, Duguid P. 1991. Organizational learning and sectors. Current Chinese engagements systemati- communities-of-practice: toward a unified view of cally recompose global MNC internal flows and working, learning, and innovation. Organization Science: Chinese offshore experiences through recursive 2(1): 40–57. mutual learning. As a result, China operations supply Chang SJ, Xu D. 2008. Spillovers and competition among MNC home country product development and pro- foreign and local firms in China. Strategic Management duction processes with useful benefits. The benefits Journal 29(5): 495–518. are likely to be all the greater, moreover, as the gap Chesbrough H. 2002. Towards a dynamics of modularity: a between Chinese and developed country market cyclical model of technical advance. In The Business of Systems Integration, Prencipe A, Hobday M, Davies A sophistication narrows—in part, as a result of the 11 (eds). Oxford University Press: Oxford, U.K.; 174–200. mutual learning processes this article outlines. Chu WW. 2011. How the Chinese government promoted a global automobile industry. Industrial and Corporate Change 20(5): 1235–1276. 11 We develop this argument more extensively in Herrigel, Dedrick J, Kramer K, Linden G. 2010. Who profits from Wittke, and Voskamp (forthcoming). innovation in global value chains? A study of the iPod and

notebook PCs. Industrial and Corporate Change 19(1): Humphrey J, Memedovic O. 2003. The global automotive 81–116. industry value chain: what prospects for upgrading by Dyer JH, Chu W. 2011. The determinants of trust in developing countries. Working paper, UNIDO Sectorial supplier-automaker relationships in the U.S., Japan, and Studies Series, Vienna, Austria. Korea. Journal of International Business Studies 42(1): Ivarsson I, Alvstam C. 2010. Upgrading in global value 10–27. chains: a case study of technology learning among IKEA Elkjaer B. 2003. Organizational learning with a pragmatic suppliers in China and Southeast Asia. Journal of Eco- slant. International Journal of Lifelong Education 22(5): nomic Geography 11(4): 731–752. 481–494. Kumaraswamy A, Mudambi R, Saranga H, Tripathy A. Elkjaer B. 2004. Organizational learning: the ‘third way.’ 2012. Catch-up strategies in the Indian auto components Management Learning 35(4): 419–434. industry: domestic firms’ responses to market liberaliza- Elkjaer B, Simpson B. 2006. Towards a pragmatic theory of tion. Journal of International Business Studies 43(4): creative practice. Paper presented at the Second Organi- 368–395. zation Studies Summer Workshop on Re-turn to Practice: Lave J, Wenger E. 1991. Situated Learning: Legitimate Understanding Organization as it Happens, 15–16 June, Peripheral Participation. Cambridge University Press: Mykonos, Greece. New York. Ernst D. 2005. Limits to modularity: reflections on recent Lemoine F. 2000. FDI and the opening up of China’s developments in chip design. Industry and Innovation economy. Working paper, Centre d’Études Prospectives et 12(3): 303–335. d’Informations Internationales, París, France. Ernst D. 2007. Beyond the ‘global factory’ model: innova- Li Z. 2009. The role of international technology transfer in tive capabilities for upgrading China’s IT industry. Inter- the Chinese automotive industry. Discussion paper 269, national Journal of Technology and Globalisation 3(4): Manufacturing Management Research Center, University 437–459. of Tokyo, Tokyo, Japan. Fixon S, Park J-K. 2007. The power of integrality: linkages Lin N. 2001. Building a network theory of social capital. In between product architecture, innovation, and industry Social Capital: Theory and Research, Lin N, Cook K, structure. Working paper 4642–07, Sloan School of Man- Burt RS (eds). Aldine de Gruyter: New York; 3–30. agement, Massachusetts Institute of Technology, Cam- Ljungwall C, Tingvall P. 2010. Is China different? A meta- bridge, MA. analysis of the effects of foreign direct investment on Fujimoto T. 1999. The Evolution of a Manufacturing System domestic firms. Journal of Chinese Economic and Busi- at Toyota. Oxford University Press: Oxford, U.K. ness Studies 8(4): 353–371. Gereffi G, Humphrey J, Sturgeon T. 2005. The governance Lockström M, Schadel J, Harrison N, Moser R, Malhotra of global value chains. Review of International Political MK. 2010. Antecedents to supplier integration in the Economy 12(1): 78–104. automotive industry: a multiple-case study of foreign Giuliani E, Pietrobelli C, Rabellotti R. 2005. Upgrading in subsidiaries in China. Journal of Operations Manage- global value chains: lessons from Latin American clus- ment 28(3): 240–256. ters. World Development 33(4): 549–573. Lorenzen M, Mudambi R. Clusters, connectivity, and catch- Gulati R, Nohria N, Zaheer A. 2000. Strategic networks. up: Bollywood and Bangalore in the global economy. Strategic Management Journal 21(3): 203–215. Journal of Economic Geography. Forthcoming. Hawes C, Chew E. 2011. The cultural transformation of Lui C. Collaboration on thin ground: contract production large Chinese enterprises into internationally competitive arrangements between Taiwanese firms and their Ameri- corporations: case studies of Haier and Huawei. Journal can MNC customers in the personal computer industry. of Chinese Economic and Business Studies 9(1): 67–83. PhD dissertation, Department of Sociology, University of Herrigel G. 2010. Manufacturing Possibilities: Creative Wisconsin, Madison, WI. Forthcoming. Action and Industrial Recomposition in the U.S., Lüthje B, Schumm W, Sproll M. 2002. Contract Manufac- Germany, and Japan. Oxford University Press: Oxford, turing: Transnationale Produktion und Industriearbeit in U.K. der IT-Branche. Campus Verlag: Frankfurt am Main, Herrigel G, Wittke V, Voskamp U. Governance architec- Germany. tures for learning and self-recomposition in Chinese MacDuffie JP. 2007. Modularity and the geography of inno- industrial upgrading. In Political Creativity: The Mangle vation. Paper presented at the Sloan Industry Studies of Institutional Order, Agency, and Change, Berk G, Conference, 25–27 April, Cambridge, MA. Galvan D, Hattam V (eds). University of Pennsylvania March JG. 1991. Exploration and exploitation in organiza- Press: Philadelphia, PA. Forthcoming. tional learning. Organization Science 2(1): 71–87. Huang Y. 2003. Selling China. Harvard University Press: Marsh P. 2004. Bosch plans big drive into China. Financial Cambridge, MA. Times, 23 December: 1–1. Huang Y. 2008. Capitalism with Chinese Characteristics. Marsh P. 2005a. Stihl to open plant in China. Financial Cambridge University Press: New York. Times, 28 February: 1–1.

Marsh P. 2005b. Companies move to computerise plants in Sociology, Smelser NJ, Swedberg R (eds). Princeton Uni- China. Financial Times, 1 June: 1–1. versity Press: Princeton, NJ; 137–165. Marsh P. 2005c. Caterpillar weighs China options. Financial Sabel C. 2005. Real time revolution in routines. In The Times, 4 October: 1–1. Corporation as a Collaborative Community, Heckscher Marsh P. 2007. Agco chief in drive to gain second place. C., Adler P. (eds). Oxford University Press: New York; Financial Times, 21 March: 1–2. 105–151. Marsh P. 2008. China’s new formula: manufacturers begin Sabel C, Zeitlin J. 2004. Neither modularity nor relational to move beyond low cost. Financial Times, 28 May: 1–6. contracting. Enterprise & Society 5(3): 388–403. Marsh P. 2009. Emerging economies: still fertile ground. Sabel C., Zeitlin J. 2012. Experimentalist governance. In Financial Times, 24 March: 1–3. The Oxford Handbook of Governance, Levi-Faur D. (ed). Marsh P. 2010a. Steel chief hits out at U.S.-China plant Oxford University Press: Oxford, U.K. Forthcoming. critics. Financial Times, 27 September: 1–1. Shafaeddin M, Pizzarro J. 2010. The evolution of value Marsh P. 2010b. Eaton chief hits out at Beijing’s critics. added in assembly operations: the case of China and Financial Times, 27 December: 1–1. Mexico. Journal of Chinese Economic and Business Marsh P. 2011a. Emerging economies flex manufacturing Studies 8(4): 373–397. muscle. Financial Times,14 March: 1–4. Simpson B. 2010. Pragmatism, Mead, and the practice turn. Marsh P. 2011b. Focus on deals high up value chain. Finan- Organization Studies 30(12): 1329–1347. cial Times, 25 April: 1–2. Soder J. 2006. Ziel: prozessberater-pool in allen feldern McDermott GA, Corredoira RA. 2010. Network composi- aufbauen. Produktion interview mit Johann Soder, SEW- tion, collaborative ties, and upgrading in emerging- Eurodrive. Produktion 7: 27. market firms: lessons from the Argentine autoparts sector. Spear SJ. 2009. Chasing the Rabbit. McGraw-Hill: New Journal of International Business Studies 41(2): 308– York. 329. Steinfeld E. 2010. Playing Our Game. Oxford University McEvily B, Marcus A. 2005. Embedded ties and the acqui- Press: New York. sition of competitive capabilities. Strategic Management Sturgeon T. 2002. Modular production networks: a new Journal 26(11): 1033–1055. American model of industrial organization. Industrial Moran TH, Graham EM, Blomström M. 2005. Does FDI and Corporate Change 11(3): 451–496. Promote Development? New Measurements, Outcomes, Sturgeon T, Florida R. 2005. Globalization, deverticaliza- and Policy Approaches. Institute for International Eco- tion, and employment in the motor vehicle industry. In nomics: Washington, D.C. Locating Global Advantage: Industry Dynamics in the Mudambi R. 2008. Location, control, and innovation in International Economy, Kenney M, Florida R (eds). Stan- knowledge intensive industries. Journal of Economic ford University Press: Palo Alto, CA; 52–81. Geography 8(5): 699–725. Sturgeon T, Lester R. 2003. The new global supply base: Nam K-M. 2011. Learning through the international new challenges for local suppliers in East Asia. Working joint venture: lessons from the experience of China’s paper MIT-IPC-03–006, Industrial Performance Center, automotive sector. Industrial and Corporate Change Massachusetts Institute of Technology, Cambridge, MA. 20(3): 855–907. Sun P, Thun E, Mellahi K. 2010. The dynamic value of Nolan PF, Zhang J, Liu C. 2007. The global business revo- MNE political embeddedness: the case of the Chinese lution, the cascade effect, and the challenge for firms. automobile industry. Journal of International Business Cambridge Journal of Economics 32(1): 29–47. Studies 41(7): 1161–1182. Qian Y. 2010. How reform worked in China. In In Search of Thun E. 2006. Changing Lanes in China: Foreign Direct Prosperity: Analytic Narratives on Economic Growth, Investment, Local Governments, and Auto Sector Devel- Rodrik D (ed). Princeton University Press: Princeton, NJ; opment. Cambridge University Press: New York. 297–333. Uzzi B, Lancaster R. 2003. Relational embeddedness and Romme AGL. 1996. Making organizational learning work: learning: the case of bank loan managers and their clients. consent and double linking between circles. European Management Science 49(4): 383–399. Management Journal 14(1): 69–75. Winter SG, Szulanski G. 2001. Replication as strategy. Romme AGL. 1997. Organizational learning, circularity, Organization Science 12(6): 730–743. and double linking. Management Learning 28(2): 149– Xi L, Lei L, Guisheng W. 2009. Evolution of the 160. Chinese automobile industry from a sectoral system of Sabel C. 1994. Learning by monitoring: the institutions of innovation perspective. Industry & Innovation 16(4): economic development. In The Handbook of Economic 463–478.