Wang, Jici (2006): Industrial clusters in China: the low road vs. the high road in cluster development, Chapter 7 in Allen Scott (ed.): Development on the Ground: Clusters, networks and Regions in Emerging Economies, Security Sales and Integration （Forthcoming）
INDUSTRIAL CLUSTERS IN CHINA: THE LOW ROAD VERSUS THE HIGH ROAD IN CLUSTER DEVELOPMENT
Jici WANG
1. Dilemmas in China’s Industrial Clusters
In recent years, there has emerged a voluminous research literature, taken from many different fields, that addresses the importance of industrial agglomeration and clusters. These concepts have also received considerable attention in China, and this paper provides an account of the cluster concept and its uses by Chinese scholars and policymakers.
The term “industrial cluster” refers to a group of firms and associated institutions that are both geographically proximate and functionally related. The idea has two key elements. First, firms in the group must be linked by traded interdependencies (input-output relationships) or by untraded (social and cultural) relationships that form around a core activity, if not by both. Second, these interrelated firms must be located in close geographical proximity to one another. The group generates external economies, and may create opportunities for mutual learning. In successful cases, the group also enhances levels of cooperation, trust, and innovation. The competitiveness of the cluster can in many cases be strengthened by appropriate public intervention.
A big difference between clusters in developed and developing countries appears to revolve around the kinds of market niches they focus on. Innovative and dynamic clusters in developed countries tend to specialize in higher-value niches, while clusters in developing countries tend to serve the lower end of the market where competitiveness is determined by price. In the latter case, entrepreneurs seldom share information or discuss common problems that they meet. Levels of trust in these clusters are not very high, leading to cut-throat competition. Nevertheless, industrial agglomeration means that firms are usually able to find skilled and experienced labor, secure up-to-date and relevant market information, make contact with diverse suppliers, and interact with the local institutional environment. All these factors reduce production costs, thus enhancing the competitiveness of individual firms. Industrial upgrading is another important consideration of developing country clusters. For clusters in developing countries, upgrading within global value chains is important. Only by the careful adoption of coherent policies and strategies can a cluster upgrade, avoiding the “low road” in order to take the “high road” to development.
1.1 Sources of ambiguity relating to the cluster concept in China
Growth via agglomeration is a concept long familiar to economists, rooted in the writings of Alfred Marshall (1890, 1919) and subsequent heterodox economists (Hirschman, 1958; Myrdal, 1959; Perroux, 1961). Since the early 1980s, an enormous surge of scholarship in economic geography and allied fields has greatly expanded on this earlier work, emphasizing that selected regions are capable of exerting powerful push effects on national economic development. In recent years, no one has been more influential in promoting the concept of specialized industrial localization than Michael Porter (1990, 1998, 2001), whose notion of industrial clusters has rapidly become a central analytical concept and policy tool.
In a theoretical context, the term “industrial cluster” differs from the agglomeration model and industrial complex concept. There is a belief that such clusters reflect not only economic responses to the pattern of available opportunities and complementarities, but also an unusual level of embeddedness and social integration (Gordon, 2003).
In practice, although embedding a dynamic cluster within an urban center is a complex process, attempts at cluster development have generally taken the form of conventional urban planning approaches, such as localized infrastructure projects, preferential policies and financial assistance. They usually focus on urbanization via the construction of high-cost infrastructure to attract outsiders, rather than emphasizing localization to create learning effects.
a) Localization and urbanization economies
The first major source of ambiguity confusion over the concept of clusters is the confusion of localization and urbanization economies. According to Hoover (1937, 1948), the benefits of agglomeration can be grouped together into three categories: internal returns to scale, localization economies and urbanization economies (Gordon, 2000). Internal returns to scale result from larger groupings of firms and industries being concentrated in one location. Urbanization economies result from concentrations of population and infrastructure, a common labor pool, and shared quality of life. These externalities accrue to all local firms, irrespective of sector. Localization effects, in contrast, stem from the agglomeration of specific activities which favor specialized infrastructure and human capital, including vocational training, and political lobbies (Goodall; Healey & Ilbery). This type of agglomeration economy is confined to a group of local firms in a given sector (Gordon, 2000).
〈邓：我在你的bibliography找不到Goodall与Healey/Ilbery的文章。他们在吗？〉
Scott and Storper (2003) state three basic factor underlying agglomeration in their analysis of agglomeration processes. Their statements obviously draw a clear connection between urbanization and localization economies:
“Cities always appear as privileged sites for economic growth because they economize on capital-intensive infrastructure (which is particularly scarce in developing areas), thus permitting significant economies of scale to be reaped at selected locations. But to this obvious basic factor underlying agglomeration, we must add three further sets of phenomena that complement and intensify its effects, namely (a) the dynamics of backward and forward inter-linkage of firms in industrial systems, (b) the formation of dense local labor markets around multiple workplaces, and (c) the emergence of localized relational assets promoting learning and innovation effects. Some brief commentary on these points is now in order.”
Storper and Venables (2002) use the term “buzz” to identify an important sub-set of urbanization economies, which refers to the information and communication benefits that result from face-to-face contacts, and the co-location of people and firms within the same industry and place or region (Bathelt, Malmberg and Maskell, 2002). In developing countries, however, the concept of a sector-specific cluster is largely confused with a pan-sector agglomeration of firms – a concept related more to Hoover’s theories about urbanization..
b) Dimensions of proximity: Geographical versus organizational
The confusion between the notion of geographical proximity and a true industrial cluster can also be found in the literature. This is the second source of ambiguity in cluster concept. For example, Crevoisier (1999) observed differing views on proximity between Camagni and Scott. For Camagni (1995), proximity and agglomeration do not necessarily coincide. Scott (1988), however, takes proximity as a necessary feature of industrial clustering in his analysis of agglomeration in California. Crevoisier postulates that this might be a consequence of differences between European and North American geographies. In North America, he argues, few regions other than major metropolitan areas,feature the same population density and activity characteristics that are found in corresponding areas in Europe (e.g. in Germany, North Italy, Switzerland, and the Benelux countries).
The impact of geographical proximity on interactive learning and innovation is a key issue in economic geography. In view of the French School of Proximity Dynamics, distinction is made between organizational and geographical proximity. While geographical proximity is defined as spatial distance between actors, organizational proximity is associated with the organizational relationships between economic actors. Economic geographers sometimes add a third form of proximity, institutional proximity, to account for the fact that interactions between players are influenced, shaped and constrained by the institutional environment (Torre and Gilly, 1990; Boschma, 2005).
Georgaphy, however, may not be enough to produce positive externalities. Boschma argues that “geographical proximity cannot be considered a sufficient condition for the exchange of tacit knowledge. This may be illustrated by the experience of multinational corporations when they try to get access to the knowledge base of a host region through the setting up of a local plant. They regularly fail to do so, because it turns out to be hard to become a member of tight networks of personal relationships through which local knowledge circulates. Another illustration is the role of gatekeepers that bring external information into their home region, but this new information diffuses only to those local agents that form part of that local network” (Boschma, 2005).
Boschma claimed that geographical proximity per se is neither a necessary nor a sufficient condition for shared learning. Proximity may facilitate interactive learning by strengthening the other dimensions of proximity, but can also have negative impacts on innovation due to the problem of “lock-in” Accordingly, too much proximity may be as detrimental to interactive learning and innovation as too little. Boschma describes five dimensions of proximity: cognitive, organizational, social, institutional and geographical (Boschma, 2005). Agglomeration that facilitates the “learning economy” will possess all five dimensions of proximity.
For China, however, it is still unclear whether geographical proximity alone is sufficient for the needs of regional development. This issue in related to the ambiguity issued from urbanization in industrial agglomeration.
c) Global-local discourse on industrial linkages
Thirdly, ambiguity about cluster concept arises from the global-local discourse on industrial linkages. Yeung et al. (2005) found that benefits of agglomeration derive from ”non-cluster economies of traded interdependencies among firms in clusters,” referring to external economies of collective bargaining and production capacity derived from trading relationships between institutions, firms and customers outside local clusters. As it stands in developing countries, comparative advantage in the endowment of skilled labor can be magnified if specific industries are organized in clusters. Even low-cost skilled labor can benefit from clustering, as it is highly attractive for foreign direct investment in high-tech industries. As Hendry et al. (2000) found that there existed ‘proximity without intimacy or interaction’, national and international linkages are to be more significant than local relationships.
Much of the research on the innovation process, to date, has been concerned with local networks. However, recent literature has pointed to the importance of extra-local interconnections in the context of globalization. More connections should be made between these types of networks, as embodied in the production networks of global firms, and regional development in specific territorial formations. Local clustering and global production networks are not only compatible, but can also be mutually reinforcing. Latecomer firms in developing countries can become both vertically integrated in global value chains and horizontally integrated in local clusters.
This questions are key for China, because many local governments in China are eager to build local linkages, but have ignored the benefits of external linkages.
d) Site of operation versus site of interaction
The evolution of cluster theory has grown from a basic focus on the site of operation to a broader view of site of interaction, but these two are still used ambiguously. Initial work on the issue of agglomeration held that links between firms, institutions and other economic agents, located in geographical proximity, tend to generate advantages of scale and scope – development of general labor markets and specialized skills – as well as enhanced linkages between suppliers and customers (Lloyd and Dicken, 1990). These economic efficiencies were concentrated at the site of operation, and did not involve local capability for learning and innovation. Over time, however, researchers concluded that agglomeration may also facilitate the “learning economy” through interaction among actors in a cluster. These new views claim that industrialization is a territorial process, while innovation is a social process (Asheim and Cooke, 1999). In this model, the relationship between innovation and regional economic development is given a far more prominent place. From this dialogue emerged a number of theoretical perspectives, such as the theory of the “new industrial spaces,” the “district” theory, the innovative milieu approach, and “regional innovation systems.” In developing countries, however, the cluster concept continued to refer, by and large, to the basic concept of agglomeration, and focus on its operational efficiencies.
In some traditional “operational clusters,” the spirit of the Schumpeterian entrepreneur might dwindle, due to increasing industrial concentration and the domination of large companies. In response, these clusters often also lobby for sectoral interventions often at a national or supranational level, which hamper the restructuring process more than they support it, as they remove the incentives for entrepreneurs to take the initiative, thus paralyzing competition and tranquillizing large industries. In these kinds of networks, status is privileged over knowledge, power over learning, and the past over the present (Eich-Born and Hassink, 2003).
e) Concentration and dispersion
A fifth source of ambiguity regarding to the cluster concept is the confusion about agglomeration, which is often tied to knowledge creation, and dispersion, which is often viewed as a way to lower labor costs. Globalization and the worldwide increase in contract manufacturing have created new trends towards concentration and dispersion in the global economy. Despite the scattering of global industry, industrial agglomeration is still a major force driving industrial geography, because of external transaction costs such as transport and administrative duties, which increase in the event of greater vertical disintegration. Recent trends towards flexible specialization and associated small-scale, volatile and highly specific contract-manufacturing have also increased the incentives to agglomerate is accentuated (Leung, 1993). There is a strong presumption that high-wage and more knowledge-intensive activities are more likely to cluster, and hence more resistant to geographic dispersion. In reality, however, there is no clear-cut separation between the trend that low-end activities are highly dispersed and the trend that knowledge-intensive ones require localized clusters; the reality is considerably messier. Globalization enhances the dispersion of knowledge across firm boundaries and national borders, and can push industries to seek lower labor costs around the world. Such dispersion however has remained concentrated, due to the continuous impact of agglomeration economies. The latter are no longer restricted to the nation state. These dispersed clusters at difference locations can be integrated through global production network. In such a context, industrial clusters may only be considered as facilitating factor for a number of subsequent developments (which may or may not occur): division and specialization of labor, the construction of a wide network of suppliers, the emergence of agents who sell to distant national and international markets, provision of specialized services, the materialization of a pool of specialized and skilled labor, and the formation of business association (Giuliani, Rabellotti and van Dijk, 2005).