TSpace Research Repository tspace.library.utoronto.ca

Conceptualizing Multiple Clusters in Mega-City Regions: The Case of the Biomedical

Industry in Beijing

Harald Bathelt & Jingyuan Zhao

Version Post-print/accepted manuscript

Citation (published version)

Bathelt, H., & Zhao, J. (2016). Conceptualizing multiple clusters in mega-city regions: the case of the biomedical industry in Beijing. Geoforum, 75, 186-198.

Copyright / License © 2011. This manuscript version is made available under the CC-BY-NC-ND 4.0 license. http://creativecommons.org/licenses/by-nc-nd/4.0/

Publisher’s Statement The version of record [Bathelt, H., & Zhao, J. (2016). Conceptualizing multiple clusters in mega-city regions: the case of the biomedical industry in Beijing. Geoforum, 75, 186-198.] is available online at: http://www.sciencedirect.com/science/article/pii/S0016718516300653[doi: 10.1016/j.geoforum.2016.07.016]

How to cite TSpace items Always cite the published version, so the author(s) will receive recognition through services that track citation counts, e.g. Scopus. If you need to cite the page number of the TSpace version (original manuscript or accepted manuscript) because you cannot access the published version, then cite the TSpace version in addition to the published version using the permanent URI (handle) found on the record page.

1

May 31, 2016 Word count: 9,935 (main only)

Conceptualizing multiple clusters in mega-

city regions: The case of the biomedical

industry in Beijing

by

Harald Bathelt

University of Toronto, Department of Political Science and Department of Geography and

Planning, Sidney Smith Hall, 100 St. George Street, Toronto ON M5S 3G3, Canada;

E-mail: harald.bathelt@utoronto.ca, URL: http://www.harald-bathelt.com

and

Jingyuan Zhao

University of Toronto, Department of Political Science,

Sidney Smith Hall, 100 St. George Street, Toronto ON M5S 3G3, Canada;

E-mail: jingyuan.zhao@utoronto.ca

To be re-submitted to

Geoforum

2

Conceptualizing multiple clusters in mega-

city regions: The case of the biomedical

industry in Beijing Abstract (ca. 150 words): This paper introduces a unique industrial configuration

that has emerged in Beijing, where three economic clusters in the biomedical industry, originally established as industrial/research parks, have developed parallel to each other. This configuration of multiple co-located clusters of the same industry, which has not been discussed before, raises the question of whether the industrial/research parks are competing for the same resources, or whether they are complementary to each other and can collectively be viewed as a new type of industrial configuration. The paper conceptualizes a framework of multiple clusters in mega-city regions that distinguishes between collaborating and competing clusters and presents initial empirical evidence for the Beijing case. As such, this research aims to unravel the phenomenon of multiple clusters in mega-city regions and to understand the complex spatial interrelationships that exist within and beyond multiple co-located clusters in the same industry.

Keywords: Beijing, biomedical clusters, cross-cluster linkages, mega-city regions, multiple clusters

JEL codes: D22, D85, L14, R11

1. Introduction

Debate about industry clusters has become widespread over the last few decades. A

cluster is a localized economic context in which many firms from a value chain

simultaneously compete against each other and also collaborate to gain economic advantages.

Since Porter’s (1990) study on ‘the competitive advantage of nations’, cluster strategies have

become popular approaches to fostering economic growth among policymakers and

economic development practitioners worldwide (Lagendijk and Cornford, 2000).

Academic interest in clusters has proliferated. Studies since the 2000s have

investigated the potential benefits of clusters in terms of their contribution to metropolitan

competitiveness, the key role of clusters in the generation and effective transmission of

innovations, and the networks of internal interactions and external linkages by which clusters

3

generate synergies and introduce new knowledge (e.g. Bathelt et al., 2004). And yet, only a

few studies have explicitly investigated interrelationships and networks – whether local,

national or international – between industrial clusters (e.g. Hsu and Saxenian, 2000; Chen,

2004; Blundel and Thatcher, 2005; Zhou, 2008; Lu and Cao, 2012; Conlé and Taube, 2012;

Lu et al., 2013; Bathelt and Li, 2014).

There is, in particular, very little work about multiple clusters within a single city.

This is somewhat surprising, not least because there is clear evidence that metropolitan areas

and mega-city regions have become the bases of multiple industrial clusters. The co-location

of two or more clusters in a metropolitan region is, in fact, quite common: it occurs in both

world city regions, such as New York, London and Toronto, and much smaller urban areas,

such as Wuppertal or Nürnberg in Germany. Montreal’s metropolitan region alone, for

instance, is home to seven organized cluster initiatives (Montreal Clusters, 2014). While not

all of these may qualify as ‘true clusters’, similar co-agglomerations do exist and support

each other through urbanization economies, especially related to labor market effects

(Crevoisier, 2001).

There have been no studies at all of two or more clusters that are situated within in a

single city region and are also within the same industry. And yet, this situation may not be as

unusual in the world’s largest cities as one may think. As urbanization processes in

developing countries continue to accelerate, such cases could become more prominent in

countries like China and other developing economies with strong states (Wong 2004), where

cluster development often occurs through the planned establishment of multiple industrial or

research parks. A configuration of multiple clusters in the same industry that are located in

the same city-region and yet socially and spatially separated requires a very large urban

context. We would not expect this kind of development in small urban environments because

4

the separate infrastructures, research networks, labor markets and input-output linkages

required for such a configuration would be too costly to establish and maintain.

We understand a cluster, in what follows, as an agglomeration of firms in an industry

along with their suppliers and service providers that are linked through traded and untraded

linkages and create their own labor markets, research networks and on so on (Bathelt and

Taylor 2002). To be able to identify multiple clusters in a city region, it is not enough to

identify spatially distinct agglomerations of firms in different parts of the city; they must also

be socially separated. This does not mean that there are no linkages between these clusters,

but that each has its own labor market dynamic and linkage network and could potentially

exist without the others.

In this paper we hope to initiate a debate about multiple co-located clusters in the

same industry by investigating a specific industrial configuration that has evolved in Beijing,

where three economic clusters in the biomedical industry have developed, originally

established as industrial/research parks: Zhongguancun Life Science Park (ZLS Park),

Beijing Economic & Technology Developing Area (Yizhuang Park) and Beijing

Bioengineering & Pharmaceutical Industry Base (Daxing Park). These three planned

developments in Beijing can be viewed as industrial clusters because each is characterized by

an agglomeration of firms in the same industry, a substantial infrastructure of suppliers and

service firms and an institutional environment that includes universities and specialized

research facilities which support their reproduction.

Since there is no research about multiple co-located clusters in the same industry,

numerous questions about the nature and origin of such development have yet to be answered.

Such questions form the basis of our analysis in this paper. In particular, we address the

following research questions: do the three industrial/research parks have a similar structure

and are they set up as rivals that compete for the same resources? Or do they perform

5

complementary tasks and can be viewed as a new type of industrial configuration? We

initiate a theoretical debate about multiple co-located clusters within the same industry by

comparing two types of cross-cluster relationship which are differentiated according to the

nature of linkages within and between the respective clusters, namely collaborating and

competing co-located clusters.

In the empirical part of the paper, we present qualitative evidence from our ongoing

research in the Beijing biomedical industry. In particular, information will be drawn from

telephone interviews with biomedical firms that aimed to identify research, input-output,

labor market and knowledge linkages within, between and beyond the three clusters under

consideration. This study also draws from prior empirical research conducted over the past

decade (+Zhao, 2006; 2008; 2010; 2012). By focusing on the linkage structures between

firms, we aim to understand three clusters that exist in one mega-city region and the complex

interrelationships between them. Our primary goal in this part of the paper is to identify

whether the type multiple-cluster configuration in the Beijing biomedical industry is closer to

the ideal type of competing or collaborating clusters by looking at the cluster structures that

have emerged and the linkages that have developed between them.

Our analysis will be structured as follows. Section 2 introduces the theoretical and

empirical background of cross-cluster relationships and multiple-cluster structures, which

then serves as the basis for the distinction between collaborating and competing co-located

clusters, discussed in section 3. In section 4, the methodological basis of the study is

discussed, while section 5 describes the context of biomedical clusters in Beijing. Section 6

presents the empirical evidence that allows us to identify three collaborating clusters in

Beijing. Finally, section 7 concludes our argument and spells out some of its implications.

6

2. Theoretical and empirical background

This section discusses the general phenomenon of multiple clusters and inter-cluster

relationships. While this topic has not been analyzed systematically in conceptual or

empirical terms, numerous studies provide hints as to how to proceed.

2.1 Cross-cluster linkages

While most studies of clusters have only focused on linkages within clusters, a limited

number of studies have investigated the linkages and networks between clusters that are

located in adjacent regions or in different cities and different countries. Some empirical work

has shown, for instance, that similar industrial clusters are sometimes located in relative

geographical proximity in neighboring regions. One example of such a spatial structure can

be found in the ‘Third Italy’: Veneto, Emilia-Romagna and Tuscany are close-by

administrative regions with a similar industry focus and structure which have strong textile

clusters that compete in the global market and share similar business cultures and social

networks (Asheim, 2000). Similarly, Delgado et al. (2010) found in their longitudinal study

of clusters in the US that agglomeration tendencies in adjoining regions increase the

likelihood that these regions develop similar or related clusters.

Most studies on the topic of cross-cluster linkages have focused on connections

between agglomerations in different regions or countries, often over large distances.

Examples have been found of both competition and collaboration between clusters in

different regions and countries. For example, Blundel and Thatcher (2005) describe how the

yacht manufacturing cluster in Southern England gradually declined over time as its market

was eroded by other yacht manufacturing clusters in France, Sweden and Germany. By

contrast, it has been found that the information technology clusters located in the Beijing,

Yangtze River Delta and Pearl River Delta regions are connected with each other through

firms engaged in inter-cluster collaboration (Zhou, 2008). Close network linkages and a

7

specific spatial and social division of labor also exist between information technology

clusters in Taiwan and Mainland China. Clusters in Taiwan, such as Hsinchu, focus on

research and development (R&D), while clusters in Mainland China, such as Shenzhen,

receive crucial technologies from Taiwan and focus on the manufacturing and assembly of

final products (Hsu and Saxenian, 2000; Chen, 2004; Lin et al., 2011; Wang and Lin, 2013;

Lu et al., 2013). Conlé and Taube (2012) found in their study of the emergence of clusters in

China that different types of biotechnology clusters are linked with each other and exchange

knowledge and technologies. They present evidence of numerous examples of linkages

between China’s most important domestic biomedical firms, located in different clusters

across the country.

Some initial evidence also reveals how leading clusters establish networks with

similar clusters in other countries based on foreign-direct investment linkages. Recent studies

have, for instance, identified linkages between the ceramic tile clusters in Emilia, Italy and

Castellon, Spain (Oliver et al., 2008) and between the film industry clusters in Hollywood

and Vancouver (Scott and Pope, 2007). In a recent study of FDI linkages between Canada

and China, Bathelt and Li (2014) developed a conception of global cluster networks and

global city region networks by understanding multinational firms from a network perspective,

which explains the existence of cross-cluster knowledge networks between those two

countries. Similarly, Turkina et al. (2016) present a study of the network of linkages within

and across 52 aerospace clusters in order to provide new insights into the structure and

dynamics of such networks and of the way they evolve from a geographically partitioned to a

hierarchical cross-cluster network structure, stratified along value chain stages.

Interestingly, almost all of these studies focus on cross-cluster linkages among

different regions or different countries. Although not investigated extensively and

systematically, such linkages may, as Phelps (2004) puts it, be expressions of a trend toward

8

new extended forms of agglomeration that are characterized by linkages that go beyond the

cluster itself.

2.2 Multiple clusters in metropolitan regions

The phenomenon of industrial clusters located within a single region or in adjacent regions is

only just beginning to receive some attention in the literature, even though spatial

configurations with more than one industrial cluster are quite common. Chakravorty et al.

(2003) analyze eight industrial sectors with evidence of co-located clustering and suggest that

land use policies are a key factor in influencing the intra-metropolitan spatial distribution of

industry clusters. Lu et al. (2013) suggest that the co-presence of clusters in different

industries within a large city region can trigger innovation and economic growth through

spillovers associated with the exchange of complementary knowledge between firms from

different clusters (Lu and Cao, 2012). An economy with a large degree of diversity naturally

has the ability to push knowledge exchanges further towards the development of new

research fields (Beaudry and Schiffauerova, 2009); and the more diversified a regional

economy is, the more likely it can create inter-industry knowledge bases and product

combinations (Neffke et al., 2011). This points to important diversity advantages of co-

located clusters in metropolitan regions. The proponents of the French-Swiss milieu school,

who originally studied innovation processes focusing on small and medium-sized firms

outside of urban areas (e.g. Camagni, 1991; Keeble and Wilkinson, 1999), found that the

effects of innovative milieus are more pronounced in large urban regions, within which broad

sophisticated labor markets and different overlapping milieus can develop that cross-fertilize

each other and generate important urbanization economies (Crevoisier, 2001). Since regional

social networks structure the interactions between agents and provide a framework for

knowledge exchange and cooperation between firms, knowledge may spread more easily in a

region that is home to multiple complementary clusters. In this case, transaction costs in

9

regional exchange relations may be lower and the efficiency of knowledge transfers higher

than in a region without multiple clusters (Fornahl and Tran, 2010).

These and other examples in the literature are primarily concerned with the co-

location of clusters from different industries in metropolitan regions. However mega-city

regions may also be characterized by a rather different spatial configuration: namely, the

existence of two or more similar clusters in the same industry. As far as we know, no

academic study has yet focused on investigating configurations of clusters that are part of the

same industry and are co-located in the same mega-city region. Firms in multiple clusters of

this kind can benefit from relative geographical proximity within the same institutional

context, compared with firms that are located in faraway cities or regions. Since opportunities

for localized learning derived from spatial proximity are often based on joint or similar

cognitive, institutional, social and cultural settings, multiple clusters in mega-city regions

may be close enough to develop effective cross-cluster linkages, while the social distance

between them may be large enough to support cluster-specific networks and labor market

dynamics.

One could argue that in knowledge-based industries, such as biomedicine, firms must

generate linkages to the most recent scientific and technical knowledge in order to

successfully develop new products. Such knowledge is not only complex, but also often tacit,

and thus difficult and expensive to transfer (Sorenson et al., 2006). Intensive face-to-face

interaction between firms that have the same specialized research focus may thus become a

necessity (Owen-Smith and Powell, 2004; Stuart et al., 2007; Fornahl et al., 2011). From this

perspective, one might assume that multiple clusters in mega-city regions would need to

engage with one another in order to develop close-by linkage networks, especially with

leading biomedical research organizations, such as universities and existing firms (Haug,

1995; Zucker and Darby, 1996). Crevoisier and Jeannerat (2009) go beyond simple

10

arguments of proximity relations and emphasize in their analysis of changing innovation

contexts that milieus increasingly need to mobilize external knowledge and anchor

themselves in multi-location networks and environments. This line of analysis suggests the

possibility of conceiving the phenomenon of multiple clusters in mega-city regions as a

spatially-compressed form of such multi-location dynamics.

However, the situation of multiple clusters in mega-city regions is not so straight-

forward; it requires re-conceptualization. We have to keep in mind that strong competitive

relationships may develop between these clusters if they require access to similar sets of

infrastructure and other localized resources, such as specialized research facilities.

Furthermore, while studies often emphasize the role of automatic knowledge flows and

networking in clusters, there is good reason to believe that, in some industry contexts,

information and trust are not necessarily easily shared among cluster members (Lissoni and

Pagani, 2003). Whether firms in multiple co-located clusters in the same industry in one

mega-city region would easily share information and trust is thus not clear. In other words, it

is not self-evident whether primarily collaborative and complementary or competitive and

exclusive linkages are likely to develop between the multiple clusters in such a setting. This

suggests that we need to take a closer look at the advantages and problems of such spatial

configurations – a task that we turn to in the next section.

3. Conceptualizing multiple clusters in the same industry

in mega-city regions

Mega-city regions with multiple clusters may allow firms to benefit from synergies

due to relative geographical proximity and potentially strong social, institutional and cultural

association (Torre and Rallet, 1999; 2005; Boschma 2005; Moodysson et al., 2009; Mattes,

11

2012). Yet at the same time, the clusters may be distant enough from each other to develop

labor markets, research networks and knowledge ecologies of their own.

While it is too early to state how widespread the phenomenon of co-located same-

industry clusters is, it seems likely that similar structures may be identifiable in mega-city

regions other than Beijing. For example, a group of industrial clusters within the same

industry can be found in South Korea’s Seoul metropolitan area. Seoul is known for its

information and communication technology (ICT) industry and highly-skilled labor market.

About 43 percent of the country’s ICT patents originate from Seoul, where the existing

Teheran Valley (Gangnam), Poi Valley Cluster (Yongsan and Yeoido) and the Seoul Digital

Industrial Complex (Guro) have become critical locations for the country’s ICT success.

These locations have developed multiple clusters, with Gangnam, Seocho and Yeongdeungpo

specializing in ICT services, while Yongsan specializes in wholesale/retail, and Guro and

Geumcheon in manufacturing activities (Hamaguchi and Kameyama, 2007).

3.1 Advantages and challenges

In both cases, Beijing and Seoul, strong state capacity (Wade, 1990) clearly played an

important role in the establishment of the clusters, but it is also possible that such

configurations can result as a side product of decentralized economic development in a fast-

growing metropolitan context where a city quickly spreads into its surrounding area. And

neither in South Korea nor in China is economic development completely controlled by the

state (Tan, 2006; Wang and Lin, 2013). Whatever the specific origin of the emergence of

multiple co-located clusters in the same industry, it is a configuration that has not been

systematically discussed before and can be associated with both advantages and challenges.

12

(i) Advantages

From the perspective of cluster linkages, multiple clusters in mega-city regions can

provide numerous benefits to firms, including advantages associated with distributed

knowledge assets, opportunities for collaboration, government support and land use.

Utilization of distributed knowledge assets. Different and spatially separated districts

in a mega-city region have specific advantages resulting from their proximity to specific local

assets and resources, such as university programs and research laboratories, labor markets,

buyer-supplier networks, transportation facilities, and so on. For this reason, multiple clusters

are better equipped than single clusters to provide access to the diversified resources and

knowledge assets dispersed across a mega-city region. Firms can make location decisions so

as to provide the most direct access to the required resources. Multiple clusters can thus

utilize the entire set of advantages within a mega-city region and provide firms with more

technological opportunities, if they are part of complementary networks. Decentralized

knowledge resources may, of course, also stimulate rivalry, product differentiation and

dynamic innovation processes between the firms.

Collaboration between clusters. If firms locate closer to specific knowledge resources

and technology parks, different districts in a city develop that are characterized by different

capabilities. As firms within a district become focused on similar products, limited interaction

may develop between them as they become competitors. Along with competition and rivalry,

firms may also try to avoid unintended knowledge transfers and imitation of products. With

different clusters of this type, however, complementary specializations can develop that

generate opportunities for collaboration between these entities. Especially when thinking

about potential research linkages, it may be attractive to generate connections across clusters

between firms that utilize somewhat different, but related, resources.

Specialized government support. Under these conditions, multiple industrial clusters

located within a mega-city region can develop different foci. For instance in the case of

13

Beijing, we find that one cluster specializes on technological innovation with a strong R&D

base, a second on pilot production and business incubation, and the third on standardized

manufacturing. These specialization effects are more pronounced if they are supported by

policy support or even initiated by such initiatives. In any case, once these clusters reach a

certain size, a scale effect will attract further specialized firms as well as targeted government

funds that fit the specific needs of the firms and research organizations in each district. In

establishing collaborations with neighboring clusters or in setting up specialized facilities in

several clusters, firms can benefit simultaneously from different government programs.

Land use and spatial expansion. In emerging market economies, high economic

growth in metropolitan regions attracts people and industries and leads to centrifugal urban

expansion which brings with it competing demands for infrastructure, residential and

industrial land uses. In such high-growth contexts, it is almost impossible to reserve a large

territory as a potential base for the future growth of one specific industry. Under conditions of

uncertainty, a configuration with two or three smaller territories for an industry within a

mega-city region is a more appropriate way of coping with both the industry’s demand for

growth and the need to make efficient use of land and urban infrastructure under strong

competitive and developmental pressures. It is not surprising under these conditions that we

find multiple clusters of a growing industry, such as the biomedical industry, in different

districts of the Beijing mega-city region.

(ii) Challenges

Multiple co-located same-industry clusters can, of course, also face challenges that

could have a negative impact on urban development due to, for example, competitive

relations between the respective districts or technology parks (Zhao and Zhang, 2008).

Competition between clusters. Multiple clusters in mega-city regions can provide

synergies and spark corporate rivalry with positive impact on innovation; however, if

14

multiple clusters develop the same specialization, strong competition between the districts

may be the outcome. In this case, firms from different clusters will strive for access to the

same resources, such as specialized research facilities, collaborators and highly-skilled labor

force and for the same government funds. Firm-level rivalry could quickly lead to cluster-

level competition, with attempts to block off local resources from other districts and the

potential development of separated clusters that are not interlinked.

Lower spillovers effects. Knowledge spillovers associated with a dense information

and communication ecology are often viewed as stimulating innovation in clusters (Saxenian,

1994; Bathelt et al., 2004; Sorenson et al., 2006). If such knowledge spillovers are highly

sensitive to distance, then innovation processes may be restricted or face additional hurdles

when faced with a situation of separate clusters in a single mega-city region. This may be

especially the case in highly dynamic research contexts, such as the biomedical industry,

where spillovers are dependent on fast and precise knowledge transfers and face-to-face

contacts.

Waste from duplication. Under conditions of limited government funding, multiple

clusters are at a disadvantage compared to one large cluster because each district will be in

need of some basic infrastructure. While a single large technology park would benefit from a

concentration of government funds and fully-fledged support policies to provide

technological infrastructure, research funds and incentives, in a situation of multiple clusters

such funds would have to be split up to some degree in order to avoid wasteful duplication of

infrastructure and service platforms. In any case, multiple clusters in the same industry within

a mega-city region would have to compete for funds and could suffer from funding shortages.

3.2 Collaborating versus competing clusters

When multiple-cluster configurations emerge, as is particularly likely in the context of

a quickly-developing economy and a strong state, they can take a variety of forms. The

15

primary processes that drive technological learning within clusters are associated with

interactions between firms and suppliers, customers, universities, research institutes and

KIBS. Other important linkages include those with government agencies, financial

organizations and intermediary agencies that support networking. Depending on the way in

which co-located clusters within the same industry develop, we can expect to find different

types of linkage and interaction patterns between the clusters in terms input-output flows,

labor market relations and knowledge networks (Delgado et al., 2016).

Specifically, we identify two ideal-type scenarios of linkage and interaction patterns

below: (i) collaborating clusters and (ii) competing clusters (Figure 1).

******************************

Insert Figure 1 about here

******************************

(i) Collaborating clusters. In this scenario, multiple industrial clusters in the same

industry within a mega-city region establish a system of collaborating clusters that

successfully exploit proximity advantages based on social, institutional, cultural and

cognitive affinities. While automated knowledge flows and communications, accompanied by

mutual interactions and transactions, develop within each specialized industrial cluster, close

learning processes and inter-firm linkages also develop across clusters based on

complementarities. Such linkages can be generic or constructed by policy. They can generally

be expected to have characteristics typical of ‘pipelines’ (Bathelt et al., 2004). This situation

may develop in a growing mega-city region characterized by islands of coherent industry

infrastructure, such as specialized research institutes. Initial start-up and location decisions or

targeted industrial policies lead to a functional and spatial division of labor, with different

parts of an industry’s value chain being drawn to different locations that fit their needs best.

While originally being part of a single development, separate clusters can develop over time

16

if these districts are able to attract specific networks of suppliers and services providers and

to develop their own labor markets.

Because of complementarities, the clusters in this scenario do not remain separate but

develop linkages to one another in order to access each other’s related but different

capabilities. Expressions of such linkages are investment activities across clusters and firms

through which specialized branch facilities are established in multiple districts that each

benefit from localized resources. Besides these internal and cross-cluster interactions,

external linkages to other technology hubs and the world market also, of course, exist.

Overall, this scenario can be described as one of multiple clusters that are characterized by

functional specialization. They are co-located in the same region and develop close

collaborative linkages to each other based on complementarities.

(ii) Competing clusters. In this scenario, the clusters in a mega-city region focus on

the same resources and their relations are predominantly competitive. Firms across the

different districts specialize on similar tasks, develop similar capabilities and become rivals.

As a result, they may not be interested in collaborating with each other or sharing knowledge.

If the different districts grow to a certain size, they attract similar sets of services and supply

functions and develop their own internal networks. This does not affect cluster-specific

knowledge and innovation dynamics, but few incentives exist to encourage the establishment

of cross-cluster linkages (e.g. Maskell and Lorenzen, 2004; Lu et al., 2013). The clusters thus

become separate entities with few linkages and little interaction between them. They are

characterized by localized buzz dynamics as well as pipelines with other regions and world

markets, but there are no systematic inter-connections between the various clusters. Since

these clusters specialize in procuring and competing for the same resources, such as specific

government funds and preferential access to the innovation base, a number of potential

threats or negative consequences can result. Factor prices can be high and knowledge

17

spillovers limited, basic infrastructures have to be duplicated, and parallel investments may

occur. While this may have positive individual effects at the corporate level and stimulate

product differentiation, the overall effect on the city region may be costly and lead to

detrimental effects. To draw further conclusions regarding this scenario, of course, first

requires identifying and investigating such a configuration.

4. Methodology

The goal of this paper is to identify whether the type multiple-cluster configuration in

the Beijing biomedical industry is closer to the ideal type of competing clusters or

collaborating clusters by investigating the cluster structures that have emerged and the

linkages that have developed between them. In an attempt to develop a better understanding

of the phenomenon of multiple clusters in mega-city regions – and inspired by the idea of

grounded theory which provides a set of useful research strategies for the discovery of theory

through the analysis of data (Glaser and Strauss, 1967; Eisenhardt, 1989) – we adopted a

qualitative explorative approach as the basis of our research into the three biomedical

industrial/research parks in Beijing. In doing so, we utilized a package of heterogeneous

research methods to gather and analyze primary and secondary data that were collected over

an extended time period and carefully triangulated and verified, supported by a comparative

approach (Glaser, 1998; Yin, 2003; Tokatli, 2015).

(i) Database of firms. First, we developed a database of all firms in the three

industrial/research parks located in Beijing. This database was established from data

published in industrial park directories, through various government organizations, China’s

National Bureau of Statistics, as well as websites of firms and organizations in the Beijing

biomedical industry. Our final database for the year 2012 consists of 390 biomedical firms in

Yizhuang Park, 112 firms in ZLS Park and 337 firms in Daxing Park.

18

(ii) Firm and case selection. Our database and previous research provided the basis

for the identification of intra- and cross-cluster linkages and networks in production and

research. Previous investigations, which permitted us to identify important cases of

individuals, firms and other organizations and their linkage patterns, consisted of four

projects conducted between July 2003 and December 2012 (+Wang et al., 2006; Zhao, 2006;

2010; 2012; Niosi and Zhao, 2012; Zhao and Richards, 2012). Crucial-case information was

complemented by and triangulated with data received in telephone interviews (Seawright and

Gerring, 2008).

(iii) Semi-structured interviews. Between February 2013 and March 2014, we

conducted relatively short semi-structured telephone interviews with 84 biomedical firms

across the three industrial/research parks. The firms were selected as a systematic random

sample from our database, stratified by location, size, type, and country of origin of the firms.

In each interview, we highlighted the confidential nature of this research and decided not to

record the conversations to maximize our response rate. The interviews were with human

resources, technology and product managers and took on average 15 to 20 minutes.

While the case information from previous research provided information about firm-

specific linkages, ownership patterns and investments within and between the clusters, we did

not have direct evidence about the occurrence and intensity of intra- and cross-cluster

interaction patterns. During the telephone interviews, we verified the firm information

collected in our database and asked question about interactions and linkages within, between

and beyond the three biomedical industrial/research parks. The questions focused on the

nature and frequency of cooperation and interaction patterns with suppliers and knowledge-

intensive business services (KIBS), customers, universities, research organizations, other

biomedical firms, training schools and government agencies. Since we were only able to

write-up ex-post summaries of these interviews, we cannot provide direct quotations.

19

In the empirical analysis outlined below, we mostly refer to cases of cluster, firm and

organizational linkages identified in prior research; our interview data serves the primary

function of helping us verify our conclusions regarding the existence of intra- and cross-

cluster interaction patterns.

5. Context: Three biomedical clusters in Beijing

The development of the biomedical industry in Beijing can be traced back to the

1980s. Despite this early start, it took until 2002 before the industry took off, benefitting from

strong government support through the “Beijing Bioengineering and Pharmaceutical Industry

Development Promotion Program”. In October 2006, the National Development and Reform

Commission (NDRC) assigned Beijing the status of a national biomedical industry base and

provided financial resources for its development. This included support for the three

biomedical industrial/research parks: ZLS Park, Yizhuang Park and Daxing Park (NDRC,

2006). Due to its rapid growth, the Chinese Academy of Social Sciences (CASS) identified

the industry as one of Beijing’s five pillar industries in 2010 (CASS, 2010). Around the three

industrial/technology parks, three economic clusters developed over time (Table 1):

Yizhuang Park is located in the Southeast of Beijing. The park was established as a

biomedical industry park in 2009 and became part of the Beijing Economic and

Technological Development Area (BETDA). Founded in 1994, the area developed into an

important base for high-technology and other modern manufacturing industries. This was the

site where biomedical firms first started to agglomerate. Firms in Yizhuang Park enjoy

twofold preferential policies as they are part of both a national economic and technology

development area and a national high-technology industry park. In 2012, the park hosted

more than 1,300 firms, 390 of which were related to the biopharmaceutical industry. Many of

the firms are well-known multinational firms from over 30 different countries that invested in

20

the city region’s growing bioengineering and biomedical fields. The growth potential of this

industry/research park is, however, limited to the territory of BETDA.

Daxing Park, established in 2002, is located in the Beijing Daxing Industry

Development Area in the South of the city region. The park has excellent infrastructure

conditions and is well-connected to the city region’s traffic networks. It is characterized by

biopharmaceutical production, especially generic firms with integrated technological

innovation capabilities, as well as by a range of other biomedical activities from traditional

Chinese medicine to animal vaccines. The park focuses on manufacturing facilities combined

with R&D and has developed into a world-class location for modern bioengineering and

pharmaceutical production. In 2012, 337 biomedical firms were located here and the park

spanned an area of about 960 hectares (about 2,400 acres). Daxing Park has become the

largest biomedical industry base in the city region in terms of the number of firms and the

size of occupied land. It still has some peripheral areas that are available for future expansion.

ZLS Park was founded in 2000 and is located in the Beijing Zhongguancun

Development Area in the North of Beijing, close to crucial knowledge resources for the

biomedical industry. The park has the highest density of advanced university and research

organizations in China (Zhou, 2008). It was established with the intention of creating a

leading international science park that aims to link life science research with enterprise

incubating, pilot production, biotechnology exhibitions, venture capital funding and specific

training opportunities. ZLS Park is the primary research base for the national life science and

new pharmaceutical high-technology industry. In 2012, the park hosted 112 firms, including

18 subsidiaries/branches of foreign firms. ZLS Park covers an area of nearly 250 hectares and

will be extended eastwards and westwards to a total of 500 hectares (about 1,250 acres) in the

future.

21

Overall, the three biotechnology parks experienced substantial growth. By 2014,

Beijing’s biomedical industry had an astonishing 80,000 employees (BMBS, 2015) and was

one of the largest agglomerations of this industry worldwide. The three biomedical parks can

be viewed as individual clusters (and will be referred to as such below) because they integrate

large parts of the respective value chains and support infrastructure, consisting of specialized

suppliers and service providers, research and training facilities, and government organizations

and associations (Zhao, 2010). In our telephone interviews, firms indicated that they had

strong labor market, supplier and research linkages with other firms and organizations within

their parks.

The three clusters, however, are not identical in their structure but have developed

different specializations and functions over time: ZLS Park puts particular emphasis on

upstream R&D activities, pilot production and business incubation; Daxing Park is an

industrial manufacturing base engaged in specialized production activities; and Yizhuang

Park focuses on biopharmaceutical contract manufacturing (Table 1). As will be

demonstrated in the next section, the Beijing biomedical cluster configuration is similar to the

‘collaborating clusters’ scenario depicted in Figure 1, being characterized by internal

production, research and knowledge dynamics but also strong patterns of cross-cluster

networks. While our analysis focuses on intra-metropolitan linkages, it should be noted that

each of the three biomedical clusters also has strong linkages with other regions nationally

and internationally.

******************************

Insert Table 1 about here

******************************

22

6. Relational ties within and between the biomedical

clusters in Beijing

This section takes a closer look at the relational ties (Bathelt et al., 2004; Li et al.,

2012) that exist within and between the three biomedical clusters in Beijing. Qualitative

evidence is presented from prior and current empirical observations that enable us to identify

a specific cluster configuration in the Beijing biomedical industry. In the first subsection, we

will illustrate that the three biomedical parks in Beijing can indeed be viewed as clusters with

strong internal supplier and service networks. The clusters, although being part of the same

industry, have each developed their own functional specialization with corresponding

knowledge-related services and nearby research facilities. The following subsections present

evidence of systematic cross-cluster linkages and support our hypothesis that the Beijing

biomedical industry is best understood as a configuration of collaborating instead of

competing clusters. This will be shown by investigating networks of firms, government-led

collaborations and co-operations with universities and other research institutes.

6.1 Localized biomedical cluster dynamics

When investigating whether the three industrial/research parks in Beijing can indeed

be viewed as separate clusters that generate their own internal dynamics and networks, the

question arises as to why the multiple-parks structure developed to begin with. In the case of

Beijing, this development depended in large part on industrial policies pursued at various

levels of government, and was therefore a consequence of the strong state capacity

characteristic of a developmental state (Wade, 1990). As Guo Li, the General Manager of

ZLS Park, emphasized in an interview conducted in 2010, the three industrial areas were

purposely developed with the goal of generating specialized biomedical parks:

“(1) The Beijing government applied a strategic policy of multiple clusters as

follows. ZLS Park is characterized by biomedical R&D. Yizhuang Park is

23

located in the national economic development zone. Here, the costs of land

and other resources are expensive compared with the other two parks, which is

why we find an orientation on high-end manufacturing activities. Daxing Park,

which has a larger territory compared with the other parks, concentrates on

pharmaceutical manufacturing, especially the core activities of manufacturing

traditional Chinese medicine, medical apparatus and instruments, as well as

animal vaccines. (2) The output value of these three clusters accounts for more

than 60 percent of the Beijing biomedical industry and will reach 70 percent

by the end of the Twelfth Five-Year Plan” (translated and revised from

Chinese).

Dedicated industrial and technology policies induced start-up processes and provided

incentives for the development of specialized KIBS and of other suppliers and service

providers in each of the parks. This was supported by close-by research institutes and training

facilities that fit the specific needs of the different clusters. From this basis, a functional

division of labor developed between the parks that resulted in the establishment of three

distinct functional clusters with specific local labor market dynamics. The telephone

interviews strongly confirmed the existence of close intra-cluster linkages and knowledge

networks. According to our interviews, the following patterns emerged over time:

(1) Almost all of the firms interviewed, whether large or small and medium-sized

enterprises, had close relationships with universities, research organizations, training schools

and KIBS in their respective parks. Many large firms had post-doctoral centers and offered

temporary positions for coop students from near-by universities and colleges.

(2) Many of the firms described themselves as being part of the same park-specific

research networks or alliances and specializing in particular functions and product segments

that were different from those in the other parks. These specializations supported internal

24

cluster dynamics with respect to service providers and suppliers, but also generated

complementarities between the clusters. These, in turn, provided the basis for the

establishment of cross-cluster linkages.

(3) Our telephone interviews also provided evidence of close localized knowledge

ecologies in the three parks associated with regular personal meetings and knowledge

exchanges. The park administrations, for instance, frequently organized events, such as

academic presentations, government workshops, product promotions and even entertainment

and sporting activities. These events were all regularly and well attended by local firms and

generated a constant flow of knowledge about what was going on in the industry, the

regulatory environment and at the park in general.

These interview findings show that the biomedical industry in Beijing is organized in

the form of separate clusters that developed their own dynamics. Our interviewees described

highly localized cluster dynamics and identified local biomedical innovation networks. It is,

of course, not unusual that the biomedical industry is highly clustered and shaped by network

forms of organization (Powell, 1990; Podolny and Page, 1998). Similar structures and

tendencies have developed in each of Beijing’s biomedical clusters, driven by linkages

between firms, research organizations and government institutes. Important relational nodes

that encourage such interactions and knowledge exchanges within the three clusters include

the following:

(1) Bacterial cultures originate from key national laboratories and research institutes,

while pre-clinical trials are given to local contract research organizations such as JOINN

Laboratories in Yizhuang Park.

(2) Park-specific industry alliances and associations have developed, such as the

Alliance of Biotech Outsourcing, the BDA Alliance of Innovative Technology Service for

25

Biopharmaceutical Industry, the Alliance of Zhongguancun Contract Research Organizations

and the Northern Antibody Industry Alliance.

(3) Technical service platforms have been established in the parks supported by

government programs to provide localized services, including delivery and market

development services, to firms.

(4) Incubators, for instance in ZLS Park, have actively formed innovation chains that

include basic research, technological innovation, industrial support and clinical resources.

(5) Specialized KIBS in the various parks provide crucial services to and are linked

with many, especially local, biomedical firms.

These organizations establish the basic architecture for cluster and network formation

in the three biomedical parks (Zhou and Dai, 2008; Zhao and Richards, 2012). With respect

to KIBS, for instance, more than 100 firms and organizations are located in Beijing that focus

on R&D services for the biomedical industry. One-sixth of all international clinical trials in

China are conducted in the city region, one-fifth of all Chinese drug safety evaluation centers

are located here, and two-thirds of all biopharmaceutical scientists and engineers in China

live in Beijing. This large concentration of research activities goes along with highly

localized linkages and supports local innovation networks. To benefit from this, biomedical

firms locate in those areas that fit their functional and product specialization best. It is not

surprising that many of the firms in a park are attached to the same industry associations and

alliances, a similar mix of shareholders, related university or research institutes and the same

group of leading biomedical innovators.

While the above evidence suggests that we are faced with three separate clusters that

are characterized by specific localized linkages and networks, the next subsections take a

closer look at the relationships between the three biomedical clusters.

26

6.2 Complementary specialization and cross-cluster cooperation

Compared with other biomedical clusters discussed in the literature, the Beijing

situation is a special one because it involves both close intra-cluster networking along with

strong cross-cluster linkages that connect three biomedical parks in the same city-region. Our

findings suggest that this case resembles the scenario of ‘collaborating clusters in the same

industry’ (Figure 1). Systematic cross-cluster linkages have been enabled by the development

of biomedical parks that are characterized by complementary specialization, as identified in

our firm database (Table 1). When investigating the biomedical producers, suppliers and

service providers across the three parks, we found a clear specialization pattern: more than

half of the firms from ZLS Park were biomedical R&D firms; almost 60 percent of the firms

from Daxing Park were active in the areas of traditional Chinese medicine, medical

apparatus/instruments, animal vaccine, agricultural biotechnology and veterinary medicine;

and about 70 percent of the firms in Yizhuang Park were specialized on contract

manufacturing with a focus on large-scale medical equipment, chemicals and biological

pharmaceuticals. While ZLS Park focused on attracting research facilities from international

biomedical research groups, Yizhuang Park concentrated on transnational manufacturing

investments and Daxing Park on domestic and multinational innovative firms with

specialized production activities.

Whereas these functional specializations stimulated cluster development in each park,

they were also the basis for the establishment of cross-cluster linkages. Responses in the

telephone interviews suggested that each park developed closely-knit internal value chain

linkages. However, due to complementarities, firms extended their activities also to other

parks in order to benefit from related knowledge pools and sets of competencies.

Interviewees frequently emphasized that such linkages created regular interaction and

knowledge exchanges between the clusters. For example, research facilities and incubators

located in ZLS Park focused on biomedical research and related engineering, but were also

27

open to firms located in other Beijing parks, creating cross-cluster knowledge flows.

Similarly, Daxing Park was home to a concentration of national-level organizations, such as

the National Institutes for Food and Drug Control, the Institute of Materia Medica (IMM), the

Chinese Academy of Chinese Medical Science, the China Animal Disease Control Center

and the National Veterinary Microorganism Center. While these organizations played an

important role for Daxing firms, they were also important contacts and partners for firms

from the other parks and generated interactions between the other parks and Daxing.

Altogether, plenty of opportunities for collaborations and complementary linkages between

the clusters were generated through the specialization of the three parks.

To make full use of the advantages of the parks and their complementarities, firms

established or co-founded subsidiaries or branches in several of the parks – a practice which

promoted close interaction and flows of material and knowledge. We identified numerous

firms with such multiple locations, especially large biomedical firms. Due to their

specialization, each of the clusters offered a set of proximity advantages to different types of

resources, such as technological capabilities, skills sets, research streams or transaction

partners. An example of such linkages is shown in Figure 2. The firm Beijing Medical

University United Biological Engineering Co. (BMUC), which is at the center of this case,

established two research-driven firms in ZLS Park and two manufacturing-centered

biomedical firms in Yizhuang Park. In addition, one of BMUC’s subsidiary firms established

two other firms: one located in Yizhuang Park which focused on production and sales

activities, the other one as an incubator located in ZLS Park to benefit from the park’s unique

innovation and R&D resources. Interviewees pointed out that these corporate linkages created

frequent interactions, material flows and site visits between the parks.

We identified a number of similar cases in our interviews where larger biomedical

firms had invested in several clusters simultaneously in order to take advantage of specific

28

talents, preferential policies, land prices and research specializations that were targeted at

different stages of the value chain. In our interviews, these investments were closely linked to

one another and created value-chain-based interactions between the clusters (see, also,

Turkina et al. 2016). They also supported joint research projects between firms at similar

stages of the value chain that belonged to different clusters.

******************************

Insert Figure 2 about here

******************************

These findings clearly differ from the scenario of competing clusters in Figure 1. If

this configuration was prevalent, we would not expect to find such intensive linkages across

the clusters as the different clusters would perform similar tasks and compete against each

other. In contrast, our study provided strong support that collaborative relations developed

between the three clusters, as firms established pipelines in research and manufacturing

across the parks to simultaneously take benefit of various specialized resources. From our

interviews, we identified both intra-cluster dynamics and systematic inter-cluster networks. A

typical pattern of knowledge linkages according to type, closeness and frequency of contacts

for a large biomedical firm in Beijing is shown in anonymized form in Figure 3. The

exemplary biomedical firm in Figure 3 was located in Park A and had important linkages

within the park, as well as with Parks B and C. Intra-cluster linkages involved collaborations

with the administrative committee, a training school and close relations with local KIBS.

Collaborations with training schools and KIBS also extended to Parks B and C and resulted

in cross-cluster training and hiring schedules and knowledge linkages along with

collaborations in research projects across clusters. In terms of cross-cluster interaction, the

biomedical firm also engaged in collaborations with an industrial association, an incubator

and one R&D firm in Park C, as well as with one co-founded subsidiary, one university and

29

research unit and one national institute located in Park B. In our interviews, we found

evidence of numerous similar cases of firms with intra- and cross-cluster linkages, suggesting

that the biomedical industry in Beijing developed into a configuration of multiple

collaborating clusters as indicated in Figure 1.

******************************

Insert Figure 3 about here

******************************

6.3 Government-led development and opportunities for local buzz

Government policies in the development of the biomedical industry in Beijing seemed

to pursue two sets of goals. On the one hand, there was a plan to establish a multi-nodal

structure in the growing biomedical industry that would enable collaboration, rather than

competition, between different localities. On the other hand, targeted policies were put in

place to support the development of self-sufficient cluster relations in each of the industry

parks. Delivery of these policies was facilitated by the strong capacity of the Chinese state,

yet we should not assume that the development was perfectly coordinated and under full state

control (Conlé and Taube, 2012; Zhao and Richards, 2012). The various policies also did not

all start at the same time. While ZLS Park and Daxing Park were founded as government-led

industrial clusters in the years 2000 and 2002, respectively, support in Yizhuang was

launched in 2009 to improve the park’s infrastructure and service platform for biomedical

firms. However, government policies and programs tried to consider the overarching

industrial development plan, as well as complex factors such as land use policies and aspects

of resource access (proximity).

Specific government initiatives drove the development of local cluster dynamics and

encouraged cross-cluster interaction and knowledge linkages. For example, academic

conferences, seminars/forums, government workshops (often associated with the

30

announcement of new support programs), product promotions and other collective activities

were frequently organized by Beijing government institutes and the management committees

of the parks. These government events were viewed by firms as important because they

provided access to crucial information about new developments. Not surprisingly, over 80

percent of the firms interviewed took advantage of this access by sending their employees

regularly to engage in this kind of professional activity. Three-quarters of firms sent

employees to government workshops. Several interviewees suggested that these get-togethers

also generated manifold opportunities for knowledge exchanges among firms and led to new

collaborations. At the same time, these events also enabled cross-cluster knowledge flows,

since firms sent their employees regularly to attend professional activities held in the other

parks.

The most important effect of these kinds of government policies was the facilitation

of linkages within the three parks in a way that generated local cluster dynamics. Xiaochen

Zhao, the deputy director for Daxing Park Administrative Committee, explained this for his

park as follows: “We made agreements with two institutes: the National Veterinary

Microorganism Center – a veterinary drug inspection institute – and the National Institutes

for Food and Drug Control – a top-tier inspection institute for human pharmaceutical and

medical equipment. When the two institutes were established in Daxing Park, they had to

offer specialized professional services which prioritized firms located in the park” (translated

and revised from Chinese). The firms we interviewed suggested that the resulting knowledge

economies had the dynamism typical of ‘local buzz’ (Bathelt et al., 2004).

To support cluster development, government policies were designed to attract

specialized biomedical firms to the three parks by establishing research organizations,

regulatory agencies and other facilities conducive to the development of knowledge networks

and inter-organizational collaborations. Daxing Park, for instance, became an important

31

reference point for firms from the other parks due its local concentration of government-

related institutes and organizations. These included technology transfer organizations,

industry associations, industry alliances, venture capital organizations, professional services,

training organizations and research institutes. These facilities were beneficial to the entire

range of biomedical activities and therefore for firms in all of the parks. Our interviewees

pointed out that complementary organizations were also set up in the other parks and

generated intensive interactions and mobility between the clusters, leading to longer-term or

project-based collaboration between firms. The combination of these influences supported the

development of a multiple-cluster configuration with systematic collaborative linkages across

the individual clusters.

6.4 Universities and research institutes as technology sources and linkage nodes

A crucial feature of the configuration of collaborating clusters in the same industry is

the development of durable linkages between the individual entities. While shaped and

planned by government programs, the linkage patterns were not controlled by such programs.

They were fundamentally supported and stabilized in decentralized ways by the activities of

universities and research institutes. These provided services to nearby park firms, but had

much wider effects across all three clusters and hence also provided incentives for firms to

establish linkages between the parks. As shown below, these facilities were not just enablers

but played an active role in providing technologies across the clusters and generating cross-

cluster networks themselves.

Since the healthcare, biotechnology and biomedical industries generally have a strong

science base, university research is intimately interwoven with this sector (Pisano, 2006). In

fields with well-developed academic infrastructure, university-related start-up and spin-off

activities have generally become important drivers of industry development (Conlé and

Taube, 2012). Some scholars have found that personal ties and involvement, aside from these

32

organizational ties, also play an important role in cluster development and catch-up processes

(Lorenzen and Mudambi, 2013). In China, these kinds of developments have occurred

particularly in the biomedical industries in Beijing and Shanghai. The majority of biomedical

firms have ties with research institutes, sometimes based on technology transfer and

sponsored research. A significant number of firms also have capital relations with academic

organizations. In one type of spin-off, the sponsoring academic institute actively participates

in the development of a firm; in the other type, the institute primarily supports autonomous

entrepreneurial activities by staff members or graduates (Conlé and Taube, 2012). This

contributes to the development of localized biomedical labor markets, as spin-off firms

engage with students from close-by research facilities, for instance by offering temporary

workplaces for doctoral and post-doctoral research.

In Beijing’s biomedical industry, we found that universities and other research

institutes were in similar ways crucial sources for technology development and technology

transfer, and that they were important linkage nodes in the firms’ networks. Their capacity to

perform these functions was driven by the establishment of spin-off firms and the active

licensing and sale of specific technologies to local firms.

Our telephone interviewees pointed out that some leading university scientists with a

strong academic reputation, who had made innovative discoveries, contributed in important

ways to the establishment of cooperation patterns between firms within and across the

clusters. For example, Yunde Hou, the director of the Institute of Virology at the Chinese

Academy of Preventive Medicine, actively contributed to intra- and cross-cluster knowledge

flows and collaborations. Technologies developed under his supervision at the Chinese

Academy of Preventive Medicine provided the basis for him to start two new firms and

establish one share-holding biotech firm. These were located in different parks and created

knowledge pipelines between them: Beijing Yuance Pharmaceutical and Beijing Kawin

33

Technology were established in Yizhuang Park, while Beijing Tri-Prime Genetic Engineering

was founded in Daxing Park.

Another example of the centrality of universities and research institutes to the

building of cross-cluster knowledge architecture is the Institute of Materia Medica (IMM), a

part of the Chinese Academy of Medical Science and the Peking Union Medical College

(Figure 4). IMM became a very important source of biomedical technologies and a crucial

reference point for cluster firms located in the three Beijing biomedical parks. Through

wholly- or partially-owned facilities across the parks, as well as technical support activities,

IMM linked up with a large number of organizations and actors within and across the cluster

locations. IMM established a network of R&D centers and science and technology firms

across the Daxing, Yizhuang and ZLS Parks, which developed wide-ranging networks and

collaborations across the Beijing biomedical industry. IMM is major shareholder of three

firms in different parks. In addition, IMM established Peking Union Lawke as a joint venture

and co-founded the National Institute of Biological Sciences, both in ZLS Park. It also

provided technical support to the Beijing National Engineering Research Center of

CapitalBio Corporation and Biological Chip with facilities in ZLS Park and Yizhuang Park.

Additionally, IMM was linked to an affiliated hospital: the Beijing Union Medical Center in

ZLS Park. This resulted in extended interaction and the development of networks both within

and between the clusters. It supported both cluster formation as well as cross-cluster

cooperation through linkages in research, development and production.

******************************

Insert Figure 4 about here

******************************

Taken together, these findings provide strong evidence that the Beijing biomedical

industry is organized in the form of three functional clusters that are self-sufficient and

34

characterized by dynamic internal linkages networks, yet are at the same time closely linked

through systematic cross-cluster investment, manufacturing and research connections, similar

to the scenario of collaborating clusters in Figure 1.

7. Conclusion

This paper has discussed the phenomenon of multiple clusters in the same industry

that are co-located in different districts of the same mega-city region. We focus on the case of

the biomedical industry in Beijing, in which three industrial/research parks were investigated

– Daxing Park, ZLS Park and Yizhuang Park – each of which exhibited strong clustering

tendencies. We identified these as a configuration of collaborating clusters that each have

specialized cluster dynamics, yet are characterized by systematic and complementary cross-

cluster linkages and interactions. While the context of Beijing is that of a mega-city region in

one of the fastest growing economies in the world and in a country where the state has very

high capacity, the situation in other developing contexts may not be fundamentally different

and we may find similar configurations elsewhere. The example of Seoul’s ICT clusters

already indicates that we may be looking at a broader phenomenon. However, it is simply too

early at this point to establish generalizations about the conditions under which such

developments occur and whether this may also be possible in a context without dedicated

government policies.

Although this phenomenon will require more research in the future, one implication

of this paper is that cluster theory needs to be more closely aligned with urban

planning/studies, as the configuration of multiple clusters discussed in this paper is closely

linked to very large urban contexts, infrastructures, labor markets, and so on. Full

investigation of these wider issues is a task, however, that goes beyond the scope of this

paper. Multiple co-located clusters require a minimum urban scale and are associated with

localized spillovers resulting from interdependencies with other urban clusters. Such co-

35

location creates manifold opportunities to develop intra-cluster and cross-cluster linkages that

feed back into urban and regional development. Similar cluster configurations with separate

infrastructures, labor markets and supplier networks are unlikely to occur in a small-scale

urban environment.

The phenomenon of multiple-cluster configurations in mega-city regions can be

viewed, following Phelps (2004), as a new diffuse form of agglomeration. This phenomenon

certainly warrants more discussion about its advantages, as well as potential problems and

challenges, because it differs substantially from the standard single cluster cases discussed

most often in the literature. Challenges can emerge if different urban districts start competing

for the same sets of resources, such as infrastructure and support programs, while positive

externalities through cross-cluster linkages are not automatic. This paper attempts to shed

some light on this phenomenon by discussing two extreme scenarios of multiple-cluster

configurations in mega-city regions. The empirical evidence presented suggests that the

biomedical parks in Beijing can be viewed as collaborating clusters with their own cluster

networks that operate in relation to each other (Parr, 2002). The three parks are each highly

specialized in different R&D and/or production functions and attract different types of

biomedical activities. They also have an infrastructure of government organizations, training

facilities, universities/research institutes and specialized service providers, all of which

generate cluster-specific knowledge networks.

Aside from internal cluster dynamics, we identify strong tendencies for cross-cluster

linkages and knowledge networks. Large multi-location firms, universities and their leading

scientists, as well as government laboratories and organizations, do not only act as knowledge

gatekeepers inside the three biomedical parks (Cohen and Levinthal, 1990), but become

crucial reference points for interactions across the clusters and form active linkage nodes that

connect individuals and organizations across the mega-city region. Leading organizations and

36

individuals are linked to different facilities and branches in different parks and actively

develop networks between the clusters. Overall, the development in these co-located

collaborating clusters is driven by both intra-cluster specialization advantages and synergies

that trigger vibrant localized knowledge ecologies, as well as inter-cluster complementarities

that provide the basis for the establishment of cross-cluster linkages and knowledge networks.

Acknowledgements

This paper was initially presented as part of the Jena Lecture Series 2015 at the

University of Jena, Germany, organized by Sebastian Henn and the Department of

Geography. Aside from Sebastian Henn, Susann Schäfer and the participants of the Jena

Lecture Series, who provided many valuable suggestions, we would like to thank Daniel

Hutton Ferris, Sufyan Katariwala and Peng-Fei Li as well as the Reviewers for providing

critical and constructive criticism on earlier drafts. Special thanks go to the Editor Padraig

Carmody for his guidance and support. This research was funded by the Canada Research

Chair in Innovation and Governance at the University of Toronto and a Postdoctoral

Fellowship by the University of Toronto’s Department of Mathematics. We particularly wish

to thank Professor William Weiss for his generous support.

References

Asheim, B. T. (2000) Industrial districts: The contributions of Marshall and beyond, in G. L. Clark, M. P. Feldman and M. S. Gertler (eds), The Oxford Handbook of Economic Geography, Oxford University Press, Oxford, 413-431.

Bathelt, H. and Li, P.-F. (2014) Global cluster networks – Foreign direct investment flows from Canada to China, Journal of Economic Geography, 14, 45-71.

Bathelt, H. and Taylor, M. (2002) Clusters, power and place: Inequality and local growth in time-space, Geografiska Annaler B, 84, 93-109.

Bathelt, H., Malmberg, A. and Maskell, P. (2004) Clusters and knowledge: Local buzz, global pipelines and the process of knowledge creation, Progress in Human Geography, 28, 31-56.

Beaudry, C. and Schiffauerova, A. (2009) Who’s right, Marshall or Jacobs? The localization versus urbanization debate, Research Policy, 38, 318-337.

37

Blundel, R. and Thatcher, M. (2005) Contrasting local responses to globalization: The case of volume yacht manufacturing in Europe, Entrepreneurship and Regional Development, 17, 405-429.

BMBS – Beijing Municipal Bureau of Statistics (2015) Beijing Statistical Yearbook, China Statistics Press, Beijing (in Chinese). Available from: http://www.bjstats.gov.cn (June 28, 2015).

IMM – Institute of Materia Medica (2014) About IMM: Introduction, Beijing. Available from: http://www.imm.ac.cn/en/about.asp (November 29, 2014).

Boschma, R. A. (2005) Proximity and innovation: A critical assessment, Regional Studies, 39, 1-14.

Camagni, R. (ed.) (1991) Innovation Networks: Spatial Perspective, Belhaven-Pinter, London.

CASS – Chinese Academy of Social Sciences (2010) Blue Book of the Industry: Annual Report on Industrial Competitiveness of China, Social Sciences Academic Press, Beijing (in Chinese).

Chakravorty, S., Koo, J. and Lall, S. V. (2003) Metropolitan Industrial Clusters, Policy Research Working Paper 3073, The World Bank Development Research Group, Infrastructure and Environment, Washington, DC.

Chen, S.-H. (2004) Taiwanese IT firms’ offshore R&D in China and the connection with the global innovation network, Research Policy, 33, 337-349.

Cohen, W. M. and Levinthal, F. A. (1989) Innovation and learning: The two faces of R&D, Economic Journal, 99 (397), 569-596.

Conlé, M. and Taube, M. (2012) Anatomy of cluster development in China: The case of health biotech clusters, Journal of Science and Technology Policy in China, 3, 124-144.

Crevoisier, O. (2001) Der Ansatz des kreativen Milieus: Bestandsaufnahme und Forschungsperspektiven am Beispiel urbaner Milieus (The creative milieu: State of the art, research perspectives and the case of urban milieus), Zeitschrift für Wirtschaftsgeographie, 45, 246-256.

Crevoisier, O. and Jeannerat, H. (2009) Territorial knowledge dynamics: From the proximity paradigm to multi-location milieus, European Planning Studies, 17, 1223-1241.

Delgado, M., Porter, M. E. and Stern, S. (2010) Clusters and entrepreneurship, Journal of Economic Geography, 10, 495-518.

Delgado, M., Porter, M. E. and Stern, S. (2016) Defining clusters of related industries, Journal of Economic Geography, 16, forthcoming.

Eisenhardt, K. M. (1989) Building theories from case study research, The Academy of Management Review, 14, 532-550.

Fornahl, D. and Tran, C. A. (2010) The development of local-global linkages in biotechnology districts in Germany: Between local embeddedness and international knowledge sourcing, in F. Belussi and A. Sammarra (eds), Business Networks in Clusters and Industrial Districts: The Governance of the Global Value Chain, Routledge, London and New York, 332-355.

38

Fornahl, D., Broekel, T. and Boschma, R. (2011) What drives patent performance of German biotech firms? The impact of R&D subsidies, knowledge networks and their location, Regional Science, 90, 395-418.

Gibson, R. and Bathelt, H. (2014) Proximity relations and global knowledge flows: Specialization and diffusion processes across capitalist varieties, in A. Torre and F. Wallet (eds), Regional Development and Proximity Relations, Edward Elgar, Cheltenham and Northampton, MA, 291-314.

Glaser, B. (1998) Doing Grounded Theory – Issues and Discussions, Sociology Press, Mill Valley, CA.

Glaser, B. and Strauss, A. (1967) The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine, Chicago.

Hamaguchi, N. and Kameyama, Y. (2007) Dense Communication and R&D in Knowledge-Based Industrial Clusters: Comparative Study of Small & Medium-Sized Firms in Korea and China. RIEB Discussion Paper, No. 207, Kobe University, Kobe. (ICSEAD Working Paper, No.2007-21, Kitakyushu: The International Centre for the Study of East Asian Development).

Haug, P. (1995) Formation of biotechnology firms in the greater Seattle region: An empirical investigation of entrepreneurial, financial and educational perspectives, Environment and Planning A, 27, 249-267.

Hsu, J.-Y. and Saxenian, A. (2000) The limits of guanxi capitalism: Transnational collaboration between Taiwan and the USA. Environment and Planning A, 32, 1991-2005.

IMM – Institute of Materia Medica (2014) About IMM: Introduction, Beijing. Available from: http://www.imm.ac.cn/en/about.asp (November 29, 2014).

Keeble, D. and Wilkinson F. (1999) Collective learning and knowledge development in the evolution of regional clusters of high-technology SMEs in Europe, Regional Studies, 33, 295-303.

Lagendijk, A. and Cornford, J. (2000) Regional institutions and knowledge – Tracking new forms of regional development policy, Geoforum, 31, 209-218.

Li, P.-F., Bathelt, H. and Wang, J. (2012) Network dynamics and cluster evolution: Different trajectories in the aluminum extrusion industry in Dali, China, Journal of Economic Geography, 12, 127-155.

Lin, G. C. S., Wang, C. C., Zhou, Y., Sun, Y. and Wei, Y. D. (2011) Placing technological innovation in globalising China: Production linkage, knowledge exchange and innovative performance of the ICT Industry in a developing economy, Urban Studies, 48, 2999-3018.

Lissoni, F. and Pagani, M. (2003) How many networks in a local cluster? Textile machine production and innovation in Brescia, in D. Fornahl and T. Brenner (eds), Cooperation, Networks and Institutions in Regional Innovation Systems, Edward Elgar, London.

Lorenzen, M. and Mudambi, R. (2013) Clusters, connectivity and catch-up: Bollywood and Bangalore in the global economy, Journal of Economic Geography, 13, 501-534.

Lu, R. and Cao, J. (2012) Relationships among homogenous clusters in adjoining cities, Regional Insights, 3, 12-14.

39

Lu, R., Zhang, R. and Reve, T. (2013) Relations among clusters in six Chinese city regions, European Planning Studies, 21, 1189-1209.

Maskell, P. and Lorenzen, M. (2004) The cluster as market organization, Urban Studies, 41, 991-1009.

Mattes, J. (2012) Dimensions of proximity and knowledge bases: Innovation between spatial and non-spatial factors, Regional Studies, 46, 1085-1099.

Montreal Clusters (2014) Montréal’s Metropolitan Clusters, Communauté métropolitaine de Montréal, Montreal. Available from: http://grappesmontreal.ca/metropolitan-clusters/montreals-metropolitan-clusters/?L=1 (November 29, 2014).

Moodysson, J., Coenen, L. and Asheim, B. (2009) Two sides of the same coin? Local and global knowledge flows in Medicon Valley, in F. Belussi and A. Sammarra (eds) Business Networks in Clusters and Industrial Districts: The Governance of the Global Value Chain. Routledge, New York, 356-376.

NDRC – National Development and Reform Commission (2006) NDRC Identified Beijing as National Biomedical Industry Base, Beijing (in Chinese). Available from: http://www.ndrc.gov.cn/gzdt/200610/t20061030_90634.html (November 29, 2014).

Neffke, F., Henning, M., Boschma, R., Lundquist, K.-J. and Olander, L.-O. (2011) The dynamics of agglomeration externalities along the life cycle of industries, Regional Studies, 45, 49-65.

Niosi, J. and Zhao, J. (2012) Pharmaceutical Biotechnology in Emerging Countries, presented to the 72rd International Atlantic Economic Conference, Istanbul, March 28-31.

Oliver, J. L. H., Garrigós, J. A. and Porta, J. I. D. (2008) External ties and the reduction of knowledge asymmetries among clusters within global value chains: The case of the ceramic tile district of Castellon, European Planning Studies, 16, 507-520.

Owen-Smith, J. and Powell, W.W. (2004) Knowledge networks as channels and conduits: The effects of spillovers in the Boston biotechnology community, Organization Science, 15, 5-21.

Parr, J. B. (2002): Missing elements in the analysis of agglomeration economies, International Regional Science Review, 5, 151-168.

Phene, A., Fladmoe-Lindquist, K. and Marsh, L. (2006) Breakthrough innovations in the US biotechnology industry: The effects of technological space and geographic origin, Strategic Management Journal, 27, 369-388.

Phelps, N. A. (2004) Clusters, dispersion and the spaces in between: For an economic geography of the banal, Urban Studies, 41, 971-989.

Pisano, G. P. (2006) Science Business: The Promise, the Reality, and the Future of Biotech, Harvard Business School Press, Boston.

Podolny, J. M., and Page, K. L. (1998) Network forms of organizations, Annual Review of Sociology, 24, 57-76.

Porter, M. E. (1990) The Competitive Advantage of Nations, The Free Press, New York.

Powell, W. W. (1990) Neither market nor hierarchy: Network forms of organization, Research in Organizational Behavior, 12, 295-336.

40

Rallet, A. and Torre, A. (1999) Is geographical proximity necessary in the innovation networks in the era of the global economy?, GeoJournal, 49, 373-380.

Rallet, A. and Torre, A. (2005) Proximity and localization, Regional Studies, 39, 47-59.

Saxenian, A. L. (1994) Regional Advantage, Harvard University Press, Cambridge, MA.

Scott, A. J. (1986) High tech industry and territorial development: The rise of the Orange County complex, 1955-1984. Urban Geography, 7, 3-45.

Scott, A. J. and Pope, N. E. (2007) Hollywood, Vancouver, and the world: Employment relocation and the emergence of satellite production centers in the motion-picture industry, Environment and Planning A, 39, 1364-1381.

Seawright, J. and Gerring, J. (2008) Case selection techniques in case study research: A menu of qualitative and quantitative options, Political Research Quarterly, 61, 294-308.

Sorenson, O., Rivkin, J. and Fleming, L. (2006) Complexity, networks and knowledge flow, Research Policy, 35, 994-1017.

Stuart, T. E., Ozdemir, S. Z. and Ding W. W. (2007) Vertical alliance networks: The case of university-biotechnology-pharmaceutical alliance chains, Research Policy, 36, 477-498.

Tan, J. (2006) Growth of industry clusters and innovation: Lessons from Beijing Zhongguancum Science Park, Journal of Business Venturing, 21, 827-850.

Tokatli, N. (2015) Single-firm case studies in economic geography: Some methodological reflections on the case of Zara, Journal of Economics Geography, 15, forthcoming.

Turkina, E., van Assche, A. and Kali, R. (2016) Network structure and industrial clustering dynamics in the aerospace industry, Journal of Economic Geography, 16, forthcoming.

Wade, R. (1990) Governing the Market, Princeton University Press, Princeton, NJ.

Wang, C. C. and Lin, G. C. S. (2013) Dynamics of innovation in a globalizing China: Regional environment, inter-firm relations and firm attributes, Journal of Economic Geography, 13, 397-418.

Wang, Z., Zhao, J., Liu, X., Mao, K. J. and Xie, S. L. (2006) An analysis of the spatial pattern of high-tech industries and related factors, Studies in Science of Science, 24, 227-232 (in Chinese).

Wong, J. (2004) The adaptive developmental state in East Asia, Journal of East Asian Studies, 4, 345-362.

Yin, R. K. (2003) Case Study Research: Design and Methods, 3rd edn., Sage, Thousand Oaks, CA.

Zhao, J. (2006) Industry–academia collaboration on technological innovation, Enterprise Management, 7, 20-21 (in Chinese).

Zhao, J. (2008) A study of university-industry cooperative pattern, Science and Technology Management Research, 6, 14-16 (in Chinese).

Zhao, J. (2010), Regional Innovation Mode of Industrial Clusters Based on Technology Diffusion – The Case of Beijing Pharmaceutical Industrial Parks, Unpublished postdoctoral report, Harbin Institute of Technology, Harbin (in Chinese).

41

Zhao, J. (2012) Internationalization strategies of Chinese pharmaceutical firms’, in T. G. Bas and J. Zhao (eds.), Comparing High Technology Firms in Developed and Developing Countries: Cluster Growth Initiatives, IGI Global Publisher, Hershey, PA, 204-216.

Zhao, J. and Richards, J. (2012) Beijing innovation system: The perspective of organizational structure and spatial distribution, International Journal of Learning and Intellectual Capital, 9, 413-428.

Zhao, Y.-D. and Zhang, W.-X. (2008) Clustering or agglomeration: A reflection on the development of local industrial parks, China Industrial Economics, 24(1), 131-138 (in Chinese).

Zhou, D.-S. and Dai, D.-S. (2008) A study on the promotion mechanisms of industry cluster innovation networks in biopharmaceutical industrialization. Chinese Journal of Pharmaceutical Technology Economics and Management, 2(6), 39-48 (in Chinese).

Zhou, Y. (2008) The Inside Story of China’s High-Tech Industry: Making Silicon Valley in Beijing, Rowman & Littlefield, Lanham, MD.

Zucker, L. G. and Darby, M. R. (1996) Star scientists and institutional transformation: Patterns of invention and innovation in the formation of the biotechnology industry, Proceedings of the National Academy of Sciences, 93, 12709-12716.

42

Table 1: Basic characteristics of three biomedical clusters in Beijing, 2012

Item

Cluster

Year/date established

Location in Beijing

Biomedical firms (number)

Specific competence

Main function

Yizhuang Park (Beijing Economic and Technological Development Area, BETDA)

August 25, 1994, BETDA; 2009, biophar-maceutical industry park

Yizhuang district in the Southeast

390 R&D outsourcing; biopharmaceu-tical manufac-turing

Global biopharmaceu-tical contract manufacturing projects

Daxing Park (Beijing Bioengineering and Medicinal Industry Base)

December 31, 2002

Daxing district in the South

337 Biopharmaceu-tical manufac-turing

National vaccine research and production base; state-level detection of biomedicine; drug evaluation center

ZLS Park (Zhongguancun Life Science Park)

August 18, 2000

Changping district in the North

112 Life-science/ drug R&D; R&D out-sourcing

Biopharmaceu-tical incubator

Sources: compiled during the research project

43

Figure 1: Conceptualizing multiple clusters in the same industry in one mega-city region

Scenario 1 Collaborating clusters Scenario 2 Competing clusters

Cluster/industry agglomeration

External pipelines between clusters

Global pipelines

Global city/mega-city region

Actors and firms Local information flows, gossip, news,

44

Figure 2: Firm networks surrounding Beijing Medical University United Biological

Engineering Co. located in ZLS Park and Yizhuang Park, 2014

Co-founded

Beijing Medical University United Biological Engineering Co. (BMUC)

Peking University Health Science Center (Affiliated Hospital in ZLS Park)

Beijing Biological Medicine High-Tech Incubator (ZLS Park)

Beijing Mingwude Biological Technology Co. (Yizhuang Park)

Beijing Medical University United Pharmaceutical Co. (Yizhuang Park)

Beijing ZLS Incubator Co. (ZLS Park)

Holding Co-founded

Beijing Economic and Technological Investment Development

ZLS and Management Committee of ZGC (ZLS Park)

Co-founded

45

Figure 3: Example of intra- and cross-cluster networks in Beijing’s biomedical industry,

2014

KIBS

Biomedicalfirm

Training school

PARK A

R&D firm

Incubator

PARK C

National institute

University and research unit

PARK B

Holding or co- founded firm

Administrative committee

Very close and very frequent linkage Close and frequent linkage Not close and not frequent linkage

Industrial association

46

Figure 4: Example of university and research organization as technology source and

linkage node, 2014

Joint venture

Wholly-owned

Shareholder

China Academy of Medical Sciences/Peking Union Medical College: Institute of Materia Medica

Beijing Union Pharmaceutical Factory I (Daxing Park)

Beijing Union-Genius Medical Technology Development (Yizhuang Park)

Beijing Collab Pharma (China Academy of Medical Sciences New Drug Development Test Base) (Daxing Park)

Beijing Lianxin Pharmaceutical (Daxing Park/Yizhuang Park)

Peking Union Lawke (ZLS Park)

Beijing Union Medical Center (ZLS Park)

National Institute of Biological Sciences (ZLS Park)

Beijing National Engineering Research Center of CapitalBio Corporation and Biological Chip (ZLS Park/Yizhuang Park)

Co-founded

Affiliated hospital

Technical support

Beijing Union Pharmaceutical Factory II (Daxing Park)

Overseas firms, universities & research institutes