Paper to be presented at

DRUID15, Rome, June 15-17, 2015

(Coorganized with LUISS)

An analysis of the co-inventor networks associated with the Chinese

pharmaceutical industryAlessandra Perri

Ca' Foscari University VeniceManagement

alessandra.perri@unive.it

Vittoria Giada ScaleraPolitecnico di Milano

Department of Management, Economics and Industrial Engineerivittoriagiada.scalera@polimi.it

Ram Mudambi

Fox School of Business - Temple UniversityDepartment of Strategic Management

rmudambi@temple.edu

AbstractWe study innovation processes in emerging markets, where foreign actors that have been identified as key sources ofknowledge spillovers as well as progenitors of industry clusters. We argue that in order to reconcile the location-centricimperative of innovation catch-up and the organization-centric objective of competence creation, we must widen our lensbeyond commercial firms to include the full range of innovative actors. Examining the co-inventor networks generatedby US PTO patents associated with the Chinese pharmaceutical industry, our results emphasize the critical role ofuniversities and research centers from advanced countries in connecting China to global knowledge networks.

Jelcodes:O32,M00

!"

"

An analysis of the co-inventor networks associated with the Chinese pharmaceutical

industry

Abstract

We study innovation processes in emerging markets, where foreign actors that have been identified as

key sources of knowledge spillovers as well as progenitors of industry clusters. We argue that in order

to reconcile the location-centric imperative of innovation catch-up and the organization-centric

objective of competence creation, we must widen our lens beyond commercial firms to include the full

range of innovative actors. Examining the co-inventor networks generated by US PTO patents

associated with the Chinese pharmaceutical industry, our results emphasize the critical role of

universities and research centers from advanced countries in connecting China to global knowledge

networks.

#"

"

1. Introduction

Emerging market multinational enterprises (EMNEs) have risen to occupy important positions

in a wide range of global industries (Cuervo-Cazurra and Genc, 2008; Hobday, 2010;

Kumaraswamy, et al. 2012; Lorenzen and Mudambi, 2013). A key to understanding the rapid

pace with which many of these EMNEs have achieved such significant positions on the global

stage is to distinguish between output and innovation capabilities (Bell and Pavitt, 1993).

While output capabilities depict a firm’s expertise in delivering the current generation of

products and services, innovation capabilities refer to its inherent proficiency in extending and

enriching existing technological knowledge. In this regard, recent studies have demonstrated

that emerging country firms are quick in catching up on output capabilities, but not as quick in

terms of innovative capabilities (Awate et al., 2012).

While this finding confirms that the process behind the development of output and innovation

capabilities is inherently different, it also reveals that our understanding of how innovation

catch-up happens is underdeveloped. Innovation capabilities are critical for EMNEs, as a

persistent lack of such skills would prevent these firms from fully participating in the creation

of knowledge-based intangibles that account for the bulk of all value creation in today’s

global economy (Mudambi, 2008; Corrado and Hulten, 2010). Aggregating up from the firm

level, this implies that a failure of EMNEs to develop innovation capabilities will hinder the

overall catch-up process of the economy at large (Abramovitz, 1986). Indeed, the contrast

between South Korean firms’ development of innovation capabilities and Brazilian firms

inability to do the same has been highlighted as one of the causes underlying the different

catch-up experiences of these two economies (Moreira 1995; Hannigan et al, 2013).

$"

"

The foregoing discussion emphasizes the critical importance of the development of innovation

capabilities for emerging economies. Our principal objective in this study is to address the

core research question related to this phenomenon to wit, what are the most promising

organizational conduits for emerging economy knowledge sourcing? Recent research has

analyzed this phenomenon in the context of EMNEs (Awate et al., 2015), showing that these

firms seek to access the knowledge required for innovation catch-up from their acquisitions in

advanced countries. In this study, we widen out lens beyond multinational enterprises (MNEs)

to include other actors that may play a critical role in the build-up of innovation capabilities:

domestic firms, as well as domestic and foreign universities and research centers.

A basic requisite of international knowledge sourcing is connectivity, the full range of

potential linkages between one location and all other global locations. In other words,

connectivity provides the basis for the potential recombination of ideas from diverse

locations. It occurs through the activation of a variety of global linkages that may serve as

conduits for valuable knowledge inflows (Lorenzen and Mudambi, 2013). Several studies

suggest that it may play a central role in emerging countries’ technological catch-up (Amin,

2002; Davenport, 2005; Gertler and Levitte, 2005; Martin and Sunley, 2006).

Knowledge is context-specific (Nelson & Winter, 1982) and tends to develop in co-evolution

with distinct national characteristics (Phene et al., 2006). Hence, relying only on local

resources for innovation catch-up would expose emerging countries to the risk of being

locked into their poor and low-quality knowledge base (Martin and Sunley, 2006).

Conversely, infusions of external knowledge may provide actors in these countries with the

novelty and variety that are needed to feed and enrich local innovation processes, especially if

%"

"

the knowledge sources reside in foreign countries (Malmberg and Maskell, 2002; Bathelt et

al., 2004).

Our research context in this study is the Chinese pharmaceutical industry. We use patent data

to examine the extent to which inventor networks linked to China are geographically

dispersed. We consider an inventor network to be “linked to China” either when it includes

one or more Chinese inventors, or when it refers to a patent that has been assigned to a

Chinese organization. The phenomenon of international connectivity through geographically

dispersed inventor networks is especially relevant given the decrease in communication and

transportation costs and the notable upgrades in digital technologies, which enable real time

collaborations with peers located at great distance (Saxenian, 2005; McCann, 2005). If

knowledge flows more effectively through direct interaction and personal contacts (Saxenian,

1994), collaborating with international teams should act as a channel for the acquisition of

advanced technology and knowledge creation practices, and increase the likelihood of

knowledge sharing across locations in the future (Haas & Hansen, 2005), thus ultimately

fostering the development of innovation capabilities.

Inventors’ scientific work can be coordinated by different types of innovative organizations,

referred to as “patent assignees”, which may originate from both advanced and emerging

geographic contexts. Because organizations may differ in terms of their resource endowment,

objectives and incentives depending on they geographic origin (advanced versus emerging

countries) and typology (e.g., multinational firms, universities, research centers, single-

location firms), their ability and willingness to foster the international connectivity of their

research teams can vary. In order to explore this phenomenon, this study seeks to understand

&"

"

how the geographic origin and institutional type of innovative organization affect the

international connectivity of inventor networks in the Chinese pharmaceutical industry.

To answer this question, we collected the population of pharmaceutical patents issued by the

USPTO between 1975 and 2010 and granted to both Chinese and foreign assignees utilizing

the scientific work of Chinese inventors. We analyze the geographic dispersion of the

inventor networks and classified patent assignees based on their geographic origin, as well as

on a comprehensive taxonomy of assignee types.

We believe the empirical setting of our research is appropriate for several reasons. First, the

pharmaceutical industry one of the most technology intensive sector, but simultaneously

displays a significant gap, in terms of knowledge-based activities, between advanced and

emerging countries (National Science Board, 2014), thus representing an interesting field for

exploring innovation catch-up strategies of emerging countries. Second, agents affiliated to

this industry extensively use patents as a way to protect their innovation output and

intellectual property (IP), thus making patent information a reliable and rather comprehensive

data source. Finally, due to both the increasing interest of foreign multinational firms and the

manifold reforms that have occurred in the industry in the last decades, the Chinese

pharmaceutical sector is populated by various types of actors – both domestic and foreign –

that actively participate in the innovation process (Thomson Reuters, 2010). This provides us

with the opportunity to investigate the role that different organizational, institutional and

geographic characteristics of the innovative agents may play in a country’s innovation catch-

up process.

'"

"

Our paper contributes to the literature on the technological catch up of emerging countries by

focusing on innovation catch-up, the most sophisticated component of the catch-up process. It

does so by investigating the specific channel of international co-inventor networks. Building

on previous research suggesting that internationalization is critical for the upgrade of

emerging countries (Chittoor, et al., 2009), we focus on the internationalization of inventors’

collaborative activities. Specifically, we analyze the geographic reach of collaboration

behaviors in emerging countries, under different organizational arrangements. We also add to

the research stream on the networking behavior of inventors (Balconi et al. 2004), by

simultaneously accounting for the role of the geographic origin and institutional type of

innovative actors in a previously unexplored economic context, i.e. the emerging country

context. We discuss both institutional and managerial implications of our study.

2. Conceptual framework and research

2.1 International knowledge sourcing through global linkages and connectivity

Firms attempt to develop their knowledge internally, but often they also leverage external

sources using formal and informal arrangements. Internal R&D and external knowledge

sourcing are often complementary parts of an organization’s innovation strategy (Cassiman

and Veugelers, 2006). External knowledge sourcing occurs through numerous mans,

including direct channels like acquisitions and alliances, as well as indirect ones like

spillovers, personal relationships, scientist mobility in the labor market, buyer-supplier

relationships, and other means (Chung and Yeaple, 2008; Laursen and Salter, 2004).

("

"

Considering that innovation activities and knowledge resources differ across countries, firms

can increase their knowledge base by sourcing technical capabilities internationally (Cantwell,

1989; Kuemmerle 1999). The disaggregation of global value chains has accelerated

geographically dispersed knowledge sourcing as organizations increasingly use their foreign

subsidiaries to tap into global centers of excellence. The orchestration of fine sliced value

chains by orchestrating organizations has played a key role in the creation of global linkages

between firms and individuals located in both advanced and emerging economies (Jensen and

Pedersen, 2011; Mudambi, 2008; Mudambi and Venzin, 2010). As value chains increasingly

span national borders, national systems of innovation have become interconnected in global

innovation networks (Narula and Guimòn, 2010).

Global value chains have not only created and intensified global linkages, but they have also

changed the configuration of the social networks. Traditionally global innovation networks

were concentrated in advanced market economies with relatively low levels of geographical

dispersion. However, both these characteristics of global innovation networks are rapidly

changing. MNEs based in emerging economies are increasingly entering global innovation

networks, so the dominance of advanced market economies is declining. Further, the extent of

global innovation networks’ geographical dispersion is rising, driven by two processes. First,

MNEs based in advanced market economies are offshoring knowledge creation to subsidiaries

in emerging market economies in order to leverage the high quality, low cost resources

available there (D’Agostino et al., 2013; Govindarajan and Ramamurti, 2011).. Second,

EMNEs are implementing “catch-up” processes by strategically acquiring knowledge assets

in advanced market economies (Awate et al., 2012; Athreye and Kapur, 2009, Deng, 2009; Li

)"

"

et al., 2012). Both of these processes lead to geographically dispersed innovation networks

spanning advanced and emerging economies.

In this context, global linkages can be defined as channels that allow for the efficient

transmission of different types of resources from geographically dispersed locations.

Lorenzen and Mudambi (2013) distinguish between two forms of global linkages: pipelines,

which are “organization-based linkages” (Bathelt et al., 2004) and personal relationships that

are often leveraged through global diasporas. This latter form of global linkage is particularly

important for the diffusion of knowledge, since extant literature demonstrates that the

complexity of knowledge acquisition and transfer can be overcome through personal

interactions between those who are willing to learn and those who have generated the

knowledge to be transmitted (Breschi and Lissoni, 2001).

The concept of connectivity is defined by integrating the nature of global linkages with their

network structure. It has been proposed that the connectivity that arises a decentralized

network structure is most conducive to knowledge creation and innovation in emerging

economies. Since a network structure is characterized by the lack of a strong “gatekeeper”

controlling access into and out of emerging economy locations so that local inventors and

scientists can interact and share knowledge. International connectivity facilitates external

knowledge infusions that can nourish (local) internal innovative activities, especially because

it encourages the recombination of knowledge from different sources and countries (Bathelt et

al., 2004).

*"

"

2.2 Emerging economies and technological catch-up

Firms from emerging economies often lack the competencies and knowledge to successfully

compete with their counterparts from advanced economies (Awate et al., 2015; Luo and Tung,

2007). EMNEs tend to lag behind in terms of technological expertise (Kumar and Russell,

2002), and therefore cannot rely only on their own resources to reduce the gap. Although the

literature has documented the ongoing process of technological upgrading and the relatively

large pools of high skilled human capital in emerging economies (Athreye and Cantwell,

2007; Lewin at al., 2009), progress seems to be slow. Moreover, it is primarily focalized on

output capabilities (Awate et al., 2012).

Extant literature has shown that foreign direct investment (FDI) from advanced economies

often generates knowledge spillovers (Blomstrom and Kokko, 1998). More importantly, it is

the genesis of spillover processes that work through the joint activities of MNE subsidiaries

and local firms (Mudambi, 2008) and these processes are typically the spark that ignites

wide-ranging technological catch-up (Li et al. 2012, Wei and Liu, 2006). Personal

relationships and the mobility of skilled workers act in concert with inter-organizational ties

as knowledge channels that operationalize spillovers (Filatotchev et al., 2011). Given that

knowledge flows more effectively through direct interaction and personal contacts (Saxenian,

1994), emerging country inventors collaborating with international teams should act as an

efficient channel for the acquisition of advanced technology and knowledge creation

practices. This should ultimately promote the development of superior innovative capabilities.

This discussion suggests that international connectivity has a critical role in fostering the

technological catch-up process of emerging economies. It activates global linkages that

!+"

"

conduct knowledge inflows from worldwide sources (Amin, 2002; Davenport, 2005; Gertler

and Levitte, 2005), offsetting distances and allowing inventors from emerging countries to

learn and ultimately to implement catch-up processes. It is important to recognize that

connectivity is bi-directional, and generates higher awareness and mutual interdependence.

In the context of emerging economies, it appears that there are two drivers of international

connectivity: institutions from emerging markets “reaching out”, and foreign organizations

from advanced economies “reaching in”. The major modalities for reaching out are

technology licensing and joint ventures, followed by cross-border acquisitions. The goals of

these knowledge-seeking acquisitions are both R&D knowledge (technology and know-how)

and marketing knowledge (brand building and global legitimacy) (Buckley at al., 2007; Luo

and Tung, 2007). R&D-based FDI in particular helps emerging countries organizations to

develop collaborations with foreign inventors, thus accessing diverse pools of knowledge and

cutting-edge technologies.

The second driver of connectivity is embodied in foreign-based institutions that reach into

emerging economies. Typically, this happens through the offshoring of manufacturing

activities, production processes, and services (Contractor et al., 2010). Asian countries, such

as China, offer a substantial pool of qualified workers and expertise at a competitive cost, to

which firms from other countries are increasingly willing to access (Lewin et al., 2009). As a

result, emerging economies are the preferred offshoring location for R&D activities

(D’Agostino et al., 2013). Thus, the presence of advanced institutions in emerging economies

promotes the involvement of local inventors in their knowledge networks, facilitating the

chances of cooperation and transfer of ideas and know-how (Lorenzen and Mudambi, 2013).

!!"

"

While both emerging countries institutions reaching out and advanced market organizations

reaching in generate connectivity, we claim that there is a systematic difference between these

two sets of actors in their ability to create effective connectivity, and so to spawn

geographically dispersed inventor networks. In particular, we posit that emerging economies

innovative institutions more likely generate less internationally dispersed networks, compared

to their advanced economy counterparts.

Geographic origin, in fact, is crucial in determining access to knowledge, resources and

networks (Bartholomew, 1997; Phene et al., 2006; Porter, 1990). Advanced economy

innovative institutions are traditionally recognized as the engine of innovation and have been

the major players in the production of cutting-edge R&D for a long period of time (Cantwell,

1989). The leading innovators from advanced economies possess complex and established

organizational capabilities that they leverage to orchestrate geographically dispersed activities

effectively (Cantwell and Mudambi, 2011; Tallman and Chacar, 2011). Such firms are able to

transmit and integrate even the most tacit knowledge across borders and over long distances

(Cantwell and Santangelo, 1999). Conversely, emerging economy organizations have only

recently started to invest in innovation and the globalization of their R&D is nascent

(UNCTAD, 2005). In contrast to their counterparts from advanced economies, emerging

economies innovative institutions have been positioned in the peripheries of international

innovative networks, playing supporting role in the exchange and creation of knowledge at

international level (Awate et al., 2012).

Hence emerging economy institutions are likely to face higher barriers when attempting to

connect to inventors or institutions from advanced markets (Madhok and Keyhani, 2012). In

!#"

"

spite of their increasing international openness, cultural and institutional distances may hinder

the ability of emerging market organizations in developing collaborations with foreign

inventors and thereby limit the international connectivity of their networks. Further, emerging

markets innovative institutions may not possess enough international experience to manage

the complexity associated with international knowledge networks, which embrace a range of

different sources of heterogeneity (Hansen, 2002). Further, they typically lack a knowledge

base that is strong enough to be leveraged in the support of effective R&D collaborations with

more skilled partners (Deng, 2009; Lane et al., 2001). In fact, they may not even possess

sufficient know-how to successfully interact with their peers and to orchestrate dispersed and

heterogeneous networks.

We therefore expect that:

Hypothesis 1 (H1): In emerging economies, domestic innovating organizations (and

those based in other emerging economies) spawn less internationally dispersed

networks than innovating organizations based in advanced economies.

2.3 The role of university and commercial knowledge pipelines

The geographic origin of innovative actors is not the only variable that may influence the

international dispersion of inventor networks. Organizations involved in innovative activities

are highly heterogeneous in terms of their institutional types. Since different types of

organizations differ systematically in terms of their objectives, leading to differences in their

patterns of knowledge sourcing and knowledge creation. In order to explore this issue, we

!$"

"

distinguish between (foreign) MNEs, (domestic) single-location firms and universities and

research centers, and elaborate on their ability to drive connectivity. More specifically,

assuming single-location firms as the benchmark to which comparing the other institutional

types, we develop hypotheses on the role of MNEs and universities and research centers.

Single-location firms have limited opportunities in terms of resource access. In fact, as

suggested by existing literature, they tend to rely almost exclusively on their local cluster for

linkages creation, thus being isolated from international networks (Henderson, 2003).

Conversely, MNEs have the potential to access to two different knowledge networks

(Almeida and Phene, 2004), both of which are geographically dispersed. First, MNEs by

definition are organized as networks of subsidiaries established worldwide. Hence, they can

exploit firm-internal networks and develop substantial internal linkages (Alcacer and Zhao,

2012; Meyer et al., 2011). Second, by embedding in their localities (Andersson et al., 2002),

foreign subsidiaries develop external linkages that grant access to pools of localized

knowledge and resources in different host-regions (Almeida and Phene, 2004).

As far as universities and research centers are concerned, previous literature has demonstrated

that inventors working for this type of institutions are better in “connecting individuals and

network components” (Balconi et al. 2004, p. 144) compared to non-academic inventors.

Universities and research centers act as the sources of basic knowledge for industrial scientists

(Cohen et al., 2000) but, unlike the latter, they are characterized by an “open” approach to

science and technology (Balconi et al. 2004). While industrial innovators have a strong

incentive to protect the outcomes of their innovative activities as these represent a source of

!%"

"

rents, scientists operating in universities and research centers are typically not interested in the

commercialization of ideas, as this falls beyond the scope of their activity.

In other words, academic inventors are driven by “primacy” while commercial inventors are

driven by “secrecy” (Mudambi and Swift, 2009). Rather, they pursue research with the goal of

advancing the knowledge frontier, often driven by their individual motivation. In addition, the

social and professional environment to which they belong stimulates their willingness to share

the results of their innovative processes, as this increases their personal reputation (Siegel et

al., 2003). The dissemination of the research results is in fact a central component of

universities’ and research centers’ scientific activity (Fabrizio, 2007), which for this reason is

often referred to as “public science”. In sum, contrary to what happens in commercial firms,

inventors working for universities and research centers retain strong incentives to collaborate

over long geographical distances as they seek linkages with experts and the broadest possible

diffusion of their ideas. It follows that the community of scientists tends to be highly

connected in spite of geographic distance, which stimulates the collaboration among inventors

located worldwide.

We therefore expect that:

Hypothesis 2 (H2). In emerging economies, both MNEs and universities and research

centers spawn more internationally dispersed inventor networks compared to

(domestic) single-location firms.

In an advanced country context, one could argue on which institutional type, among MNEs

and universities and research centers, is able to drive the most internationally dispersed

!&"

"

inventor networks. The innovative activities of MNEs in advanced economies tend to be

central and linked to their widely dispersed subsidiary networks. Conversely, MNE

innovation occurring in emerging countries is usually peripheral since it is rooted in a local

context where the industrial knowledge base is backward and the institutional infrastructure,

including the intellectual property protection system, is rather poor. As MNEs are strongly

committed to protect their proprietary knowledge from external appropriation (Mariotti,

Piscitello, & Elia, 2010; Perri and Andersson, 2014), they can be expected to have lower

incentives to develop internationally dispersed inventor teams in such locations. Giving a

central role to innovation teams in these locations could decrease the control over their assets

while granting access to low-quality knowledge.

In contrast, as we have argued, universities and research centers are less sensitive to

knowledge protection imperatives (Balconi et al., 2004), and are thus less concerned about

threats arising from weak intellectual property rights protection. Combining these arguments,

we suggest that, in emerging country contexts, universities and research centers generate

higher international connectivity than MNEs do:

Hypothesis 2a (H2a). In emerging economies, inventor networks coordinated by

universities and research centers are more internationally dispersed than inventor

networks coordinated by MNEs.

2.4 The impact of geographic origin on the role of universities and firms pipelines

In order to fully appreciate the impact of the geographic origin of innovative organizations

and their institutional type, it is important to consider these factors jointly. In fact, the effects

!'"

"

predicted in H1a and H2a may be sensitive to whether the innovative organization

coordinating the scientific work of the inventor team is based in an advanced or an emerging

country.

As far emerging country MNEs and universities are concerned, the literature suggests that

they can rely on the return migration of skilled graduate students and engineers educated in

advanced countries and returned to their home economies endowed with critical professional

linkages to peers in their former host-countries (Saxenian, 2005; Jonkers and Tijssen, 2008).

While this is an important channel for enabling these actors to connect to international

networks, their counterparts from advanced countries have many other critical advantages that

may place them in privileged position to spawn internationally dispersed inventor teams.

On the one hand, in spite of the idea of the academic community as a small world

characterized by high interconnectedness, not all actors belonging to this world are likely to

be equally central or to share the same privileged position within the network (Newman,

2000; 2001; Fleming and Marx, 2006). Compared to their foreign peers, universities and

research centers from emerging countries are likely to be marginalized, peripheral

components of the scientific community, less able to connect to the global academic network.

On the other hand, compared to advanced country MNEs, those originating from emerging

countries are likely to have a less geographically dispersed network of foreign affiliates and

their internal linkages may be hampered by the opportunism of acquired subsidiaries (Awate

et al., 2015). Moreover, EMNEs tend to be endowed with a narrower capability base

(Mathews, 2006), which decreases their ability to connect to the rest of the world. The relative

backwardness and peripheral position of their locality may also play a role in reducing the

!("

"

opportunities for the creation of knowledge linkages with partners from more technologically

advanced regions.

Compared to their advanced counterparts, they should therefore drive a lower degree of

connectivity. Based on this reasoning, we expect that:

Hypothesis 3a (H3a). In emerging economies, the higher connectivity of universities

and research centers is less accentuated when they originate from emerging countries.

Hypothesis 3b (H3b). In emerging economies, the higher connectivity of MNEs is less

accentuated when they originate from emerging countries.

3. Empirical analysis

3.1 Empirical setting

Traditionally regarded as a highly profitable context (Ghemawat, 2010), the global

pharmaceutical sector has experienced a number of major changes in the last decades, which

have strongly modified the industry’s competitive dynamics leading to a gradual shrinking of

profit opportunities (Scalera et al., 2015). After the discovery of important “blockbuster”

drugs which have fed the big pharmaceutical companies for several decades, the odds of

coming up with new high-potential molecules decrease over time, as those “easy targets are

being steadily exhausted” (Bruche, 2012; p.5). Major institutional changes also contributed to

modify the industry’s structure and competitive approaches, inducing the rise of a generic

!)"

"

market segment and, accordingly, of an increasing number of imitative manufacturers

(Bruche, 2012). Finally, starting from the early 1980s, the successful emergence of

biotechnology and genetics demanded incumbent firms to develop new skills in terms of

rational drug design, as opposed to traditional trial and error discovery processes, thus making

the exploitation of external technology sources compelling (Cockburn, 2004).

Faced with these competitive challenges, big pharmaceutical companies had to significantly

amend their business model over time. One opportunity for managing these challenges arise

from emerging countries, whose enormous populations, growing awareness of the importance

of healthcare, and increasing GDP have readily attracted companies by then used to deal with

mature and stagnant markets. Originally regarded exclusively as final markets where Western

pharmaceutical companies could manufacture and sell their products, emerging countries have

progressively become the target of knowledge intensive FDI, hosting an increasing number of

foreign MNEs’ R&D facilities (Scalera et al., 2015). Among these locations, emerging

markets like India and China take the lead. In these geographical contexts, the pharmaceutical

industry tends to be a turbulent and discontinuous environment. In fact, the effects of the

economic liberalization process and IP reforms combine with the specificities of a highly

regulated industrial setting. Internal heterogeneity is also a major issue in these countries, as

there are profound gaps in terms of access to healthcare services among rural and more central

geographic areas.

We choose the Chinese pharmaceutical industry as the empirical setting of our study, as China

represents an ideal test-bed for our hypotheses. China is one of the largest pharmaceutical

markets in the world, but its prominence arises from the size of its population, rather than

!*"

"

from the sophistication of its offer. Although in the last decades it has experienced a reform of

the healthcare system and a gradual transformation of local pharmaceutical industry, which is

increasingly populated by research-based companies, the market is still highly fragmented,

and characterized by a complex system of sub-national segments dominated by small to

medium-sized generics and over the counter drugs (OTC) manufacturers. Moreover, in spite

of China’ adhesion to the in the World Trade Organization (WTO) in 2001, the country is still

regarded as an unsafe context for intellectual property rights protection (Zhao, 2006).

3.2 Data

In order to study innovative activities in the Chinese pharmaceutical industry, we used patent

data as a proxy for innovative output. Accordingly to other studies about innovative activities

in China (e.g. Branstetter et al., 2013; Scalera et al., 2015; Zhao, 2006), we focus on the

United States Patent and Trademark Office (USPTO) data, in order to assure the originality

and high quality of the innovations analyzed (Archibugi and Coco, 2005)1.

Extant literature has already explored collaboration patterns of inventors employing patent co-

inventorship (e.g. Ejermo & Karlsson, 2006). In fact, patents provide detailed information on

the team of inventors, favoring the identification of collaborations among inventors and their

geographical distribution. In addition, patents represent a more refined measure of the

efficient utilization of human capital, compared to input measures. Finally, they also give the

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""1 We decided to not consider data from the State Intellectual Property Office of the Public Republic of China (SIPO) for an

array of reasons. First, the Chinese patent system has been frequently changed over time, and so Chinese data may be

inconsistent in our time horizon. Second, it is common practice to use data from a single patent office in order to avoid

problems related to lack of consistent quality across national patent systems (Jaffe and Trajtenberg, 2002).

#+"

"

possibility to account only for innovations considered valuable by the innovative

organizations that make the application to the patent office (Griliches, 2000; Zhao, 2006).

In order to build our sample, we selected all USPTO patents granted between 1975 and 2010,

that report at least one Chinese inventor or applied for by a Chinese organization. From the

initial sample, we only included patents representative of the pharmaceutical industry, so

referring to the Drug and Medical technological fields defined by Hall et al. (2001)2. We also

included design patents containing the technological class “Pharmaceutical Devices” (D24).

Finally, we exclude patents assigned to individuals, or unassigned. The sample thus generated

consists of 1026 patents, emanating from 516 different assignee organizations.

In order to ensure the validity of our analysis, it is crucial that we identify each individual

inventor and determine their address. To do so, we complemented our patent data gathered

directly from USPTO website using the “Disambiguation and co-authorship networks of the

U.S. patent inventor database (1975 - 2010)” distributed by The Harvard Dataverse Network

(see Li et al., 2014). In particular, we used this database to source the information concerning

the inventors, since the database focuses on USPTO patent data and disambiguates each

patent inventor, providing – among other information – their addressees, together with cities’

ZIP codes and latitude and longitude3.

3.3 Variables and model

3.3.1 Dependent variable: Geographical dispersion

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""2 The Drug and Medical category as defined by Hall et al. (2001) includes four sub-categories: Drugs (sub-category code 31);

Surgery and Medical Instruments (32); Biotechnology (33); and Miscellaneous – Drugs and Medicine (39). 3 As “Disambiguation and co-authorship networks of the U.S. patent inventor database (1975 - 2010)” does not include all

the USPTO design patents, we manually checked the address information of the inventors of the design patents not found in

the database. "

#!"

"

To measure the degree of connectedness of the innovative actors we employ the geographical

dispersion of their inventors’ team, using the UPSTO and “Disambiguation and co-authorship

networks of the U.S. patent inventor database (1975 - 2010)” location data of the address of

patent inventors. In fact, often an invention underlying a patent assigned to a company or a

university may be the result of R&D activities distributed in different locations. The focus on

inventors’ location enables us to determine the actual geographic origin of the knowledge4

(Lahiri, 2010).

Thus, the dependent variable, Geographical dispersion, is the geographical dispersion of the

network of inventors measured at patent level, following the approach of Cano-Kollmann and

colleagues (2014). The construction of Geographical dispersion is based on the Herfindahl

index, also known as Herfindahl–Hirschman Index, which is commonly used in industrial

organization to measure of concentration of an industry (e.g., Tallman and Li, 1996). Since

we are interested in the dispersion (and not in the concentration) of the inventor network at

patent level, the Geographical dispersion i for patent i is constructed as follows:

!"#$%&'!!"#$!!"#$%&#"'(! ! !!! !!"#!!!!!"#!!!

!

!!!

where !"#!!! is the number of inventors of patent i located in country n (N is the total number

of inventors’ locations mentioned in patent i), !"#! is the total number of inventors of patent i.

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""4 The locations are identified at country level.

##"

"

As a result, we obtained a censored dependent variable, which takes the minimum value of 1

when all inventors are located in the same country (i.e. China in our analysis), and an upper

limit asymptotically approaching 1 as the inventors network is more dispersed across different

countries.

We employ a multiple regression approach to test our hypotheses. Given that our dependent

variable is censored, taking a minimum value of 0 and an upper limit asymptotically

approaching 1 (the maximum value is 0.82), we adopted a robust Tobit regression model,

which allows controlling for heteroskedasticity of the sample (Greene, 2000: 905-926).

The baseline model to be tested is the following:

Geographical dispersioni =ƒ(Emergingi ; Universityi ; MNEi ; controlsi)

where i= 1, … , 1026 patents.

To test our HP3a (HP3b) we add to the baseline model the interaction effect between

Emerging and University (MNE), as follows:

Geographical dispersioni =ƒ(Emergingi ; Universityi ; MNEi ; Universityi * Emergingi ;

controlsi)

Geographical dispersioni =ƒ(Emergingi ; Universityi ; MNEi ; MNEi * Emergingi;

controlsi)

3.3.2 Explanatory variables

#$"

"

3.3.2.1 Emerging economies innovative institutions

In order to test our first hypothesis, we introduced the independent variable Emerging, which

is a dummy variable equal to 1 if the assignee is located in an emerging country (mostly in

China), and 0 otherwise. For the purposes of the current work, we have relied on the World

Bank classification of emerging countries, which is based on the level of per capita income.

Consequently, all countries belonging to the lower and upper middle-income groups have

been classified as emerging (http://data.worldbank.org/about/country-and-lending-groups).

If the assignee is an MNE’s foreign subsidiary, we built the variable using the location of its

headquarter (Almeida and Phene, 2004; Phene and Almeida, 2008). We used BvD Orbis to

identify the locations of the MNE headquarter.

Further, our sample includes 108 (10.53%) co-assigned patents, i.e. patents that have more

than one assignee. In these cases, the variable Emerging takes the value of 1, if at least one of

the assignee is from an emerging country.

3.3.2.2 Typology of innovative institutions

In order to categorize the organization contributing to the innovative activities in emerging

economies, we distinguished between (1) universities and research centers, (2) MNEs and (3)

single-location firms. We have carried out a thorough work of cleaning and standardizing

assignees’ names and addresses. Assignees have been identified first by assigning a unique

code to all assignees with the same name and address. Then, using BvD Orbis, we

consolidated the code to assignees with the same address and with very similar names, when

#%"

"

inconsistences arise from misspelling, presence/absence of extensions, or presence/absence of

spaces between parts of the names.

For each assignee mentioned in the patent document, and univocally identified, we analyzed

first the institutional typology and then, in the case of commercial firms, the ownership

structure, manually inspecting the assignees’ name and relying on information from BvD

Orbis, companies’ websites and other on-line resources, i.e. Bloomberg website. As regards

the commercial firms, we defined as MNE any firm that has at least one foreign subsidiary by

looking at the family tree for every firm in our sample. Because patents may be assigned

either to the MNE parent company or to one of its subsidiaries for unobservable reasons

(Cantwell and Mudambi, 2005), we considered each multiunit firm as an integrated strategic

agent, following the approach of Zhao (2006).

The categorization of the assignee type is time variant, since we checked the status of each

assignee in correspondence to the year of the patent application."This procedure enables us to

take into account changes in the firm ownership structure (e.g. merge and acquisitions), which

are very frequent especially in the pharmaceutical industry.

After the assignees’ categorization, for each patent we created three dummy variables:

University, if the patent’s assignee is a university or a research center, MNE, in case the patent

has been assigned to an MNE or one of its subsidiaries, and Single_location, otherwise. For

the analysis, we used Single location as the benchmark. Finally, in case of co-assigned

patents, we take into consideration the categories of all the co-assignees. For instance if a

patent has been assigned to a university and an MNE, both University and MNE take the value

of 1.

#&"

"

3.3.2.3 Controls

Assignee innovative leadership: The dummy variable Leader takes the value of 1 for

assignees that are in the upper quartile (or 75th percentile) of the pharmaceutical patent pool

in terms of patent production in the year previous the patent application (t-1). To determine

the pharmaceutical patent pool we considered all UPSTO patents granted in Drug and

Medical technological fields defined by Hall et al. (2001). We measured patent production as

the natural logarithm of the cumulative number of USPTO pharmaceutical patents issued by

each assignee in the period 1975 – (t-1). Data come from “Disambiguation and co-authorship

networks of the U.S. patent inventor database (1975 - 2010)” (Li et al., 2014). If the company

is part of a group or is the subsidiary of an MNE, we used the pharmaceutical patent stock of

its global ultimate owner to calculate the variable. In case of co-assigned patent, Leader takes

the value of 1, if at least one of the co-assignees is in the upper quartile.

Number of inventors of the inventor team: Team size represents the number of inventors for

each patent.

Design patent: the variable Design is a dummy and takes the value of 1, if the patent is

classified by the USPTO as a design patent and 0 in case it is a utility patent. Relying on the

USPTO definition, “[…] “utility patent” protects the way an article is used and works, while

a “design patent” protects the way an article looks. The ornamental appearance for an

article includes its shape/configuration or surface ornamentation applied to the article, or

both” (http://www.uspto.gov/).

#'"

"

Primary pharmaceutical technological class: Pharma is a dummy variable equal to 1, if the

first technological class of the focal patent is included in the pharmaceutical category, as

defined in section 3.2; otherwise it takes the value of 0.

Technological composition of the patent: The variable Tech composition is built adapting to

the Cubbin-Leech index (Cubbin and Leech, 1983) to the case of the patents’ technological

composition5. First, we computed the Herfindal index of the patent technological

concentration (H tech), using the three digit technological classes to which the USPTO has

assigned the patent:

!!!"#!! ! ! !!"#!!!"#$$!!!!!

!

!!!

where Tech composition,m is the percentage of the technological class m represented in patent

i on the total number of technological classes mentioned in patent i, i.e. M. Tech composition

is defined as follows:

!"#!!!"#$"%&'&"(! ! !!!!"#!!!"#$$!!!!!!!!!"#!! ! !"#!!!"#$$!!!!!!!!!

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""5 For a different approach measuring the ownership concentration shares in a firm, see Mudambi and Nicosia

(1998).

#("

"

where F[.] is the standard normal distribution function and !"#!!!"#$$!!! is the percentage of

the technological class most representative in patent i6.

Co-assigned patent: Co-assigned is a dummy variable that takes the value of 1 if the patent

has more than one assignee, and 0 otherwise.

IP policy changes: Since we pool patent data over a 35-year period characterized by

regulatory turbulence in Chinese IP regime, we include a year dummy that is equal to 1 if the

patent has been applied for from 2005 onwards. Since 2005, in fact, the Chinese government

fully has fully complied with the requirements of the TRIPS agreement.

4. Findings

Table 2 shows the estimated coefficients of the Tobit regressions applied to the equation

models described above.

[Insert Table 2 about here]

All models produced statistically significant results (LR chi2=283.05 and p<.0 in Model 1, LR

chi2=726.35 and p<.0 in Model 2, LR chi

2=744.09 and p<.0 in Model 3, LR chi

2=747.29 and

p<.0 in Model 4, LR chi2=744.56 and p<.0 in Model 5).

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""6 For patent with only one technological class, so with highest level of technological concentration, we proxy the

limit case for which it is possible to calculate a compute value of Tech composition, i.e, !"#!!!"#$$!!! ! !"#!

#)"

"

We employed Model 1 as the baseline that includes all our controls. As expected, the

technological leadership of innovative organizations (Leader) has a positive (0.446) and

significant effect (p<.001) on inventor teams’ geographic dispersion, as technologically

advanced actors are endowed with appropriate knowledge and relational resources to develop

global linkages. Moreover, larger inventor teams are also more likely to encompass a higher

geographic dispersion, as highlighted by the positive (0.021) and significant (p<.001)

coefficient of the Team size control.

In order to test our HP1, we ran Model 2 and found confirmation of our first hypothesis. As

predicted, the dummy variable identifying emerging economies innovative institutions

(Emerging) exhibits a negative and significant coefficient (p<.001 also in Model 3, 4 and 5),

thus showing that local innovative actors spawn less internationally dispersed inventor

networks compared to advanced economy innovative actors.

In order to test our second set of hypotheses, we employed Model 3 and added the variables

for universities and research centers and multinational firms. The University dummy variable

shows positive and significant coefficient (p<.001 in Model 3, 4 and 5), confirming the

central role of this type of innovative institution in the technological catch-up process of

emerging economies. The MNE dummy variable is also positive, but much less significant

across the different specifications (p<.05 in Model 3, p<.05 in Model 4, p<.1 in Model 5).

Overall, these results suggest that, compared to single-location firms, both universities and

research centers and MNEs establish more internationally dispersed investor networks, thus

offering support for HP2. Moreover, further analysis (Table 3) rejects the hypothesis of

#*"

"

equality of the University and MNE coefficients, confirming our HP2a and providing evidence

of the higher connectivity associated with University patents.

[Insert Table 3 about here]

To test HP3a and HP3b, Model 4 and 5 include, respectively, the interaction terms that reflect

our theoretical argumentations, i.e. University*Emerging and MNE*Emerging. In particular,

in Model 4 the coefficient of the interaction between University and Emerging (emerging

innovative institutions) is negative and statistically significant, although at a low level (p<.1).

This lends some support to HP3a. On the other hand, in Model 5, the interaction between

MNE and Emerging (emerging innovative institutions) turns out to be non significant. Hence,

HP3b is not supported.

To further explore the effects of geographic origin on the institutional type of innovative

organizations, we split our original sample between patents developed by emerging

economies institutions and patents developed by advanced economies institutions, and run

regressions including both the University and MNE dummy variables (results are available

upon request). The analyses performed on the two subsamples seem to confirm our

expectations that both universities and research centers and MNEs are capable of driving

international connectivity when they originate from advanced countries. In fact, the

coefficients of the University (p<0.001) and MNE (p<0.05) variables are significant in the

$+"

"

advanced economy subsample. Conversely, they turn out to be non-significant in the

subsample of patents belonging to emerging economy organizations.

5. Discussion and conclusions

Our results show that" innovative organizations from emerging countries spawn less

internationally dispersed inventor networks than innovative organizations from advanced

countries. This suggests that, in spite of the economic growth that many emerging economies

have achieved in the last years, their process of technological catch-up still hinges upon

innovative actors from the “advanced world”, at least when it comes to the channel of

international connectivity. Hence, these contexts are not independent as regards the

development of innovation capabilities.

Moreover, contrary to the expectations that ascribe to MNEs a central role in the

technological development of emerging countries, in the context of our analysis, it is the

institutional category of universities and research centers to act as the most effective pipeline

to connect China to global knowledge networks. This finding confirms the critical role

universities play as growth engines (Charles, 2003; Cooke, 2001; Dasgupta and David, 1994;

Kitagawa, 2004; Lundvall, 1992; Nelson, 1993, 2004; Salter and Martin, 2001). Our result

pointing to the major contribution universities and research centers can offer to the

development of innovation capabilities in emerging countries is also consistent with previous

research suggesting that these actors are seldom able to impact directly on commercial firms’

creation of new products and services (Pavitt, 2001). In fact, while firms are successful on this

$!"

"

latter dimension, given their strong motivation toward the accumulation of financial wealth,

universities and research centers are further from markets and, due to their focus on

fundamental investigation, provide a greater contribution to the development of more

sophisticated, long-term innovation capabilities, by offering a better understanding of

underlying phenomena (Fleming and Sorenson, 2004; Fabrizio, 2007). Not surprisingly,

previous research has shown that only firms that adopt very “open” search strategies and

invest in R&D draw upon university knowledge for their innovative activities (Laursen and

Salter, 2004).

Overall, our results offer renewed support for the statement according to which “at best,

foreign investment from the core might contribute to the incremental mastery of

manufacturing techniques and upgrading of local suppliers. Even the most successful newly

industrializing countries are destined to remain imitators as long as leading-edge skill and

technology reside in the corporate research labs and universities in the core.” (Saxenian,

2005; p. 38).

Of course, it should be noted that universities and research centers play a primary role in

fostering internationally dispersed inventor networks mainly when they originate from

advanced countries, suggesting that the periphery’s scientific environment has not yet

developed the required skills to continually connect with the core. This finding is also

consistent with research suggesting that, among the different emerging countries, China has

been the least successful in fostering the return migration of its skilled scientist and

knowledge workers (Jonkers and Tijssen, 2008), likely due to an ineffective return incentives

policy (Cao, 1996).

$#"

"

These results have some implications for both local and global policy-makers aiming to

expedite the process of technological catch-up of emerging countries, who should (1)

continuously engage innovative actors from advanced countries, and (2) design incentive

policies involving universities and research centers, along with FDI attraction strategies.

Moreover, our results have some bearings also for emerging country firms. First, they should

be aware that connecting to global knowledge networks is a very complex activity for

organizations that are not familiar with the global innovation system. Second, they should

consider that in order to gain access to global knowledge, they may choose between two

alternative strategies: the pursuit of internationally distributed inventor networks, or the

leverage of global cities’ inventors. The latter strategy may result easier, as it does not require

geographic dispersion, but rather focuses on “selected” locations.

$$"

"

References

Abramovitz, M. (1986), ‘Catching up, forging ahead and falling behind,’ Journal of Economic

History, 46, 385-406.

Almeida, P. and A. Phene (2004), ‘Subsidiaries and knowledge creation: the influence of the

MNC and host country on innovation,’ Strategic Management Journal, 25, 847-864.

Amin, A. (2002), ‘Spatialities of globalization,’ Environment and Planning A, 34, 385–399.

Archibugi, D. and A. Coco (2005), ‘Measuring technological capabilities at the country level:

a survey and a menu for choice,’ Research Policy, 34, 175-194.

Athreye, S. and J. Cantwell (2007),‘Creating competition?: globalisation and the emergence

of new technology producers,’ Research Policy, 36(2), 209-226.

Athreye, S. and S. Kapur (2009), ‘Introduction: the internationalization of Chinese and Indian

firms - trends, motivations and strategy,’ Industrial and Corporate Change, doi:

10.1093/icc/dtp007.

Awate, S., M. M. Larsen and R. Mudambi (2012), ‘EMNE catch-up strategies in the wind

turbine industry: is there a trade-off between output and innovation capabilities?’

Global Strategy Journal, 2, 205-223.

Awate, S., M. M. Larsen and R. Mudambi (2015), ’Accessing vs. Sourcing knowledge: a

comparative study of R&D internationalization between emerging and advanced

economy firms,’ Journal of International Business Studies, 46, 63-86.

$%"

"

Balconi, M., S. Breschi and F. Lissoni (2004), ‘Networks of inventors and the role of

academia: an exploration of Italian patent data,’ Research Policy, 33, 127-145.

Bathelt, H., A. Malmberg and P. Maskell (2004), ‘Clusters and knowledge: local buzz, global

pipelines and the process of knowledge creation,’ Progress in Human Geography, 28,

31-56.

Bartholomew, S. (1997), ‘National systems of biotechnology innovation: complex

interdependence in the global system,’ Journal of International Business Studies, 28,

241-266.

Bell, M. and K. Pavitt (1993), ‘Technological accumulation and industrial growth: contrasts

between developed and developing countries,’ Industrial and Corporate Change, 2,

157–210.

Blomstrom, M. and A. Kokko (1998), ‘Multinational corporations and spillovers,’ Journal of

Economic Surveys, 12, 247-277.

Branstetter, L., G. Li and F. Veloso (2013), ‘The rise of international co-invention,’ in A. B.

Jaffe and B. Jones (eds.), The Changing Frontier: Rethinking Science and Innovation

Policy. University of Chicago Press: Chicago, IL, forthcoming.

Breschi, S. and F. Lissoni (2009), ‘Mobility of skilled workers and co-invention networks: an

anatomy of localized knowledge flows,’ Journal of Economic Geography, 9, 439-468.

$&"

"

Bruche, G. (2012). ‘Emerging Indian pharma multinationals: latecomer catch–up strategies in

a globalised high–tech industry,’ European Journal of International Management, 6,

300-322.

Buckley, P., J. Clegg and C. Wang (2007), ‘Is the relationship between inward FDI and

spillover effect linear? An empirical examination of the case of China,’ Journal of

International Business Studies, 38, 447-59.

Cano-Kollmann, M., R. Mudambi and A. T. Tavares-Lehmann (2013), ‘Innovation networks

in peripheral economies: the case of Portugal’, Paper presented at the 35th DRUID

Celebration Conference. Barcelona, June 19-17, 2013.

Cantwell, J. (1989), Technological Innovation and Multinational Corporations. Basil

Blackwell: London.

Cantwell, J. and G. Santangelo (1999), ‘The frontier of information technology networks:

sourcing abroad the most highly tacit capabilities,’ Information Economics and Policy,

11(1), 101-123.

Cantwell, J. and R. Mudambi (2005), ‘MNE competence-creating subsidiary mandates,’

Strategic Management Journal, 26, 1109-1128.

Cantwell, J. and R. Mudambi (2011), ‘Physical attraction and the geography of knowledge

sourcing in multinational enterprises,’ Global Strategy Journal, 1, 206-232.

Cao, X. (1996), ‘Debating ‘brain drain’ in the context of globalization,’ Compare, 26, 269–

283

$'"

"

Cassiman, B. and R. Veugelers (2006), ‘In search of complementarity in innovation strategy:

internal R&D and external knowledge acquisition,’ Management science, 52, 68-82.

Charles, P. (2003), ‘Universities and territorial development: reshaping the regional role of

UK universities,’ Local Economy, 18, 7–20.

Chittoor, R., M. B. Sarkar, S. Ray and P. S. Aulakh (2009), ‘Third-world copycats to

emerging multinationals: institutional changes and organizational transformation in the

Indian pharmaceutical industry,’ Organization Science, 20, 187-205.

Chung, W. and S. Yeaple (2008). ‘International knowledge sourcing: evidence from US firms

expanding abroad,’ Strategic Management Journal, 29, 1207-1224.

Cockburn, I. M. (2004), ‘The changing structure of the pharmaceutical industry,’ Health

Affairs, 23, 10-22.

Cohen, W. M., R. R. Nelson and J. P. Walsh (2000), ‘Protecting their intellectual assets:

appropriability conditions and why US manufacturing firms patent (or not),’ NBER

Working Paper, No. 7552.

Contractor, F. J., V. Kumar, S. K. Kundu and T. Pedersen (2010), ‘Reconceptualizing the firm

in a world of outsourcing and offshoring: the organizational and geographical

relocation of high!value company functions,’ Journal of Management Studies, 47,

1417-1433.

Cooke, P. (2001), ‘Regional innovation systems, clusters and the knowledge economy,’

Industrial and Corporate Change, 10, 945–974.

$("

"

Corrado, C. and C. Hulten (2010), ‘How do you measure a technological revolution?’

American Economic Review, 100, 99-104.

Cubbin, J. and D. Leech (1983), ‘The effect of shareholding dispersion on the degree of

control in British companies: theory and measurement,’ The Economic Journal, 93,

351-369.

Cuervo-Cazurra, A. and M. Genc (2008), ‘Transforming disadvantages into advantages:

developing-country MNEs in the least developed countries,’ Journal of International

Business Studies, 39, 957-979.

D’Agostino, L. M., K. Laursen and G. D. Santangelo (2013), ‘The impact of R&D offshoring

on the home knowledge production of OECD investing regions,’ Journal of Economic

Geography, 13, 145-175.

Dasgupta, P. and P. David (1994), ‘Toward a new economics of science,’ Research Policy,

23, 487–521.

Davenport, S. (2005) ‘Exploring the role of proximity in SME knowledge-acquisition,’

Research Policy, 34, 683–702.

Deng, P. (2009), ‘Why do Chinese firms tend to acquire strategic assets in international

expansion?’ Journal of World Business, 44, 74-84.

Ejermo, O. and C. Karlsson (2006), ‘Interregional inventor networks as studied by patent

coinventorships,’ Research Policy, 35, 412-430.

$)"

"

Fabrizio, K. R. (2007), ‘University patenting and the pace of industrial innovation,’ Industrial

and Corporate Change, 16, 505-535.

Filatotchev, I., X. Liu, J. Lu and M. Wright (2011), ‘Knowledge spillovers through human

mobility across national borders: evidence from Zhongguancun Science Park in China,’

Research Policy, 40, 453-462.

Fleming, L. and M. Marx, (2006), ‘Managing creativity in small worlds’, California

Management Review, 48(4): 6-27.

Fleming, L. and O. Sorenson (2004), ‘Science as a map in technological search,’ Strategic

Management Journal, 25, 909–928.

Gereffi, G. (1999), ‘International trade and industrial upgrading in the apparel commodity

chain,’ Journal of International Economics, 48, 37-70.

Gertler, M. S. and Y. M. Levitte (2005), ‘Local nodes in global networks: the geography of

knowledge flows in biotechnology innovation,’ Industry and Innovation, 12, 487–507

Ghemawat, P. (2010), Strategy and the Business Landscape. Pearson: Upper Saddle River,

N.J.

Govindarajan, V. and R. Ramamurti (2011), ‘Reverse innovation, emerging markets, and

global strategy,’ Global Strategy Journal, 1, 191-205.

Greene, W. (2000). Econometric Analysis, 4th

edn. Prentice-Hall: Upper Saddle River, NJ.

Griliches, Z. (2000). R & D, Education, and Productivity. Harvard University Press: Boston,

MA.

$*"

"

Haas, M. R. and M. T. Hansen (2005), ‘When using knowledge can hurt performance: the

value of organizational capabilities in management consulting,’ Strategic Management

Journal, 26: 1–24.

Hall, B. H., A. B. Jaffe and M. Trajtenberg (2001), ‘The NBER patent citation data file:

lessons, insights and methodological tools,’ NBER Working Paper, No. 8498.

Hannigan T.J., A. Lee and R. Mudambi (2013), ‘The pitfalls of an inward-oriented economy:

lessons from the evolution of Brazil and the Republic of Korea,’ Transnational

Corporations, 22, 1-24.

Hansen, M. T. (2002), ‘Knowledge networks: explaining effective knowledge sharing in

multiunit companies,’ Organization science, 13, 232-248.

Hobday, M. (2010), ‘Latecomer catch-up strategies in electronics: Samsung of Korea and

ACER of Taiwan,’ Asia Pacific Business Review, 4, 48-83.

Jaffe, A. B. and M. Trajtenberg (2002), Patents, Citations, and Innovations: A Window on the

Knowledge Economy. MIT Press: Cambridge, MA.

Jensen, P. D. Ø. and T. Pedersen (2011), ‘The economic geography of offshoring: the fit

between activities and local context,’ Journal of Management Studies, 48, 352-372.

Jonkers, K. and R. Tijssen (2008), ‘Chinese researchers returning home: impacts on

international mobility on research collaboration and scientific productivity,’

Scientometrics, 77, 309-333.

%+"

"

Kitagawa, F. (2004), ‘Universities and regional advantage: higher education and innovation

policies in English regions,’ European Planning Studies, 12, 835–852.

Kuemmerle, W. (1999), ‘The drivers of foreign direct investment into research and

development: an empirical investigation,’ Journal of International Business Studies,

30,1-24.

Kumar, S. and R. R. Russell (2002), ‘Technological change, technological catch-up, and

capital deepening: relative contributions to growth and convergence,’ American

Economic Review, 92, 527-548.

Kumaraswamy, A., R. Mudambi, H. Saranga and A. Tripathy (2012), ‘Catch-up strategies in

the Indian auto components industry: domestic firms’ responses to market

liberalization,’ Journal of International Business Studies, 43, 368-395

Lahiri, N. (2010), ‘Geographic dispersion of R&D units: how does it help innovation?’

Academy of Management Journal, 53, 1194-1209.

Lane, P. J., J. E. Salk and M. A. Lyles (2001), ‘Absorptive capacity, learning, and

performance in international joint ventures,’ Strategic Management Journal, 22, 1139-

1161.

Laursen, K. and A. Salter (2004), ‘Searching high and low: what types of firms use

universities as a source of innovation?’ Research Policy, 33, 1201-1215.

%!"

"

Laursen, K. and A. Salter (2006), ‘Open for innovation: the role of openness in explaining

innovation performance among U.K. manufacturing firms,’ Strategic Management

Journal, 27, 131-150.

Lewin, A. Y., S. Massini and C. Peeters (2009), ‘Why are companies offshoring innovation?

The emerging global race for talent,’ Journal of International Business Studies, 40,

901-925.

Li, G. C., R. Lai, A. D’Amour, D. M. Doolin, Y. Sun, V. L. Torvik and L. Fleming (2014),

‘Disambiguation and co-authorship networks of the US patent inventor database

(1975–2010),’ Research Policy, 43, 941-955.

Li, J., Y. Li and D. Shapiro (2012), ‘Knowledge seeking and outward FDI of emerging market

firms: the moderating effect of inward FDI,’ Global Strategy Journal, 2, 277-295.

Lorenzen, M. and R. Mudambi (2013), ‘Clusters, connectivity and catch-up: Bollywood and

Bangalore in the global economy,’ Journal of Economic Geography, 13, 501-534

Lundvall, B.-Å. (1992), National Systems of Innovation Towards a Theory of Innovation and

Interactive Learning. Anthem Press: London and New York.

Luo, Y. And R. L. Tung (2007), International expansion of emerging market enterprises: a

springboard perspective,’ Journal of International Business Studies, 38, 481-498.

Madhok, A. and M. Keyhani (2012), ‘Acquisitions as entrepreneurship: asymmetries,

opportunities, and the internationalization of multinationals from emerging economies,’

Global Strategy Journal, 2, 26-40.

%#"

"

Malmberg, A. and P. Maskell (2002), ‘The elusive concept of localization economies:

towards a knowledge-based theory of spatial clustering,’ Environment and Planning A,

34, 429-450.

Mariotti, S., L. Piscitello and S. Elia (2010), ‘Spatial agglomeration of multinational

enterprises: the role of information externalities and knowledge spillovers,’ Journal of

Economic Geography, 10, 519-38.

Martin, R. and P. Sunley (2006), ‘Path dependence and regional economic evolution,’ Journal

of Economic Geography, 6, 395–437.

Mathews, J. A. (2006), ‘Dragon multinationals: new players in 21st century globalization,’

Asia Pacific Journal of Management, 23, 5-27.

McCann, P. (2005), ‘Transport costs and the new economic geography,’ Journal of Economic

Geography, 5, 305-318.

Moreira, M. (1995), Industrialization, Trade and Market Failures: The Role of Government

Intervention in Brazil and South Korea. Palgrave Macmillan: London.

Mudambi, R. (2008), ‘Location, control and innovation in knowledge-intensive industries,’

Journal of Economic Geography, 8, 699-725.

Mudambi, R. and C. Nicosia (1998), ‘Ownership structure and firm performance: evidence

from the UK financial services industry’ Applied Financial Economics, 8, 175-180.

Mudambi, R. and T. Swift (2009), ‘Professional guilds, tension and knowledge management,’

Research Policy, 38(5), 736-745.

%$"

"

Mudambi, R. and M. Venzin (2010), ‘The strategic nexus of offshoring and outsourcing

decisions,’ Journal of Management Studies, 47, 1510-1533.

Narula, R. and J. Guimón (2010), ‘The R&D activity of multinational enterprises in peripheral

economies: evidence from the EU new member states,’ UNU-MERIT Working Paper

Series No. 2010-048.

National Science Board (2014), ‘Industry, technology, and the global marketplace,’ in Science

and Engineering Indicators 2014. National Science Foundation: Arlington, VA,

Chapter 6.

Nelson, R. R. (1993), National Innovation Systems: A Comparative Analysis. Oxford

University Press: New York.

Nelson, R. R. (2004) ‘The market economy, and the scientific commons,’ Research Policy,

33, 455–471.

Nelson, R. R. and S. G. Winter (1982), ‘The Schumpeterian tradeoff revisited,’ American

Economic Review, 72,114-132.

Newman, M. E. J. (2000), ‘Who is the best connected scientists? A study of scientific

coauthorship networks,’ SFI Working Paper 00-12-64.

Pavitt, K. L. R. (2001). Public policies to support basic research: what can the rest of the

world learn from US theory and practice? (and what they should not learn),’ Industrial

and Corporate Change, 10, 761–779.

Perri, A. and U. Andersson (2014), ‘Knowledge outflows from foreign subsidiaries and the

%%"

"

tension between knowledge creation and knowledge protection: evidence from the

semiconductor industry,’ International Business Review, 23, 63-75.

Phene, A. and P. Almeida (2008), ‘Innovation in multinational subsidiaries: the role of

knowledge assimilation and subsidiary capabilities,’ Journal of International Business

Studies, 39, 901-919.

Phene, A., K. Fladmoe-Lindquist and L. Marsh (2006), ‘Breakthrough innovations in the U.S.

biotechnology industry: the effects of technological space and geographic origin,’

Strategic Management Journal, 27, 369– 388.

Porter, M. E. (1990), ‘The competitive advantage of notions,’ Harvard Business Review, 68,

73-93

Salter, A. and B. R. Martin (2001), ‘The economic benefits of publicly funded basic research:

a critical review,’ Research Policy, 30, 509–532.

Saxenian, A. L. (1994), Regional Advantage. Harvard University Press: Cambridge, MA.

Saxenian, A. (2005), ‘From brain drain to brain circulation: transnational communities and

regional upgrading in India and China,’ Studies in Comparative International

Development, 20, 35–61.

Scalera, V. G., A. Perri and R. Mudambi (2015), ‘Managing innovation in emerging

economies: organizational arrangements and resources of foreign MNEs in the Chinese

pharmaceutical industry,’ Advances in International Management, forthcoming.

%&"

"

Siegel, D. S., D. A. Waldman, L. E. Atwater and A. N. Link (2003), ‘Commercial knowledge

transfers from universities to firms: improving the effectiveness of university-industry

collaboration,’ Journal of High Technology Management Research, 14, 111–133.

Tallman, S. and A. S. Chacar (2011), ‘Communities, alliances, networks and knowledge in

multinational firms: a micro-analytic framework’ Journal of International

Management, 17, 201-210.

Tallman, S. and J. Li (1996), ‘Effects of international diversity and product diversity on the

performance of multinational firms,’ Academy of Management Journal, 39, 179-196.

Thomson Reuters (2010), Chinese Patenting. Report on the Current State of Innovation in

China, available at http://fordhamipconference.com.

UNCTAD (2005), Transnational Corporations and the Internationalization of R&D. World

Investment Report 2005. United Nations Conference for Trade and Development: New

York and Geneva.

Wei, Y., & Liu, X. (2006). Productivity spillovers from R&D, exports and FDI in China's

manufacturing sector. Journal of international business studies, 37(4), 544-557.

Zhao, M. (2006). Conducting R&D in countries with weak intellectual property rights

protection. Management Science, 52(8): 1185-1199.

!"#

#

Table 1. Descriptive statistics and correlation table

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

(1)Geographic dispersion 1

(2)Emerging -0.665 1

(3)MNE 0.257 -0.320 1

(4)University -0.045 0.250 -0.432 1

(5)Leader 0.471 -0.470 0.480 -0.043 1

(6)Team size 0.146 -0.090 0.129 0.073 0.137 1

(7)Design -0.113 0.059 -0.022 -0.230 -0.127 -0.217 1

(8)Pharma -0.040 0.012 -0.023 -0.107 -0.023 0.057 0.164 1

(9)Tech composition -0.114 0.101 0.007 -0.013 -0.088 -0.128 0.155 0.035 1

(19)Co-assigned -0.017 0.121 0.062 0.265 0.052 0.210 -0.105 0.050 -0.019 1

(11)Year dummy (2005) 0.065 -0.090 -0.061 0.114 0.059 -0.074 -0.085 0.092 0.008 0.014 1

Obs. 1026 1026 1026 1026 1026 1026 1026 1026 1026 1026 1026

Mean 0.199 0.340 0.361 0.472 0.319 3.845 0.086 0.682 0.940 0.324 0.105

Std. Dev. 0.235 0.474 0.480 0.499 0.466 3.073 0.280 0.466 0.075 0.468 0.307

Min 0.000 0.000 0.000 0.000 0.000 1.000 0.000 0.000 0.673 0.000 0.000

Max 0.82 1 1 1 1 31 1 1 1 1 1

!"#

#

Table 2. Tobit regression results (Dependent variable = Geographical dispersion).

Model (1) Model (2) Model (3) Model (4) Model (5)

# # # # #

Leader 0.446*** 0.143*** 0.104*** 0.0868** 0.103***

(0.0311) (0.0259) (0.0290) (0.0303) (0.0290)

# # # # #

Team size 0.0211*** 0.0124** 0.0121** 0.0129*** 0.0125**

(0.00485) (0.00391) (0.00389) (0.00389) (0.00394)

# # # # #

Design -0.0675 -0.103* -0.0800 -0.0818 -0.0791

(0.0597) (0.0508) (0.0515) (0.0508) (0.0516)

# # # # #

Pharma -0.0435 -0.0391 -0.0261 -0.0248 -0.0257

(0.0311) (0.0253) (0.0252) (0.0251) (0.0252)

# # # # #

Tech. composition -0.419* -0.182 -0.218 -0.220 -0.222

(0.191) (0.155) (0.153) (0.153) (0.154)

# # # # #

Co-assigned -0.107* 0.0770† 0.0314 0.0367 0.0278

(0.0479) (0.0395) (0.0403) (0.0405) (0.0407)

Year dummy (2005) 0.051 0.016 0.007 0.009 0.007

(0.0308) (0.0249) (0.0248) (0.0248) (0.0248)

# # # # #

Emerging -0.617*** -0.646*** -0.598*** -0.657***

# (0.0331) (0.0344) (0.0418) (0.0386)

# # # # #

MNE # # 0.0665*# 0.0833*# 0.0588†#

# # (0.0318)# (0.0329)# (0.0338)#

# # # # #

University # 0.137*** 0.178*** 0.137***

# # (0.0328) (0.0399) (0.0329)

# # # # #

University * Emerging # # -0.108†

# # # # (0.0599)

# # # # # #

MNE * Emerging # # # 0.0432

# # # # (0.0629)

# # # # #

Cons 0.194 0.374* 0.358* 0.345* 0.365*

(0.183) (0.148) (0.147) (0.146) (0.147)

# # # # #

Obs. 1026 1026 1026 1026 1026

Log likelihood (LL) -551.11 -329.46 -320.59 -318.99 -320.36

!2 283.05*** 726.35*** 744.09*** 747.29*** 744.56***

Pseudo R2

0.204 0.524 0.537 0.54 0.538

Note: Standard errors in parentheses

† p<0.1, * p<0.05, ** p<0.01, *** p<0.001

!$#

#

Table 3. Test on the equality of MNEs and Uni_Rc coefficients.

F( 1, 1016) = 5.58

Prob > F = 0.0184