Connecting local and global technological knowledge sourcing

John Cantwell (Rutgers Business School, Rutgers University, Newark, New Jersey, USA)
Salma Zaman (Rutgers Business School, Rutgers University, Newark, New Jersey, USA)

Competitiveness Review

ISSN: 1059-5422

Publication date: 21 May 2018



Through increasing globalization, cities are becoming increasingly interconnected with each other. To remain competitive, it is necessary for cities to combine complementary non-local sources of knowledge with local knowledge sources. The authors contend that an increase in non-local knowledge sourcing tends to enhance local knowledge sourcing too. The purpose of this study is to examine the influence of international knowledge sources on the capacity to build upon local knowledge sources in a city region. In addition, the authors investigate whether information and communication technologies (ICT) knowledge sources have a bigger impact than do other fields of knowledge on local knowledge connectivity.


Using the US Patent and Trademark Office data, the authors study knowledge sourcing trends for the years 1980-2016 across 33 global cities. Backward patent citations from these granted patents are used to identify the location of inventors of prior knowledge sources, and the geography of knowledge building connections over time is assessed by using the inventor locations of cited (source) and citing (recipient) patents.


The authors show that international knowledge sourcing is highly connected with local knowledge sourcing. The authors also find that ICT have a significant effect on this relationship. However, there are significant differences across cities in the extent and nature of this relationship.


This study contributes to the literature on the changing geography of knowledge connections. It provides a detailed picture of changing knowledge sourcing trends at a city region level, and it improves our understanding of the role played by a variety of knowledge connections in helping a city remain competitive.



Cantwell, J. and Zaman, S. (2018), "Connecting local and global technological knowledge sourcing", Competitiveness Review, Vol. 28 No. 3, pp. 277-294.

Download as .RIS



Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited

1. Introduction

In today’s globalized information age, knowledge plays an increasingly important role. According to Grant (2002), the role of knowledge in today’s economy corresponds to that of land in agrarian economies and that of capital in the early industrial economies. Today, cities and clusters cannot rely exclusively on local knowledge sources, but they need to combine “local buzz” (Storper and Venables, 2004) with “global pipelines” (Bathelt et al., 2004).

The current information age, with its advances in information and communication technologies (ICT), has facilitated the diffusion of knowledge across regions by lowering transport and communication costs (Foss and Pedersen, 2004). In addition, contemporary ICT have allowed combinations of previously separate lines of technological development (Santangelo, 2002). As individual locations are increasingly specialized in their activity (Cantwell and Vertova, 2004), international connections are generally necessary for such new combinations of innovative activity. Because of these changes in the environment for innovation, we would expect innovative cities to be progressively more connected with each other than ever before.

Our study contributes to the literature on the changing nature of knowledge connections across space and the complementarity of external and local connections. We do so by providing a detailed picture of how the structure of connections is changing, and how intra- and inter-regional knowledge sources influence each other across a mix of developed and developing cities. In particular, we look at 33 global cities to see how their international citations have affected their local citations. These include 13 US cities (Seattle, Austin, San Diego, Pittsburgh, NY City, Los Angeles, Boston, Chicago, the Bay Area, Miami, Atlanta, Houston and Dallas), seven European cities (London, Paris, Berlin, Frankfurt, Munich, Hamburg and Stuttgart), 12 Asian cities (Tokyo, Osaka, Nagoya, Singapore, Seoul, Hong Kong, Beijing, Shanghai, Guangzhou, Mumbai, Delhi and Bangalore) and Sydney, Australia. We selected cities above a certain threshold level of patenting that were also represented in the Globalization and World Cities Research Network (GaWC) classification of global cities. Our database consists of patents granted by the US Patent and Trademark Office (USPTO) over the period 1980-2016. Patent citations were used to identify the location of knowledge sources and recipients by using the inventor locations of cited (source) and citing (recipient) patents.

We found that in all cities there is a significant correlation between international and local citations. Increases in international knowledge connectivity are more apparent in the later years, which are characterized by the diffusion phase of the information age across a wider variety of industries and activities (Alcácer et al., 2016). This is consistent with the stream of literature which emphasizes that external and local knowledge connections complement each other and are necessary for the innovativeness of a particular region (Uzzi, 1997; Bramanti and Ratti, 1997; Maillat, 1998; Scott, 1988; Bresnahan et al., 2001; Bathelt, 2007). Cities from developing countries were already highly internationalized at the beginning of our time period, which illustrates the reliance of emerging markets on global knowledge sources for development in the current era. This is also consistent with the literature which shows that actors in emerging markets benefit from greater global knowledge connectivity (Cantwell and Zhang, 2013).

The rest of the paper is structured as follows. In Section 2, we review the related literature. In Section 3, we discuss our data and methodology. Finally, we conclude our study in Section 4 with an analysis of our findings and potential directions for future research.

2. Literature review

Previous research has emphasized the importance of local and regional linkages (Cooke and Morgan, 1998; Malmberg and Maskell, 2002; Porter, 1996). Local linkages, because of their geographical proximity, facilitate the exchange of complex and valuable knowledge (Giuliani, 2013). This is because complex forms of knowledge are more difficult to communicate over distance (Sorenson, 2005). Close and face-to-face interactions among the actors reduce uncertainty associated with the transfer of tacit knowledge, and embeddedness in the local environment helps companies to devise context-specific solutions (Perez-Aleman, 2011). Close long-term supplier–customer relations are advantageous because they enable both customers and suppliers to benefit from user-producer learning processes and to react quickly to either upstream or downstream market shifts (Granovetter, 1985). Knowledge spillovers are therefore positively associated with co-location (Jaffe et al., 1993) and the size and mobility of the labor pool (Almeida and Kogut, 1999).

However, there are limitations to the effectiveness of local linkages. One issue is the concern over the possibility of adverse selection in the composition of firms attracted to locate in clusters, if better firms believe that the costs of knowledge leakage may for them exceed the potential benefits of knowledge acquisition (Shaver and Flyer, 2000; Chung and Kalnins, 2001). However, more technologically capable leading firms may also be better positioned to become insiders in already dense local cluster networks, which would encourage them to enter such places (Cantwell and Mudambi, 2011). Other limitations are discussed in a stream of literature which emphasizes the importance of complementing local ties with external linkages to boost the innovativeness of a particular region (Uzzi, 1997; Bramanti and Ratti, 1997; Maillat, 1998; Scott, 1988; Bresnahan et al., 2001; Bathelt, 2007).

This literature claims that knowledge that is based on a similar context tends to discourage diversity. Therefore, if knowledge is generated and transferred only within a local regional network, it may run the risk of lock-in processes and isolation, causing a decline in innovation in the long term (Bathelt et al., 2004; Lazer and Friedman, 2007). Moreover, close long-term supplier–customer relations may not always be beneficial, especially if a group of suppliers is strongly embedded in relations with the same set of customers and so become dependent upon them. An over-embeddedness may reinforce lock-in effects and induce over confidence and naïveté, all of which threaten to ultimately undermine the success of a region (Uzzi, 1996).

Hence, external knowledge connections are important because they offer actors within a region knowledge from diverse sources, and variety imbues a greater capacity for change and adaptation to a shift in opportunities. When knowledge is continuously reused or applied in different contexts across locations, new knowledge generation processes may be triggered (Cantwell, 1989). Geographical separation may, therefore, be conducive to innovation (Bathelt and Glückler, 2011).

Local knowledge networks and external connections complement one another and aid in the innovativeness of a region. Without external knowledge connections, there is a chance that firms will find only unsatisfactory workarounds for the problems they encounter and lose their competitive edge. External connections ameliorate such constraints on local capability development by aiding in the diffusion of knowledge within a cluster, which in turn stimulates additional local knowledge creation and serves as a basis for further innovation (Owen-Smith and Powell, 2004). On the other hand, in the absence of adequate local connections, these external connections may be of limited use, as they would then provide little basis for further local development or extensions. Local knowledge assists firms in sifting through the large volume of available opportunities to identify the knowledge that is particularly important for solving local problems or moving into new domains of application while discarding that which has little local relevance (Bathelt and Glückler, 2011).

2.1 Barriers to the transfer of knowledge

Local and external knowledge connections need to become interrelated with one another to obtain the greatest benefit from each. However, there are barriers to the transfer of knowledge and to establishing new knowledge combinations, especially when the geographical distance between knowledge sources rises (Lamorgese and Ottaviano, 2006; Lychagin et al., 2016; Jaffe et al., 1993). These difficulties are further compounded when knowledge is being transferred across national borders. The literature on global value chains and global production networks has emphasized the challenges and complexities in organizing and maintaining knowledge linkages that cut across national boundaries (Humphrey and Schmitz, 2002; Gereffi et al., 2005; Coe et al., 2010). This is because national boundaries are often proxies for cultural and language barriers, as well as political and administrative barriers. In addition, industry standards, methods of measurement and other professional conventions may also vary across national boundaries (Teece, 1977), which increases the likelihood of incompatibility of knowledge structures or misunderstandings. The likelihood of finding capable and reliable knowledge contacts also declines with distance, which diminishing strength of network ties over distance implies a spatial concentration of knowledge (Fujita and Thisse, 2013). In addition, trust between the sender and receiver and their commitment to follow up where necessary may be weaker if they are in different countries, which may inhibit the transfer process (Szulanski, 1996; Wathne et al., 1996; Albino et al., 1998).

Another barrier to the transfer of knowledge is the type of knowledge being transferred. Only the broad outlines of technical knowledge are usually codifiable (Berrill, 1964). The tacit component of knowledge is not easy to replicate across different contexts (Polanyi, 1966; Von Hippel, 1994; Szulanski and Jensen, 2004), as it is embodied in large part in localized organizational routines and the collective expertise of specific production teams (Nelson and Winter, 1982). Additionally, knowledge may be causally ambiguous, which hampers knowledge transfer as well (Lippman and Rumelt, 1982; Szulanski, 1996; Argote et al., 2003), as having a clearer grasp of the underlying rationale may become more important once the context for the application of that knowledge changes.

However, knowledge transfer across national boundaries remains feasible, even if challenging. Despite its initial context specificity, tacit knowledge may flow both locally and across longer distances (Brannen, 2004). While studying the impact of the trade-weighted R&D of other countries on a country’s productivity growth, Park (1995) and Coe and Helpman (1995) found a positive effect. This can be regarded as evidence of knowledge spillovers across international borders, at least after the passage of time. Jaffe and Trajtenberg (1996) also show that the domestic citation probabilities of local inventions are particularly high in the early years after an invention is made but decrease over time; so any shift from the domestic to the international use of knowledge may occur with a lag, once a surrounding knowledge base has been built up locally, widening the ease and scope of potential foreign applications.

2.2 The current information age

By the late 1970s, the old science-based and oil-driven electro-mechanical era was gradually replaced by the present information age. While the previous era was based on mass production, economies of scale and specialized in-house corporate R&D, the new era is characterized by economies of scope with a greater diversity and geographic dispersion of search in R&D (Cantwell and Santangelo, 2002). In this new age, we believe that forming international connections has become easier and also more necessary (Cano-Kollmann et al., 2016).

The advances in ICT have lowered transport and communication costs thereby accelerating the process of knowledge creation and diffusion (Foss and Pedersen, 2004). In addition ICT have made previously distant technological combinations possible (Cantwell and Santangelo, 2002). The structures of corporate technological competencies have become more adaptable than they once were. Their existing technological competencies may also have multiple uses both within and outside their primary sector of activity (Langlois and Robertson, 1995). Therefore, this age is characterized by an increase in inter-organizational collaboration and openness (Chesbrough, 2003).

In this current age, knowledge is becoming increasingly complex in character and firms must now possess a wider range of technological skills (Feldman and Audretsch, 1995). As a consequence, technological interrelatedness is also rising. There is also evidence that industrialized countries are becoming more technologically specialized and differentiated from each other over time (Cantwell and Vertova, 2004), thereby increasing the importance of international knowledge linkages for each location.

Because of the increasing importance of accessing external knowledge sources, we expect to observe an increase in international patent citations, despite barriers to the cross-border transfer of knowledge. This increase should be more apparent in global cities, as by definition, they have become more interconnected in global networks that form the backbone of the global economy (Goerzen et al., 2013).

3. Data and methodology

For the purpose of this study, we use patents granted by the USPTO as our primary data source. Patent citations are used to identify the location of knowledge sources and recipients by using the first named inventor locations of cited (source) and citing (recipient) patents. Our patent database consists of a panel of 33 global cities over the period 1980-2016. Using these data, we identify how in our 33 global cities the share of international (or non-local, as opposed to purely local, intra-city) sources has changed over time. We analyze whether the changes in the number of international citations have impacted the local citations of each of the cities in our data set.

For this study, we have chosen to focus on a selection of global cities. We identified 33 cities with patents above a certain threshold level that depended on the national context, to ensure coverage across a wide range of countries, which cities were also listed in the GaWC classification of global cities, as explained earlier. To define the inventor locations included in each of the 13 US cities, we used the combined statistical area as defined by the USA Office of Management and Budget. For the seven European cities, we used the European Union NUTS classification of city regions. We used the administrative divisions of China to define the four Chinese cities and Indian Government divisions to define the three Indian cities. For Tokyo, we used the National Capital Region Planning Act definition, while for Nagoya and Osaka, we used definitions by the Statistics Bureau of Japan. Seoul was defined using the Korean National Statistics Office’s definition, while for Sydney, we used the definition provided by Australia Bureau of Statistics.

We begin our analysis by generating some summary statistics to form an overall impression of how the geographic distribution of citations is changing across cities. We calculate the percentage of local citations and non-local citations in total backward citations for patents invented in each city, and then examine how these shares have changed over time. We present our findings by calculating the linear trends of local citations and non-local citations over our time period, which we display in line graphs.

3.1 US cities

Most of the US cities showed a trend toward using less local knowledge sources than before. However, it was apparent from the data that while in most cases, the share of local and non-local sources did not show a very strong trend, the number of both local and international citations were increasing. Figure 1 shows how the percentage change in local citations have changed across our 13 US cities. As we can see in the graph, all cities apart from Houston, The Bay Area and Seattle showed a downward trend since 1980 in the percentage of local citations.

Most of the US cities, apart from Houston and The Bay Area showed an upward trend in the percentage of non-local citations (which includes citations from other US cities as well as well as international citations). These trends are visible in the Figure 2.

While non-local citations showed a clear upward trend for most of the US cities, international citations showed a downward trend. These trends are visible in Figure 3, which when considered in conjunction with Figure 2, indicates that the US cities have been increasing their cross-US national connections (often with one another) more rapidly than their global knowledge connections.

3.2 European cities

The European cities were highly internationalized in their knowledge sourcing from the outset and, gradually, became even more internationalized. Figures 4 and 5 show how the share of local citations and non-local citations have changed over our time period.

3.3 Japanese cities

Japanese cities showed little overall change in the shares of non-local citations and local citations over time. Tokyo showed a slight trend towards using more localized sources, while Nagoya and Osaka showed a slight trend toward relying on more delocalized sources. These trends are illustrated in Figures 6 and 7.

3.4 Other East Asian cities

Other East Asian cities were all highly internationalized from the very beginning. Seoul and Singapore showed a trend toward using more local sources, while Hong Kong showed a trend increase in the share of non-local sources. These trends are shown in Figures 8 and 9.

3.5 Sydney, Australia

Sydney, Australia was also highly internationalized at the beginning of our period, and it showed a gradual decrease in the share of non-local citations. This trend is illustrated in Figure 10.

In this study, we analyze the impact of international citations on local citations. Our data set contains panel data ranging from the years 1976-2016. In our study, we use Least Squares regression with dummy variables (LSDV) with fixed effects for each city. The rationale for choosing fixed effects for cities was that they would help us control for those factors that are characteristic of each city and otherwise hard to control for. This could include culture and business practices. As these factors can be assumed to be largely consistent from year to year, fixed effects are the appropriate method to use. To further justify our use of a fixed effects model instead of a random effects model, we conducted a Hausman test. In the Hausman test, the null hypothesis is that the random effects model is appropriate, while the alternative hypothesis is that the fixed effects model is appropriate. We ran the Hausman test and obtained a significant p value of 0.0002. Therefore, we reject the null hypothesis and accept the fixed effects model as better suited for our purpose.

3.6 Dependent variable

3.6.1 Share of local citations.

Our dependent variable Lit is calculated as follows:

Lit= Number of local citations by city i in year tTotal number of all local citations across all cities in year t*10,000

In this equation, local citations are those cited patents in which the first inventor location is in the same city region as the citing patent. For example, for a patent whose first inventor’s address is in the New York City (NYC) region, a local citation would be a cited patent whose first inventor’s address is also in the NYC region.

We divide the total number of local citations for each city in a certain year by the total number of all citations in that year to control for the increasing number of citations. Our data show that in 1980 the total number of citations for all USPTO patents was 348,010, while in 2016, the total number of citations for all USPTO patents was 9,803,647. Hence, we need to control for this exponential rise in citations.

3.7 Independent variable

3.7.1 Share of international citations.

Our independent variable Iit is calculated as follows:

Iit= Number of international citations by city i in year tTotal number of all international citations across all cities in year t*10,000

In this equation, international citations are those cited patents in which the first inventor location is in a different country than the first inventor of the citing patent. For example, for a patent whose first inventor’s address is in the NYC region, an international citation would be a cited patent whose first inventor’s address is outside of the USA.

3.8 Moderating variable

3.8.1 Share of ICT citations.

As ICT act as connectors to link previously unrelated technologies (Cantwell and Santangelo, 2000), we expect that the effect of the international citations on local citations to be amplified by the proportion of ICT technologies. Therefore, we calculate the share of ICT citations and add this as a moderating variable in our model. The share of ICT patents is calculated as follows:

IICTit= Number of international ICT citations by city i in year tTotal number of all international ICT citations across all cities in year t*10,000

To categorize patents as ICT patents, we first classified patents using their USPTO classes and sub-classes into 56 technological fields (as set out in Cantwell, 1995). The six technological fields (out of 56) that are recognized as ICT fields are as follows: telecommunications, other electrical communication systems, special radio systems, image and sound equipment, semiconductors and office equipment and data processing systems (Cantwell and Santangelo, 1999).

We also ensured our model was robust and controlled for heteroscedasticity and autocorrelation.

Defining Ci as City i, the model we estimate is:

Lit= β0+β1Iit+β2IICTit+ β3(Iit*IICTit)+ β4i=233Ci

Taking the Seattle region as the reference city (i = 1), we obtain the results displayed in Table I.

The 13 US cities in our data set are different from other cities because even though they demonstrate a lesser amount of internationalization than other cities, they are increasingly citing other cities in the US. This can be seen in Figures 1, 2 and 3. To cater for this, we develop another model. In this model, we define our dependent variable, the share of local citations as we did in the previous model. However, our independent variable is the share of non-local citations, which includes those citations that are within the same country but outside the area of the focal city.

The new independent variable Nit is calculated as follows:

Nit= Number of nonlocal citations by city i in year tTotal number of all nonlocal citations across all cities in year t*10,000

In this equation, non-local citations are those cited patents in which the first inventor location is not in the same city region as the first inventor location of the citing patent. That is, this includes domestic patents that are not in the same city. For example, for a patent whose first inventor’s address is in the NYC region, a non-local citation could be a cited patent whose first inventor’s address is outside of the USA or in a different city in the USA.

Just like in our previous model, we include the share of ICT citations as a moderating variable. In this model, instead of calculating the share of ICT international citations we calculate the share of ICT non-local citations. The formula for the moderating variable in this case is:

NICTit= Number of nonlocal ICT citations by city i in year tTotal number of all nonlocal ICT citations across all cities in year t*10,000

This revised model becomes:

Nit= β0+β1Nit+β2NICTit+β3(Nit*NICTit)+ β4i=233Ci

The results of our regression are displayed in Table II.

4. Results and discussion

Our results show that changes over time in local citations and international citations are highly correlated, allowing for city fixed effects. This shows that being more connected worldwide is associated with increased local connectivity as well. We can also see that there is a positive and significant interaction effect of ICT knowledge sources and international knowledge sources on local knowledge sources. This is in line with the literature that suggests ICT acts as a fusion technology and enables previously separate technology sources to be increasingly connected.

If we look at the fixed effects associated with particular cities, we can see that these effects vary across cities. The base city in our regression model is Seattle. The coefficients for each city represent the difference in the proportionality or balance between the share of international connections and the share of local connections as compared to the benchmark city of Seattle. In Table I, we can see that there are quite a few cities for which the relationship between international and local connection varies significantly. This means, in cities that have positive coefficients, such as The Bay Area and Tokyo, the share of the local citations is relatively high in relation to the share of international citations, when compared with Seattle. In contrast, in cities such as London, Paris and Singapore, which have significant variation and negative coefficients, the share of local connections is relatively low in relationship to the share of international connections as compared with Seattle. These findings are in line with the more descriptive evidence for the equivalent cities presented earlier.

In our second regression model, the independent variable is the share of non-local citations instead of share of international citations. This is worth considering separately because most cities in the US show little or no internationalization trend, but they do show de-localization trends, i.e. they are increasingly citing other cities but within the US. Therefore, we conduct this second regression to see the effect of these de-localized citations on local citations. Our regression results are shown in Table II. In Table II, we see that cities such as London, Paris and Singapore which had negative coefficients in the international case now have a positive coefficient. This is because, compared with most of the leading international cities for technological development outside the US, most of the US cities including Seattle have had relatively low levels of international knowledge connections but relatively high levels of non-local connections. In other words, as an approximate generalization, the non-US cities are more likely to be embedded in a global system of knowledge ties beyond the country in which they are situated, whereas the US cities tend to be relatively more embedded in the US national system of innovation, including cross-city knowledge connections within the US.

Our results show that local and non-local knowledge sources interact with each other in various ways. While this interactive connection between local and non-local knowledge sourcing holds consistently across all cities, some cities especially outside the US rely more on international citations in proportion to local citations, whereas in many of the US cities, there is more of a three-way combination between international, non-local domestic and local citations. Further research needs to be conducted on the cross-field and geographic distribution of the knowledge sourcing of individual cities to further deepen our understanding of the similarities and the differences in knowledge exchange patterns between cities that we have outlined here.


Linear trends in percentage share of local citations of the US city patents

Figure 1.

Linear trends in percentage share of local citations of the US city patents

Linear trends in percentage share of non-local citations of the US city patents

Figure 2.

Linear trends in percentage share of non-local citations of the US city patents

Linear trends in percentage share of international citations of the US city patents

Figure 3.

Linear trends in percentage share of international citations of the US city patents

Linear trends in percentage share of local citations of European city patents

Figure 4.

Linear trends in percentage share of local citations of European city patents

Linear trends in percentage share of non-local citations of European patents

Figure 5.

Linear trends in percentage share of non-local citations of European patents

Linear trends in percentage share of local citations of Japanese city patents

Figure 6.

Linear trends in percentage share of local citations of Japanese city patents

Linear trends in percentage share of non-local citations of Japanese city patents

Figure 7.

Linear trends in percentage share of non-local citations of Japanese city patents

Linear trends in percentage share of local citations of East Asian city patents

Figure 8.

Linear trends in percentage share of local citations of East Asian city patents

Linear trends in percentage share of non-local citations of East Asian city patents

Figure 9.

Linear trends in percentage share of non-local citations of East Asian city patents

Linear trends in percentage share of local, non-local and international citations of Sydney patents

Figure 10.

Linear trends in percentage share of local, non-local and international citations of Sydney patents

Results of the regression of the share of local citations on the share of international citations

Share of local citations Coefficient Std. Err. T p > t [95% conf. interval]
Share of international citations 1.4523 0.0635252 22.86 0.000 1.327662 1.576938
Share of ICT international citations −0.625 0.0558951 −11.19 0.000 −0.7351008 −0.5157659
Share of international citations × share of ICT citations 0.0054 0.0001491 36.35 0.000 0.0051275 0.0057126
Austin −1.553018 1.354018 −1.15 0.252 −4.209632 1.103597
San Diego −3.629109 1.371169 −2.65 0.008 −6.319375 −0.9388424
Pittsburgh −2.520338 1.376312 −1.83 0.067 −5.220694 0.1800176
NYC 2.239438 1.446316 1.55 0.122 −0.5982671 5.077143
LA 0.9960848 1.391244 0.72 0.474 −1.733569 3.725739
Boston −6.513171 1.478834 −4.40 0.000 −9.414678 −3.611665
Chicago −3.423472 1.403015 −2.44 0.015 −6.17622 −0.6707239
SF (Bay Area) 11.93471 1.456484 8.19 0.000 9.077054 14.79237
Miami −3.243976 1.369591 −2.37 0.018 −5.931146 −0.556805
Atlanta −3.548404 1.369068 −2.59 0.010 −6.234547 −0.862261
Houston 2.092523 1.414805 1.48 0.139 −0.6833582 4.868404
Dallas −2.944296 1.360858 −2.16 0.031 −5.614332 −0.2742598
London −6.496712 1.367496 −4.75 0.000 −9.179771 −3.813653
Paris −6.291601 1.378204 −4.57 0.000 −8.99567 −3.587533
Tokyo 5.944818 1.490298 3.99 0.000 3.020817 8.868818
Osaka −7.435172 1.401231 −5.31 0.000 −10.18442 −4.685924
Nagoya −4.769985 1.399866 −3.41 0.001 −7.516555 −2.023415
Singapore −4.193638 1.363178 −3.08 0.002 −6.868224 −1.519051
Seoul −14.04513 1.362343 −10.31 0.000 −16.71808 −11.37218
Berlin −2.963758 1.369394 −2.16 0.031 −5.650542 −0.2769737
Frankfurt −2.333463 1.369961 −1.70 0.089 −5.021358 0.3544322
Munich −3.649897 1.361133 −2.68 0.007 −6.320473 −0.9793216
Hamburg −2.526014 1.369606 −1.84 0.065 −5.213212 0.1611848
Stuttgart −5.50344 1.382046 −3.98 0.000 −8.215048 −2.791832
Hong Kong −3.506257 1.370707 −2.56 0.011 −6.195616 −0.8168978
Sydney −7.437122 1.366517 −5.44 0.000 −10.11826 −4.755984
Beijing −2.533294 1.410494 −1.80 0.073 −5.300717 0.2341288
Shanghai −3.560693 1.430376 −2.49 0.013 −6.367125 −0.7542613
Guangzhou −3.688572 1.455918 −2.53 0.011 −6.545117 −0.8320265
Mumbai −2.173915 1.3883 −1.57 0.118 −4.897792 0.5499615
Delhi −1.990016 1.406567 −1.41 0.157 −4.749734 0.7697024
Bangalore −2.330454 1.431066 −1.63 0.104 −5.138239 0.4773313
_cons 1.825547 0.9803071 1.86 0.063 −0.0978391 3.748932
Source SS Df MS Number of observations 1,189
Model 913037.445 35 26086.7841 F(35, 1153) 773.04
Residual 38908.6455 1,153 33.7455728 Prob > F 0
Total 951946.09 1,188 801.301423 R-squared 0.9591
Adj R-squared 0.9579
Root MSE 5.8091

Results of the regression of the share of local citations on the share of non-local citations

Share of local citations Coefficient Std. Err. T p > t [95% conf. interval]
Share of non-local citations 0.8272691 0.0569273 14.53 0.000 0.7155765 0.9389618
Share of non-local ICT connections 0.0147609 0.0518195 0.28 0.776 −0.0869102 0.116432
Share of non-local citations × share of ICT citations 0.0018267 0.0001008 18.12 0.000 0.0016289 0.0020244
Austin 1.892671 1.571459 1.20 0.229 −1.190568 4.97591
San Diego 0.8501059 1.631918 0.52 0.603 −2.351756 4.051967
Pittsburgh 4.580507 1.628466 2.81 0.005 1.385419 7.775596
NYC 5.360982 1.673449 3.20 0.001 2.077635 8.644328
LA 3.04131 1.715135 1.77 0.076 −0.3238265 6.406446
Boston −1.246959 1.746153 −0.71 0.475 −4.672951 2.179034
Chicago 2.572022 1.662461 1.55 0.122 −0.6897653 5.83381
SF (Bay Area) 14.59104 1.694027 8.61 0.000 11.26732 17.91476
Miami 2.541926 1.616954 1.57 0.116 −0.630575 5.714427
Atlanta 1.311918 1.598252 0.82 0.412 −1.823891 4.447727
Houston 9.872109 1.694888 5.82 0.000 6.546699 13.19752
Dallas 1.439312 1.597354 0.90 0.368 −1.694734 4.573358
London 3.670204 1.614953 2.27 0.023 0.5016276 6.838781
Paris 4.468979 1.620957 2.76 0.006 1.288623 7.649335
Tokyo 44.49381 1.619925 27.47 0.000 41.31548 47.67214
Osaka 4.903657 1.622651 3.02 0.003 1.719978 8.087335
Nagoya 5.721883 1.63769 3.49 0.000 2.508696 8.93507
Singapore 4.243189 1.61377 2.63 0.009 1.076934 7.409444
Seoul 2.704504 1.581782 1.71 0.088 −0.3989904 5.807998
Berlin 4.80974 1.618779 2.97 0.003 1.633657 7.985823
Frankfurt 4.987445 1.619286 3.08 0.002 1.810367 8.164522
Munich 4.257093 1.612493 2.64 0.008 1.093344 7.420842
Hamburg 4.934642 1.619006 3.05 0.002 1.758114 8.11117
Stuttgart 4.481944 1.62576 2.76 0.006 1.292164 7.671724
Hong Kong 4.690707 1.619202 2.90 0.004 1.513794 7.86762
Sydney 4.577657 1.611513 2.84 0.005 1.415831 7.739484
Beijing 4.638224 1.668261 2.78 0.006 1.365056 7.911392
Shanghai 4.514315 1.688083 2.67 0.008 1.202257 7.826373
Guangzhou 4.512774 1.717121 2.63 0.009 1.143743 7.881806
Mumbai 5.036419 1.639793 3.07 0.002 1.819105 8.253732
Delhi 5.022106 1.661476 3.02 0.003 1.762251 8.281962
Bangalore 4.377141 1.69023 2.59 0.010 1.060869 7.693413
_cons −5.134624 1.184269 −4.34 0.000 −7.458188 −2.81106
Source SS df MS Number ofobservations 1,189
Model 900116.834 35 25717.6238 F(35, 1153) 572.12
Residual 51829.2564 1,153 44.9516535 Prob > F 0
Total 951946.09 1,188 801.301423 R-squared 0.9456
Adj R-squared 0.9439
Root MSE 6.7046


Albino, V., Garavelli, A.C. and Schiuma, G. (1998), “Knowledge transfer and inter-firm relationships in industrial districts: the role of the leader firm”, Technovation, Vol. 19 No. 1, pp. 53-63.

Alcácer, J., Cantwell, J.A. and Piscitello, L. (2016), “Internationalization in the information age: a new era for places, firms, and international business networks?”, Journal of International Business Studies, Vol. 47 No. 5, pp. 499-512.

Almeida, P. and Kogut, B. (1999), “Localization of knowledge and the mobility of engineers in regional networks”, Management Science, Vol. 45 No. 7, pp. 905-917.

Argote, L., Mcevily, B. and Reagans, R. (2003), “Managing knowledge in organizations: an integrative framework and review of emerging themes”, Management Science, Vol. 49 No. 4, pp. 571-582.

Bathelt, H. (2007), “Buzz-and-pipeline dynamics: towards a knowledge-based multiplier model of clusters”, Geography Compass, Vol. 1 No. 6, pp. 1282-1298.

Bathelt, H. and Glückler, J. (2011), The Relational Economy: Geographies of Knowing and Learning, Oxford University Press, Oxford.

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, Vol. 28 No. 1, pp. 31-56.

Berrill, K. (1964), Economic Development with Special Reference to East Asia: Proceedings of A Conference held by the International Economic Association, St Martin’s Press, New York, NY.

Bramanti, A. and Ratti, R. (1997), “The multi-faced dimensions of local development”, The Dynamics of Innovative Regions: The GREMI Approach, Ashgate, Aldershot, Brookfield, pp. 3-44.

Brannen, M.Y. (2004), “When Mickey loses face: recontextualization, semantic fit, and the semiotics of foreignness”, Academy of Management Review, Vol. 29 No. 4, pp. 593-616.

Bresnahan, T., Gambardella, A. and Saxenian, A. (2001), “Old economy’ inputs for ‘new economy’ outcomes: cluster formation in the new Silicon valleys”, Industrial and Corporate Change, Vol. 10 No. 4, pp. 835-860.

Cano-kollmann, M., Cantwell, J.A., Hannigan, T.J., Mudambi, R. and Song, J. (2016), “Knowledge connectivity: an agenda for innovation research in international business”, Journal of International Business Studies, Vol. 47 No. 3, pp. 255-262.

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

Cantwell, J.A. (1995), “The evolution of European industrial technology in the interwar period”, in Caron, F., Erker, P. and Fischer, W. (Eds), Innovations in the European Economy between the Wars, de Gruyter, Berlin.

Cantwell, J.A. (2017), “Innovation and international business”, Industry and Innovation, Vol. 24 No. 1, pp. 41-60.

Cantwell, J.A. and Mudambi, R. (2011), “Physical attraction and the geography of knowledge sourcing in multinational enterprises”, Global Strategy Journal, Vol. 1 Nos 3/4, pp. 206-232.

Cantwell, J.A. and Santangelo, G.D. (1999), “The frontier of international technology networks: sourcing abroad the most highly tacit capabilities”, Information Economics and Policy, Vol. 11 No. 1, pp. 101-123.

Cantwell, J.A. and Santangelo, G.D. (2000), “Capitalism, profits and innovation in the new techno-economic paradigm”, Journal of Evolutionary Economics, Vol. 10 Nos 1/2, pp. 131-157.

Cantwell, J.A. and Santangelo, G.D. (2002), “The new geography of corporate research in information and communications technology (ICT)”, Journal of Evolutionary Economics, Vol. 12 Nos 1/2, pp. 163-197.

Cantwell, J.A. and Vertova, G. (2004), “Historical evolution of technological diversification”, Research Policy, Vol. 33 No. 3, pp. 511-529.

Cantwell, J.A. and Zhang, F. (2013), “Do foreign-owned subsidiaries in China follow a distinctive pattern of technological knowledge sourcing? 外资企业是否在中国遵循特殊的科技能力发展道路?”, Management and Organization Review, Vol. 9 No. 3, pp. 489-512.

Chesbrough, H. (2003), “The logic of open innovation: managing intellectual property”, California Management Review, Vol. 45 No. 3, pp. 33-58.

Chung, W. and Kalnins, A. (2001), “Agglomeration effects and performance: a test of the Texas lodging industry”, Strategic Management Journal, Vol. 22 No. 10, pp. 969-988.

Coe, D.T. and Helpman, E. (1995), “International R&D spillovers”, European Economic Review, Vol. 39 No. 5, pp. 859-887.

Coe, N.M., Dicken, P., Hess, M. and Yeung, H.W. (2010), “Making connections: global production networks and world city networks”, Global Networks, Vol. 10 No. 1, pp. 138-149.

Cooke, P. and Morgan, K. (1998), The Associational Economy: Firms, Regions, and Innovation, Oxford University Press, Oxford.

Feldman, M.P. and Audretsch, D.B. (1995), Science-Based Diversity, Specialization, Localized Competition and Innovation, WZB, Forschungsschwerpunkt Marktprozess und Unternehmensentwicklung (IIMV), Berlin.

Foss, N.J. and Pedersen, T. (2004), “Organizing knowledge processes in the multinational corporation: an introduction”, Journal of International Business Studies, Vol. 35 No. 5, pp. 340-349.

Fujita, M. and Thisse, J. (2013), Economics of Agglomeration: Cities, Industrial Location, and Globalization, Cambridge University Press, Cambridge.

Gereffi, G., Humphrey, J. and Sturgeon, T. (2005), “The governance of global value chains”, Review of International Political Economy, Vol. 12 No. 1, pp. 78-104.

Giuliani, E. (2013), “Network dynamics in regional clusters: evidence from Chile”, Research Policy, Vol. 42 No. 8, pp. 1406-1419.

Goerzen, A., Asmussen, C. and Nielson, B. (2013), “Global cities and multinational enterprise location strategy”, Journal of International Business Studies, Vol. 44, pp. 427-450.

Granovetter, M. (1985), “Economic action and social structure: the problem of embeddedness”, American Journal of Sociology, Vol. 91 No. 3, pp. 481-510.

Grant, R. (2002), “The knowledge-based view of the firm”, in Choo, C. and Bontis, N. (Eds), The Strategic Management of Intellectual Capital and Organizational Knowledge, Oxford University Press, New York, NY, p. 133.

Humphrey, J. and Schmitz, H. (2002), “How does insertion in global value chains affect upgrading in industrial clusters?”, Regional Studies, Vol. 36 No. 9, pp. 1017-1027.

Jaffe, A.B. and Trajtenberg, M. (1996), “Flows of knowledge from universities and federal laboratories: modeling the flow of patent citations over time and across institutional and geographic boundaries”, Proceedings of the National Academy of Sciences of the United States of America, Vol. 93 No. 23, pp. 12671-12677.

Jaffe, A.B., Trajtenberg, M. and Henderson, R. (1993), “Geographic localization of knowledge spillovers as evidenced by patent citations”, The Quarterly Journal of Economics, Vol. 108 No. 3, pp. 577-598.

Lamorgese, A.R. and Ottaviano, I. (2006), “Intercity interactions: evidence from the US”, 2006 Meeting Papers Citeseer.

Langlois, R.N. and Robertson, P.L. (1995), Firms, Markets, and Economic Change: A Dynamic Theory of Business Institutions, Routledge, London.

Lazer, D. and Friedman, A. (2007), “The network structure of exploration and exploitation”, Administrative Science Quarterly, Vol. 52 No. 4, pp. 667-694.

Lippman, S.A. and Rumelt, R.P. (1982), “Uncertain imitability: an analysis of interfirm differences in efficiency under competition”, The Bell Journal of Economics, Vol. 13 No. 2, pp. 418-438.

Lychagin, S., Pinkse, J., Slade, M.E. and Van Reenen, J. (2016), “Spillovers in space: does geography matter?”, Journal of Industrial Economics, Vol. 64 No. 2, pp. 295-335.

Maillat, D. (1998), “Interactions between urban systems and localized productive systems: an approach to endogenous regional development in terms of innovative milieu”, European Planning Studies, Vol. 6 No. 2, pp. 117-129.

Malmberg, A. and Maskell, P. (2002), “The elusive concept of localization economies: towards a knowledge-based theory of spatial clustering”, Environment and Planning A, Vol. 34 No. 3, pp. 429-449.

Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic Change, Harvard University Press, Cambridge, MA.

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, Vol. 15 No. 1, pp. 5-21.

Park, W.G. (1995), “International R&D spillovers and OECD economic growth”, Economic Inquiry, Vol. 33 No. 4, pp. 571-591.

Perez-Aleman, P. (2011), “Collective learning in global diffusion: spreading quality standards in a developing country cluster”, Organization Science, Vol. 22 No. 1, pp. 173-189.

Polanyi, M. (1966), The Tacit Dimension, University of Chicago Press, Chicago.

Porter, M.E. (1996), “Competitive advantage, agglomeration economies, and regional policy”, International Regional Science Review, Vol. 19 Nos 1/2, pp. 85-90.

Santangelo, G. (2002), Innovation in Multinational Corporations in the Information Age, Edward Elgar, Chichester.

Scott, A.J. (1988), New Industrial Spaces: Flexible Production Organization and Regional Development in North America and Western Europe, Pion, Thousand Oaks, CA.

Shaver, J.M. and Flyer, F. (2000), “Agglomeration economies, firm heterogeneity, and foreign direct investment in the United States”, Strategic Management Journal, Vol. 21 No. 12, pp. 1175-1193.

Sorenson, O. (2005), “Social networks and industrial geography”, Entrepreneurships, the New Economy and Public Policy, Springer, Berlin, pp. 55-69.

Storper, M. and Venables, A.J. (2004), “Buzz: face-to-face contact and the urban economy”, Journal of Economic Geography, Vol. 4 No. 4, pp. 351-370.

Szulanski, G. (1996), “Exploring internal stickiness: impediments to the transfer of best practice within the firm”, Strategic Management Journal, Vol. 17 No. 2, pp. 27-43.

Szulanski, G. and Jensen, R.J. (2004), “Overcoming stickiness: an empirical investigation of the role of the template in the replication of organizational routines”, Managerial and Decision Economics, Vol. 25 No. 67, pp. 347-363.

Teece, D.J. (1977), “Technology transfer by multinational firms: the resource cost of transferring technological know-how”, The Economic Journal, Vol. 87 No. 346, pp. 242-261.

Uzzi, B. (1996), “The sources and consequences of embeddedness for the economic performance of organizations: the network effect”, American Sociological Review, Vol. 61 No. 4, pp. 674-698.

Uzzi, B. (1997), “Social structure and competition in interfirm networks: the paradox of embeddedness”, Administrative Science Quarterly, Vol. 42 No. 2, pp. 35-67.

Von Hippel, E. (1994), “Sticky information’ and the locus of problem solving: implications for innovation”, Management Science, Vol. 40 No. 4, pp. 429-439.

Wathne, K., Roos, J., V. and Krogh, G. (1996), Towards a Theory of Knowledge Transfer in a Cooperative Context, Sage Publications, London.

Further reading

ABS (2017), Australian Bureau of Statistics [Homepage of Australian Bureau of Statistics], available at: (accessed 6 October 2017).

Archibugi, D. and Pianta, M. (1992), “Specialization and size of technological activities in industrial countries: the analysis of patent data”, Research Policy, Vol. 21 No. 1, pp. 79-93.

CENSUS (2017a), Metropolitan and Micropolitan – Census Bureau 2015-last update [Homepage of United States Census Bureau], available at: (accessed 6 October 2017).

CENSUS (2017b), Census of India 2001-last update [Homepage of Office of the Registrar General & Census Commissioner, India], available at: (accessed 6 October 2017).

Cohen, W.M. and Levinthal, D.A. (1990), “Absorptive capacity: a new perspective on learning and innovation”, Administrative Science Quarterly, Vol. 35 No. 1, pp. 128-152.

Eskelinen, H., Hannibalsson, I., Malmberg, A., Maskell, P. and Vatne, E. (2002), Competitiveness, Localised Learning and Regional Development: specialization and Prosperity in Small Open Economies, Routledge, Abingdon.

Gupta, A.K. and Govindarajan, V. (2000), “Knowledge flows within multinational corporations”, Strategic Management Journal, Vol. 21 No. 4, pp. 473-496.

Hansen, M.T. (1999), “The search-transfer problem: the role of weak ties in sharing knowledge across organization subunits”, Administrative Science Quarterly, Vol. 44 No. 1, pp. 82-111.

Ipe, M. (2003), “Knowledge sharing in organizations: a conceptual framework”, Human Resource Development Review, Vol. 2 No. 4, pp. 337-359.

Jasimuddin, S.M. (2007), “Exploring knowledge transfer mechanisms: the case of a UK-based group within a high-tech global corporation”, International Journal of Information Management, Vol. 27 No. 4, pp. 294-300.

KOSTAT (2017), Statistics Korea 06/2017, 2017-last update, available at: (accessed 6 October 2017).

Kotabe, M., Dunlap-Hinkler, D., Parente, R. and Mishra, H.A. (2007), “Determinants of cross-national knowledge transfer and its effect on firm innovation”, Journal of International Business Studies, Vol. 38 No. 2, pp. 259-282.

Laursen, K., Reichstein, T. and Salter, A. (2011), “Exploring the effect of geographical distance and university quality on industry–university collaboration in the UK”, Regional Studies, Vol. 45 No. 4, pp. 507-523.

Lazer, D. and Friedman, A. (2007), “The network structure of exploration and exploitation”, Administrative Science Quarterly, Vol. 52 No. 4, pp. 667-694.

MCA (2017), National administrative division information query platform 04/2017, 2017-last update [Homepage of Ministry of Civil Affairs of People’s Republic of China], available at: (accessed 6 October 2017).

Miesing, P., Kriger, M.P. and Slough, N. (2007), “Towards a model of effective knowledge transfer within transnationals: the case of Chinese foreign invested enterprises”, The Journal of Technology Transfer, Vol. 32 Nos 1/2, pp. 109-122.

Mowery, D.C., Oxley, J.E. and Silverman, B.S. (1996), “Strategic alliances and interfirm knowledge transfer”, Strategic Management Journal, Vol. 17 No. 2, pp. 77-91.

Mu, J., Tang, F. and Maclachlan, D.L. (2010), “Absorptive and disseminative capacity: knowledge transfer in intra-organization networks”, Expert Systems with Applications, Vol. 37 No. 1, pp. 31-38.

Park, B.I. (2011), “Knowledge transfer capacity of multinational enterprises and technology acquisition in international joint ventures”, International Business Review, Vol. 20 No. 1, pp. 75-87.

Reagans, R. and Mcevily, B. (2003), “Network structure and knowledge transfer: the effects of cohesion and range”, Administrative Science Quarterly, Vol. 48 No. 2, pp. 240-267.

Schlegelmilch, B.B. and Chini, T.C. (2003), “Knowledge transfer between marketing functions in multinational companies: a conceptual model”, International Business Review, Vol. 12 No. 2, pp. 215-232.

Simonin, B.L. (2004), “An empirical investigation of the process of knowledge transfer in international strategic alliances”, Journal of International Business Studies, Vol. 35 No. 5, pp. 407-427.

Thomson, R. and Nelson, R.R. (1997). The internationalization of technology, 1874-1929: evidence from US, British, and German Patent Experience, Mimeo, New York, NY.

Tsai, W. (2001), “Knowledge transfer in intraorganizational networks: effects of network position and absorptive capacity on business unit innovation and performance”, Academy of Management Journal, Vol. 44 No. 5, pp. 996-1004.

Zahra, S.A. and George, G. (2002), “Absorptive capacity: a review, reconceptualization, and extension”, Academy of Management Review, Vol. 27 No. 2, pp. 185-203.

Corresponding author

John Cantwell can be contacted at: