Do global innovation networks influence the status of global value chains? Based on a patent cooperation network perspective

Abstract

Under the impact of economic globalization, the internationalization of production and trade has given rise to global value chains (GVCs) and global innovation networks (GINs); however, few studies have explored the link between these two networks. This paper constructs a global innovation cooperation network based on patent cooperation data and then analyzes the impact of different network characteristics on improving global value chain status. Empirical research shows that the characteristics of GINs significantly affect GVC status; the higher the eigenvector centrality and network connectivity, the higher the country’s GVC status. Moreover, a higher degree of constraint is associated with a lower GVC status for the country concerned. The heterogeneity analysis further reveals apparent differences in the impact of GINs on different countries, manifesting as a catalytic effect in developed countries and as a disincentive in emerging economies; the effect is less apparent in developing countries. This paper provides theoretical support for the development of technological innovation activities and provides a reference for government departments to formulate policies on cross-border technological innovation cooperation.

Introduction

Over the past few decades, economic globalization has affected the world in an unstoppable trend. In this process, international trade has been the main channel of economic globalization. Naturally, cost and multinationals have become the main drivers of economic globalization. As a result, the pattern of international trade is gradually changing, driven by both multinationals and costs (Wang et al., 2016); firms are beginning to outsource activities that were previously internally handled while maintaining in-house activities for which they have core skills. Against this backdrop, Gereffi and Kaplinsky (2001) developed the concept of global value chains (GVCs) with the increasing geographical dispersion of all activities (including design, production, marketing, logistics, distribution, and sales) undertaken by firms to bring products to market, from conception to end-use. This concept has also been explained as the so-called “vertical specialization” (Hummels et al., 2001), “trade in tasks” (Grossman and Rossi-Hansberg, 2008), “production fragmentation,” “production sharing” (Amiti, 2005; Johnson and Noguera, 2012), or “the second great unbundling” (Baldwin, 2011).

Following the internationalization of production and trade activities, research and development (R&D) activities became progressively international. To adapt products and processes to the needs of local markets, firms started to locate R&D activities to support their international production and distribution activities. Correspondingly, local firms began to support their innovation activities by tapping into foreign knowledge, technology, and human capital, giving rise to global innovation networks (GINs). There has been a recent trend toward open innovation where companies open their innovation processes to external partners and outsource their innovation activities. As global trade and technological advances increase competitiveness, product life cycles are dramatically shortened, forcing companies to innovate faster and develop products and services more efficiently. However, the convergence of different technologies has increased the demand for interdisciplinary, cross-border, and cross-industry research, along with the cost and risk of innovation. This situation makes it unlikely that innovation can be achieved by a single company; therefore, companies are increasingly looking for complementary partners to access different technologies and knowledge.

Against this backdrop, GVCs and GINs are emerging as core network systems that are complementary and interdependent in the wave of economic globalization. Surprisingly, there is little empirical evidence on the relationship between GINs and GVCs, although several studies confirming the impact of innovation on GVC upgrading. This raises the question: since technological innovation can contribute to the upgrading of GVCs, does a country’s integration into a global innovation network lead to an upgrading of its GVC position? And what are the mechanisms that cause this effect? According to our research, the impact of cross-border innovation cooperation on the position of GVCs is not as simple as direct promotion. It is a typical fact that in the early stages of China’s development, due to the massive import of external technologies, many foreign companies realized “technological feedback” and thus occupied the high end of the value chain of many Chinese industries. Not only that, but the academic community is not unanimous on this issue. Some researchers argue that the impact of global innovation cooperation on whole value chains is positive. GINs facilitate the collection and integration of global innovation resources and their optimal allocation, which drives the dynamic upgrading of a given economy’s GVC division of labor. For example, Pietrobelli and Rabellotti (2011) argue that constructing a global innovation system (GIS) facilitates international knowledge exchange and R&D collaboration opportunities for developing country firms participating in GVC activities. They argue that GIS also plays a vital role in promoting firms’ ascendancy in the GVC division of labor. Furthermore, Sachwald (2013) finds that partnerships for R&D collaboration in GINs positively affect their innovation performance and productivity. This impact is not influenced by geographical proximity but rather by the innovation partner’s degree of excellence. Some scholars have expressed the opposite view, arguing that the subordination of developed and developing countries in a GVC system and GIS dominated by developed countries leads to unequal international innovation partnerships and, thus, an unequal value derived from innovation cooperation. For example, Dedrick et al. (2010) find that Apple captures a large share of value from the iPod innovation in the distribution of the financial value of innovation in iPod and laptop global supply chains. In contrast, laptop manufacturers capture a smaller share of value from personal computer innovation. This claim is also complemented by some literature arguing that differences from a global collaborative innovation are related to the absorptive capacity of countries and cannot be generalized (De Prato et al., 2013; Fransen and Helmsing, 2017; Navaretti and Venables, 2004). In addition to the mixed findings described above, there are other shortcomings in the existing research on the impact of innovation on value chain upgrading. First, there is a preponderance of case studies and a lack of empirical studies. For example, Su et al. (2020) selected three Chinese technology firms as case studies to investigate the impact of the innovation capabilities of firms in emerging economies on the upstream degree of GVC; Park and Gachukia (2021) used a Kenyan fruit and vegetable exporter as a case study to investigate the impact of the local innovation system on GVCs. Although case studies can provide more detailed and exhaustive conclusions, they also have the disadvantage of insufficient external validity. Second, the limitations of the research perspective. As mentioned above, the globalization of trade is accompanied by the globalization of innovation, and most of the existing studies discuss the impact of technological innovation cooperation in a certain region or field on the upgrading of GVCs but neglect the globalization context of innovation and trade.

In summary, to verify the relationship between innovation and GVCs in the context of globalization of innovation and trade division of labor, this paper further examines the impact of a country’s integration into the global innovation network on the position of GVCs based on the measurement of GVC location. Specifically, the paper seeks to answer the following questions:

1. What is the relationship between GINs and GVCs? What impact does a country’s integration into a GIN have on a country’s GVC position? How does this impact arise?

2. Different countries are integrated into GINs to different degrees; what are these implications for the status of GVCs? Specifically, do the characteristics of GINs have a different impact on GVC status in different countries? If so, what are the reasons for this phenomenon?

Based on these observations, this paper adopts a research framework suggesting that trade networks influence the upgrading of global value chains. It aims to establish a global innovation network using patent cooperation data and constructs an econometric model with network centrality, network linkage strength, and network structural holes as key explanatory variables for empirical analysis. Through this approach, the paper aims to investigate the impact of a country’s integration into the global innovation network on the status of global value chains.

The main contributions of this paper are as follows: First, it situates innovation and trade globalization as the backdrop, expanding the research perspective to the international level and considering innovation behavior comprehensively as a unified entity. This approach addresses the limitations of existing research perspectives by providing a more comprehensive view. Second, the paper combines social network analysis with traditional econometric models to analyze the impact of global innovation network characteristics on the status of global value chains from the perspective of patent cooperation networks. This enriches the theoretical content and methodological approaches to the research problem, addressing deficiencies in the existing literature’s empirical research. Lastly, the paper provides network-level evidence to address the issue of a country’s lock-in at the low end of the value chain. This not only offers theoretical support for a country’s technological innovation activities but also provides a reference for government departments in formulating policies for cross-border technological innovation cooperation.

Literature review

Global innovation network

In the context of global economic integration, the processes of knowledge transfer and technology diffusion are becoming increasingly complex. Global innovation networks have emerged as an effective means of facilitating knowledge transfer and technology diffusion among diverse stakeholders. The concept of a global innovation network was first proposed by Ernst (2009) as a network that integrates decentralized engineering applications, product development, and R&D activities across organizational and regional boundaries. Subsequent scholars have offered various interpretations of this concept. For example, Liu et al. (2013) suggest that the global innovation network represents an innovation approach that transcends the limitations of knowledge stickiness. It involves the management and application of analytical knowledge and synthesis knowledge at both the “global-local” levels. Necoechea-Mondragón et al. (2017) highlight three dimensions that characterize global innovation networks: geographic breadth, network scope (internal or external), and outcomes (innovations), which are global, networked, and innovative. From a comprehensive standpoint, the global innovation network represents an organizational approach to integrating and leveraging dispersed innovation resources, thereby unlocking the value of innovation through a network structure amidst the backdrop of economic globalization and knowledge diffusion. With the emergence of global innovation networks, the traditional paradigm of technology procurement and trade has waned in prominence, making way for deeper collaborative models along the value chain. This includes practices such as global innovation development, global innovation production, and global technological cooperation, which have gained widespread adoption.

Global value chain

Since the 1960s, the division of labor in the value chain has become more prevalent, and Porter (1985) first introduced the concept of “value chain,” which considers the enterprise as a collection of production, distribution, transportation and auxiliary production processes that generate profits through a series of distinct but interrelated value-added activities, thus forming a value chain. This concept focuses on micro-analysis and lacks a global perspective for comparing international strategic advantages. In response, Gereffi (1995) proposed the concept of “global value chain,” highlighting that it represents a cross-business network organization that integrates production, sales, and recycling processes to create and realize the value of a specific commodity or service on a global scale. Subsequently, numerous scholars have interpreted and refined this concept. However, the most widely accepted definition comes from the United Nations Industrial Development Organization (UNIDO), defining the global value chain as a global cross-enterprise network organization that connects production, distribution, and recycling processes to realize the value of a commodity or service worldwide. This involves the production and distribution of raw materials, transportation, semi-finished products, finished products, as well as consumption and recycling processes.

Upgrading of global value chains

As GVCs involve a range of value-added activities and profit-sharing, the position of firms and even countries in GVCs becomes critical. Currently, the four levels of upgrading classification identified by Schmitz and Humphrey (2000) - process, product, function and chain upgrading - are widely used for upgrading GVCs. Among them, process and product upgrades are fundamental. Process upgrading is often defined as the innovation of new processes, new equipment, and new management and organizational methods that can significantly improve the productivity of a firm and even the efficiency of the industry. Product upgrading begins with the development of new products, and the role of technology is reflected in product design and manufacturing (Gereffi and Kaplinsky, 2001). Consequently, research on upgrading GVCs from an innovation perspective has garnered significant attention. For instance, Sinkovics et al. (2018) highlight that the absence of innovation capacity poses a significant constraint for emerging economies in generating greater value within GVCs. Through a case study of 10 manufacturing firms, Szalavetz (2019) suggests that innovation improves the upstream degree by increasing the productivity of firms downstream in the GVC.

Relationship between GINs and GVCs

Since the globalization of the economy, companies have continued to spread their innovation activities worldwide, and the innovation process has been increasingly broken down into more elaborate and technically complex stages and tasks. In this process, the relationship between GVCs and GINs has become increasingly close, forming a new pattern of cross-fertilization and synergistic development. This new knowledge production system is also gradually penetrating and mapping into the globalization strategies of multinational companies, which have a critical impact on the flow of knowledge and the distribution of added value in the GVC. Based on the available literature, international innovation cooperation can contribute to an economy’s rising position in the GVC division of labor through channels such as knowledge flow, industry linkage, and human resource flow effects.

First, international innovation cooperation affects the position of economies in the GVC division of labor through the knowledge flow effect. Multinational companies’ R&D centers and technology partnerships are increasingly embedded in GVC systems and innovation networks. As a result, global innovation cooperation networks provide an environment that supports the knowledge sharing and information matching required for inter-country innovation cooperation. Furthermore, global innovation cooperation networks have become an effective channel for cross-border technology transfer and a driving force for developing multinational companies’ GVCs. The resulting technological innovation and advancement have a significant role in driving firms up the GVC division of labor (Cooke, 2003; Sachwald, 2013). Second, international innovation cooperation affects the position of economies in the GVC division of labor through linkage effects. Traditionally, GVCs are seen as “physical” transfer vehicles for cross-border goods and services, while GINs are seen as international transfer vehicles for intangible or immaterial assets. The increasing degree of international innovation cooperation has facilitated the coupling and interaction between GVCs and GINs, strengthening the links between upstream and downstream industrial chains and functional links in the international division of the production system (Oecd, 2017; Pietrobelli and Rabellotti, 2009). This strengthening has helped realize the value-added knowledge in GVCs through industrial input-output linkage effects. Third, human capital flows from global innovation cooperation networks also affect the GVC division of the labor position of economies. In international innovation cooperation, the two-way flow of knowledge in the form of R&D personnel has a significant human capital accumulation effect. It may also have a corresponding knowledge spillover effect in the form of cross-firm training exchanges or job-hopping experiences of R&D personnel, which may have an important impact on the evolution of the GVC division of labor in an economy (Crestanello and Tattara, 2011).

Research hypothesis

Within the current framework of social network analysis, common indicators encompass network centrality, network connection strength, and network heterogeneity. This paper adopts the research paradigm of trade networks to delve into the theoretical mechanisms through which characteristics of the global innovation network influence the status of global value chains.

Network centrality, a key characteristic that measures the position of each node within a network, serves as an important reference for assessing a country’s centrality in a global innovation network and its frequency of interaction with other nations. In general, the closer a country is to the center of the network, the greater its capacity to absorb advantageous resources. Within a global innovation network, each country enhances its technological accumulation by exploiting the technological advantages of others, thus contributing to the upgrading of the global value chain. Countries with higher centrality are more likely to occupy top positions in the global value chain, given their extensive collaborations with multiple countries and accumulation of superior foreign technologies. Furthermore, the theory of technology absorption posits that the input and output of technology are influenced by the endowment of local innovation resources. Countries with high centrality typically possess more comprehensive innovation resources and superior technologies, affording them greater opportunities to attract other countries for innovation cooperation. This facilitates the realization of complementary technological advantages and enhances their position in the global value chain. Hence, a country with higher network centrality enjoys a significant advantage in the global innovation network, exerting greater control and influence over the entire network. By enhancing its position within the network, the country can bolster its capacity to gain a competitive advantage. Accordingly, we propose the following hypothesis:

Hypothesis 1: The higher the network degree centrality is, the higher a country’s GVC division of labor position will be.

The strength of network ties measures the strength of the ties between nodes in a network due to their interdependence; the nodes’ frequency, degree of interaction, and closeness can demonstrate this relationship. The strength of linkages reflects the amount of innovation content in a country’s global innovation cooperation network, focusing more on the depth of innovation cooperation. Specifically, strong innovation cooperation network links indicate that the country has conducted many technological innovations in global innovation cooperation. When the number of innovation cooperation between innovation subjects increases, the research between the cooperation parties improves, making it easier to develop trusting relationships between the research subjects and produce effects, such as knowledge flow, talent exchange, industrial cooperation, and upgrading; thus, the GVC status will naturally be enhanced. Therefore, we propose the following hypothesis:

Hypothesis 2: The stronger the network linkages are, the higher a country’s GVC division of labor position will be.

Network structural holes are non-redundant links between two actors. The higher a node’s position in the network core, the more structural holes it may have. Structural holes or weak ties in the network manifest as network heterogeneity. In a GIN, there may be redundant ties between nodes, resulting in loose connections between them. Therefore, the degree to which others constrain a country in the global innovation cooperation network can be reflected through the network structure’s hole-limit system. Burt argues that occupying a structural hole allows actors to access and control diverse innovation resources, thereby enhancing their competitiveness. The higher the structural hole restriction of a node in an innovation cooperation network is, the more closed the country is and the less able it is to use the structural hole. If a node’s structural hole restriction in an innovation cooperation network is low, it can connect with different actors, thus gaining more information and maintaining its superior characteristics without being eliminated or squeezed out. Thus, countries occupying structural hole positions tend to improve their GVC position by matching heterogeneous knowledge and complementary resources in GINs through knowledge flows. The wealth of knowledge and technology generated by their structural hole positions attracts a constant influx of talent from other countries. Based on this, this study proposes the following hypothesis:

Hypothesis 3: The less restrictive the network structure is, the higher a country’s GVC division of labor position will be.

Global innovation networks

Few empirical studies examine global innovation cooperation networks, and the primary means of measuring innovation cooperation is using R&D data, patent data, and innovation surveys, which are the most intuitive indication of innovation cooperation; however, data are often unavailable and limited by dependence on member state participation. In contrast, scholars are increasingly using patent data to study different aspects of innovation due to its reliability and accessibility. On the one hand, patent data is a manifestation of innovation itself; on the other hand, it contains a wealth of information, such as the nature of the patent and the participants. Thus, patent data was finally chosen to construct a global innovation cooperation network.

In the selection of the study sample, this study is based on the World Intellectual Property Organization (WIPO) innovation index ranking1. After eliminating blank data, 46 countries were retained as sample countries2. This approach includes countries that account for at least 80.00% of the global gross domestic product (GDP) and are all critical participating members of the GVC. Moreover, the R&D investment and the number of PCT international patents in these countries are high, accounting for more than 70%. Furthermore, the selected sample countries cover regions such as Asia, Europe, North America, Oceania, and Africa, including developed and developing countries. The selected sample is representative, and the cooperative innovation network formed among them can comprehensively reflect the overall development of transnational technical cooperation.

In selecting innovation relationships, it is generally accepted that co-invention patents are a better indicator of inter-country cooperation than co-patenting (Balconi et al., 2004). Therefore, this study uses patent data between two or more inventors from different countries to measure innovation cooperation between two countries and thus construct an innovation cooperation network. We construct the unweighted innovation cooperation matrix $A_t$ and the weighted innovation cooperation matrix $W_t$ for period $t$. For element $a_{\mathit{ij}}^t$ in the matrix $A_t$, $a_{\mathit{ij}}^t$ = 1 if there is innovation cooperation between country $i$ and country $j$; otherwise, $a_{\mathit{ij}}^t$ = 0. For the element in the $b_{\mathit{ij}}^t$ matrix $W_{\mathit{ij}}^t$, the number of co-inventions between countries i and j is used to represent the weights between nodes $i$ and $j$, while $W_{\mathit{ij}}^t$ is divided by the maximum value; thus, all 0 ≤ $W_{\mathit{ij}}^t$ ≤ 1.

Overall network characterization

The study of global innovation cooperation network relations allows for the examination of the scale, frequency, and degree of closeness of cooperation between countries. Additionally, it enables the analysis of the position and radiation range of countries within the network. The more edges a node has that radiate outwards, the more collaborative relationships the country has established. The thickness of the edges connected to the node indicates the closeness of innovation collaboration between the two countries; the thicker the edges are, the closer the countries are. Network density is one of the indicators of the overall closeness of the network. This study uses Ucinet 6.0 software to calculate the overall network density of 46 countries and regions worldwide from 2000 to 2018 (due to the limitation of the database, patent cooperation data can only be found up to 2018). We use the innovation cooperation network density matrix $A_t$, which in turn reflects the closeness of innovation cooperation exchanges. Calculations show that the density of global innovation cooperation networks was relatively low in 2000, at around 7.0116, and showed a general upward trend from 2000 to 2018, with the network density reaching its highest level in 2018, at around 17.4338. Figure 1 shows that innovation partnerships between individual countries have become closer, and the frequency and scale of mutual collaborative innovation have significantly grown. For example, China and Malta were on the periphery of the innovation cooperation network in 2000; however, by 2018, Malta had initiated new direct innovation cooperation with seven countries, while China was at the heart of the global innovation cooperation network, having established frequent and close cooperation with most countries (especially the United States).

Fig. 1 [Images not available. See PDF.]

The global innovation cooperation network.

Analysis of individual indicators

Degree centrality

Network centrality measures are generally a degree of centrality, intermediate centrality, proximity centrality, and eigenvector centrality. The degree of centrality can measure the extent to which a country is directly or indirectly linked to other countries in a GIN. This indicator provides a visual representation of the number of countries with which it has innovation partnerships and its ability to deploy knowledge flows in the network. The degree of centrality is calculated as follows:

1

$$de{g}_i={\sum }_j{a}_{ij}$$

The kernel density estimates for numerical centrality from 2000 to 2018 are shown in Fig. 2; in 2000, the kernel density estimation curve showed a right-skewed distribution, and most countries did not establish many international innovation partnerships. In 2010, the kernel density estimation curve showed a double-peaked tip, indicating that half of the countries had started extensively developing international partnerships. Meanwhile, in 2018, it showed an apparent left-skewed distribution characteristic, and the peak reached the maximum that year. This estimate indicates that most countries already have a wide range of international innovation partners, with only a small number of countries having a more concentrated distribution of innovation partners. Overall, the estimated nuclear density curve from 2000–2018 gradually transitions from a right- to a left-skewed distribution, showing that most countries did not develop many international innovation partnerships at the beginning of economic globalization. However, as the years progressed, countries increasingly began establishing innovation partnerships, eventually peaking in 2018 with an expanding range of international innovation cooperation.

Fig. 2 [Images not available. See PDF.]

Kernel density estimation diagram for degree centrality.

Eigenvector centrality

Eigenvector centrality examines the type of structure of the entire network, i.e., the centrality of a node is influenced by the centrality of other nodes connected to it. The more important the connected neighboring nodes are, the more important the node is. Eigenvector centrality is calculated as follows:

2

$$e_i={\lambda }^{-1}\sum _j{A}_t{e}_j$$Where $\lambda$ denotes the maximum eigenvalue of the adjacency matrix, and $e_j$ is the eigenvector of the adjacency matrix. The kernel density estimates of eigenvector centrality from 2000–2018 are shown in Fig. 3. The peaks show a decreasing trend year after year during 2000–2018, and the kernel density estimation curves show a significant right-skewed distribution. Furthermore, the GIN eigenvector centrality peaks at around 0.02 and then rapidly declines, showing a significant trailing phenomenon. This phenomenon illustrates the wide disparity in the position of countries in the GIN, with most countries having established more extensive international innovation partnerships; however, they often struggle to improve their position due to the constraints of their environment and the innovation power of their partners. This situation has improved since its peak in 2015, but most countries are still locked at the bottom of the global innovation cooperation network.

Fig. 3 [Images not available. See PDF.]

Kernel density estimation plot for eigenvector centrality.

Network connectivity

Network connectivity can be measured in terms of node strength, which reflects the strength of relationships between nodes and is used to reflect the strength of a country’s connections in the GIN. The scale of innovation cooperation between countries is given as a weight on the link between them, with higher weights indicating higher nodal strength and more significant influence of the country. The strength of the network linkages is calculated as follows:

3

$$\mathit{St}{r}_i=\sum _j{w}_{\mathit{ij}}$$

The kernel density estimates of point intensity from 2000 to 2018 are shown in Fig. 4. The kernel density estimation curve shows a significant right-skewed distribution from 2000 to 2018. The GIN point intensity peaks at approximately 200 and then rapidly declines to a long and flat tail at 1000, which indicates that only a few countries have a higher intensity of international innovation cooperation. From the peak, the kernel density of point intensity reached its highest level in 2000 and was close to its lowest level after 2015. Furthermore, the peak is slightly shifted to the right, indicating that the degree of innovation cooperation among countries has deepened in recent years; however, the gap between different countries is still significant.

Fig. 4 [Images not available. See PDF.]

Kernel density estimation diagram for point strength.

Network structure hole

The network structure hole index consists of effective size, efficiency, constraint, and hierarchy. Among them, a constraint is the most important measure of network structure hole, reflecting the heterogeneity of the GIN, and it is a combination of direct and indirect constraints on node i in the network. If another node $j$ generates many ties with other nodes, it generates more significant constraints on the node $i$. Subsequently, the point $i$ is subject to a pointwise constraint denoted by

4

$${\mathrm{Con}}_i=\sum _{\mathit{j}}{\left(p_{\mathit{ij}}+{\sum }_q{p}_{\mathit{iq}}{m}_{\mathit{jq}}\right)}^2$$Here, q represents the set of neighboring nodes of nodes i and j, and $p_{\mathit{ij}}$ represents the effort spent by node i on node j. Overall, a downward trend is shown in the restriction regimes of countries worldwide, indicating that as GINs develop, countries are becoming more liberal in their international innovation cooperation activities. 5. Econometric Specifications and Data

Model specification

To further validate the impact of structural characteristics within global innovation networks on global value chains, this paper selects network centrality, network connectedness, and network structural holes as explanatory variables. The model is constructed as follows:

5

$$\begin{array}{c}GVC\_Positio{n}_{it}={\alpha }_1\mathit{de}{\mathit{g}}_{i\mathrm{t}}+{\alpha }_{2}st{r}_{it}+{\alpha }_{3}co{n}_{it}+{\alpha }_{4}pgd{p}_{it}+{\alpha }_{5}ope{n}_{it}+{\alpha }_{6}po{p}_{it}\hfill \\ +{\alpha }_{7}fc{r}_{it}+{\alpha }_{8}r{d}_{it}+Countr{y}_i+Yea{r}_t+{\epsilon }_{it}\hfill \end{array}$$where $i$ represents the country and $t$ represents time. $\mathit{GVC\_Position}$ represents the GVC position; $\mathit{Country}$ represents country-fixed effect, $\mathit{Year}$ represents time-fixed effect, deg represents network centrality of a country, and $\mathit{str}$ represents network linkage strength of a country. $\mathit{con}$ represents network heterogeneity of a country. $\mathit{pgdp},\mathit{open},\mathit{pop},\mathit{fcr},\mathit{rd}$ are a series of control variables used in this study, representing economic development level, trade openness, human resources, physical base, R&D investment, respectively. Finally, ${\epsilon }_{\mathit{it}}$ represents random disturbance terms that obey the standard distribution.

Variables

Explained variable

GVC position

The measurement of global value chains was initially introduced by Hummels et al. who characterized GVCs as a form of vertical specialization division of labor. This concept encapsulates the trend of countries engaging in heightened trade of intermediate goods, leading to extended vertical trade chains spanning multiple nations. Within this framework, each country specializes in one or a few production segments of goods production. Hummels et al. proposed the following formula based on this definition:

6

$$\mathit{VS}{S}_r=\frac{V{S}_r}{X_r}$$$\mathit{VS}{S}_r$ represents the level of vertical specialization; $V{S}_r$ denotes the value of imported product inputs in the industry’s exports; and $X_r$ signifies the value of the industry’s exports.

The essence of the vertical specialization level formula lies in the notion that a higher proportion of imported value within a country’s exports situates the country further downstream in the division of the labor chain and signifies a greater reliance on foreign technology. In light of this concept, Koopman and other scholars have endeavored to refine the measurement of GVCs. They posit that a country’s participation in the division of labor in GVCs is both supply-side (as evidenced by the foreign value added embedded in its exports) and demand-side (as reflected in its exports of indirect value added). They further argue that a more supply-side role implies that the country is situated at the upstream end of the GVCs. The country is situated at the upstream end of the division of labor in GVCs. Consequently, the country’s GVC status can be expressed in the following form:

7

$${GVC\_Position}_{ir}=Ln\left(1+\frac{I{V}_{ir}}{E_{ir}}\right)-Ln\left(1+\frac{F{V}_{ir}}{E_{ir}}\right)$$

Equation (7) represents the formula for calculating the GVC status index, where $r$ represents an industry, $i$ represents a country, $I{V}_{\mathit{ir}}$ represents the domestic value added of indirect exports, $F{V}_{\mathit{ir}}$ represents the foreign value added of exports, and $E_{\mathit{ir}}$ represents total exports. The larger the ${\mathit{GVC\_Position}}_{\mathit{ir}}$ is, the larger the country’s status in the GVC will be. This approach effectively addresses issues concerning measurement accuracy arising from double counting of trade data, trans-exporting, re-exporting, and similar factors. Consequently, it offers a more precise depiction of a country’s involvement and status level within global value chains.

Explanatory variables

This study selected three core explanatory variables: network centrality (deg), network link strength (str), and network structure hole (con). Network centrality indicates a country’s position and centrality in the global innovation cooperation network and is initially measured by the degree of centrality. Network strength indicates how closely a country cooperates with other countries in the global innovation cooperation network and is measured by the point strength indicator. Network structure hole indicates the extent to which a country is constrained by other countries in the global innovation cooperation network.

Control variables

Drawing on Shin et al. (2009), He et al. (2021), Zhang et al. (2022), and others, this study selects five control variables: the level of economic development ($\mathit{pgdp}$), trade openness ($\mathit{open}$), human resource endowment ($\mathit{pop}$), material base ($\mathit{fcr}$), and level of science and technology ($\mathit{rd}$). The level of economic development is measured by a country’s GDP, which reflects differences in the level of economic development between countries. The higher the level of a country’s economy is, the more factors of production it tends to hold and the higher its position in the GVC is. Trade openness is measured using total exports and import trade as a share of GDP. Trade networks are currently the focus of research on upgrading GVCs, and a large body of literature demonstrates the positive impact of trade networks on GVCs. In general, the more open trade is, the more liberal the institutional operating environment is, and the more conducive it is to innovation activities. This conduciveness, in turn, enhances a country’s technological level and innovation capacity and promotes economic growth and industrial upgrading. Human resource endowment is measured using a country’s population size, indicating intrinsic value chain upgrading, which is complementary to technological development. The physical base is expressed as fixed capital as a share of GDP. Fixed capital refers to productive capital in the form of machinery, equipment, plants, and other essential labor means. It is the basis for a country’s integration into and advancement in GVCs; the greater the physical capital is, the higher its position in the GVC division of labor. The level of science and technology is measured by the share of R&D investment in GDP. Science and technology is the driving force behind an economy’s sustainable development and is measured by the share of R&D investment in GDP. The more a country invests in R&D, the higher its level of science and technology and the quality of its industries, and the stronger is its international competitiveness, which in turn helps improve its division of labor position.

Data

This study used three primary data sources. The first part of the global patent cooperation information data comes from WIPO’s PCT international patent database. This database provides completed international patent information and is reviewed by multiple patent offices according to uniform standards, making it internationally uniform and comparable. For example, we searched for patent collaboration data between China and the US and used the search formula “AD: ([01.01.2000 TO 31.12.2000] AND IADC:(CN) AND IADC:(UA))” to obtain all joint invention patents between these countries in 2000. This formula resulted in a 46 × 46 matrix of innovation relationships between countries from 2000 to 2018. The data for the second part of the GVC status measurement are obtained from the TiVA database of OECD3, while the data for the third part of the control variables are taken from the World Bank4. Table 1 presents the descriptive statistics of this study’s explanatory and control variables.

Table 1. Descriptive statistics.

Variable	Obs	Mean	SD	Min	Max
$\mathit{GVC}\mathrm{\_}\mathit{Position}$	874	0.005	0.138	−0.453	0.280
$\mathit{pgdp}$	874	1627099	3192347	9566.793	2.04*107
$\mathit{open}$	874	97.093	69.363	19.559	437.327
$\mathit{pop}$	874	9.43*107	2.6*108	281205	1.40*109
$\mathit{fcr}$	874	23.127	4.812	10.770	44.518
$\mathit{rd}$	874	1.579	1.005	0.0423	4.796

Empirical results

This section conducted a baseline regression analysis and robustness tests based on the entire sample.

Baseline analysis

This study uses a fixed-effects model for the empirical study, and Table 2 reports the results of the baseline regressions. Models (1)–(3) are the results from regressions on the three core explanatory variables, and the regression results from adding control variables are presented in Models (4)–(6). Table 2 shows that the effect of innovation network linkage strength on GVC status is significantly positive. Conversely, the effect of the innovation network structure hole-limit system on GVC status is significantly negative. Furthermore, the effect of innovation network degree centrality on GVC status is not significant when no control variables are added, and its effect on GVC status is significantly negative after controlling for the control variables. The core explanatory variables of the three innovation network characteristics, controlling for other variables, show the following characteristics. For every 1% increase in the centrality of a country’s innovation network degree, its position in the division of labor in GVCs decreases by 0.9%. Furthermore, for every 1% increase in the strength of a country’s innovation network linkages, its position in the division of labor in GVCs increases by 16.2%. For every 1% decrease in the hole-limit system of a country’s innovation network structure, its position in the division of labor in GVCs increases by 2.2%. Finally, for every 1% decrease in a country’s innovation network structure hole-limit system, its position in the division of labor in GVCs increases by 2.2%. Thus, Hypotheses 2 and 3 are proven.

Table 2. Baseline regression.

	GVC position
	(1)	(2)	(3)	(4)	(5)	(6)
deg	0.005			−0.009***
	(0.004)			(0.003)
str		0.220**			0.162***
		(0.097)			(0.051)
con			−0.029**			−0.022***
			(0.014)			(0.006)
Control	NO	NO	NO	YES	YES	YES
Country FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
N	874	874	874	874	874	874
R2	0.275	0.273	0.288	0.581	0.579	0.582

Standard errors clustered at the country level are reported in parentheses. ** and *** denote significance at 5% and 1% level, respectively.

The empirical results show that the deeper the degree of international innovation cooperation is, the more it helps a country to discover and acquire advanced technology in the process of innovation cooperation, improve its technology level and develop economies of scale in the process, and achieve the rise of its position in the GVC. Additionally, the value of the network structure hole-limit system indicates that the less other countries restrict a country’s international innovation cooperation, the higher its position in the GVC will be. Generally, the country occupying the structural hole position is often at the hub of knowledge dissemination and exchange. It has a significant advantage in acquiring key technologies and the latest knowledge, thus promoting a country’s position in the GVC.

Further analysis

The previous section proved Hypotheses 2 and 3, but Hypothesis 1 yielded the opposite result; that is, extensive international innovation partnerships are not conducive to a country’s GVC position. This finding seems counter-intuitive but ignores a key attribute, namely, the innovation capacity of countries. The innovation capacity here is said to be multifaceted, including the underlying conditions for carrying out innovation, the purpose of innovation, and the nature of innovation. This study draws on the idea of calculating eigenvector centrality to describe this capability. The more innovative cooperation and the broader the innovation target, the stronger the innovation capability of the country and the more solid the foundation for carrying out innovation. The innovation capacity of different countries significantly varies, and countries with strong innovation capacity often cooperate to bring greater benefits than those with weak innovation capacity. For example, countries downstream in GVCs tend to need more basic industry innovation, which often does not make sense for countries with strong innovation capabilities upstream in GVCs. It is conceivable that a US multinational corporation would work with multinational corporations from countries with strong innovation capabilities to achieve a breakthrough in processor performance in cell phones.

Based on this, this study introduces the feature vector centrality indicator to consider the centrality of a country in the global innovation cooperation network in terms of the importance and extensiveness of innovation cooperation targets. Therefore, Hypothesis 4 is proposed as follows:

Hypothesis 4: The greater the feature vector centrality of a country’s network is, the higher its GVC status will be.

Table 3 presents the regression results for further analysis. The symbol deg* represents the sign of eigenvector centrality in the model. Columns (1) and (4) show that the effect of innovation network feature vector centrality on GVC status is significantly positive, and this effect remains constant after adding control variables. This result indicates that broader innovation cooperation with more innovative countries can significantly contribute to their GVC position; thus, Hypothesis 4 is proven.

Table 3. Further analysis.

	GVC position
	(1)	(2)	(3)	(4)	(5)	(6)
deg*	0.203***			0.161***
	(0.044)			(0.054)
str		0.220**			0.162***
		(0.097)			(0.051)
con			−0.029**			−0.022***
			(0.014)			(0.006)
Control	NO	NO	NO	YES	YES	YES
Country FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
N	874	874	874	874	874	874
R2	0.293	0.274	0.288	0.579	0.579	0.582

Standard errors clustered at the country level are reported in parentheses. ** and *** denote significance at 5% and 1% level, respectively. deg* represents the eigenvector centrality.

Robustness analysis

In a previous analysis, we added a series of control variables and controlled for time and country effects to ensure the robustness of our main findings; however, endogeneity issues must still be addressed. A straightforward and effective way to address the endogeneity problem is to select an appropriate instrumental variable. An effective instrumental variable requires that it is highly correlated with the endogenous variable itself and does not correlate with the error. In this study, endogenous variables lagged by one year as instrumental variables, as Wooldridge (2015) suggested in his book on econometrics. The rationale is that the GIN characteristics in the current period have little impact on the GVC position in the previous period. In addition, Based on the approach of Xu et al. (2023), we introduce new instrumental variables to measure a country’s GVC position to address the endogeneity of the GVC position. As stated in the introduction, the integration of the world market has led to the production chain of product offerings being completed by different countries; therefore, the value added of intermediate goods better reflects the distribution of the benefits of international trade. This study chooses domestic value-added, which is currently internationally recognized, to measure a country’s GVC position instead (Johnson, 2014), with the original data coming from the OECD’s TiVA database.

Table 4 reports the results of robustness tests, indicating that each variable’s positive and negative coefficient signs and significance remain unchanged. Furthermore, network feature vector centrality and network linkage strength are positively related to the GVC division of labor status, and the network structure hole-limit regime is negatively related to the GVC division of labor status. These results prove the robustness of the conclusions of this study.

Table 4. Robustness analysis.

	One year lag			DVA
	(1)	(2)	(3)		(4)	(5)	(6)
deg*	0.225***				3.084***
	(0.048)				(0.957)
str		0.234***				2.546***
		(0.048)				(0.850)
con			−0.021***				−0.248***
			(0.006)				(0.078)
Control	YES	YES	YES		YES	YES	YES
Country FE	YES	YES	YES		YES	YES	YES
Year FE	YES	YES	YES		YES	YES	YES
N	828	828	828		874	874	874
R2	0.512	0.513	0.527		0.877	0.874	0.873

Standard errors clustered at the country level are reported in parentheses. ** and *** denote significance at 5% and 1% level, respectively. deg* represents the eigenvector centrality.

Heterogeneity

Innovation cooperation promotes the development and diffusion of new technologies. Thanks to Romer (1990)’s north-south technology diffusion theory and Raup (1963)’s doctrine of the technological latecomer advantage, many scholars and governments have advocated that technology latecomers directly adopt the technology introduction path to absorb technology from advanced countries and imitate the innovation to promote their industrial development. As a result, most underdeveloped countries rely more on technology importation than innovation cooperation with developed countries. Furthermore, the global innovation cooperation network in this study suggests that most marginal countries have gradually developed international innovation cooperation only in recent years, leading to a weaker innovation base in these countries. International innovation cooperation has become increasingly widespread following the realization that even less developed countries must strive to innovate; however, the process, nature, and capabilities of innovation significantly differ between developed and developing countries. It would be inappropriate to apply this study’s conclusions to all countries indiscriminately. For example, the key science and technology organizations (e.g., universities, R&D laboratories, and research institutes) where innovation cooperation occurs may be imperfect or even missing in some countries. The need for innovation cooperation differs between developed and developing countries, which often need innovation in the most basic parts to upgrade their weak industrial structures. Finally, it is important to emphasize that there are also growing differences in innovation and technological capabilities within developing countries. It is clear that in countries like China and India, the basis on which innovation cooperation occurs is getting closer to that of developed countries and even reaching world-class standards in some sectors (Schmitz and Strambach, 2008). A few developing countries (often referred to as emerging countries) have begun the transition from capturing economic benefits to building innovation capacity. Although the depth and breadth of this transition are not yet known, and there is little research on it, it has attracted the attention of governments and the media (Altenburg et al., 2008). Therefore, the conclusions drawn in this study’s baseline regressions are likely inapplicable to a few emerging and developing countries. It would be significant for all types of countries if the differences and widening gaps between them could be identified and explained in this way.

Table 5 reports the results of the heterogeneity analysis. The findings for developed countries are similar to the baseline regressions in that extensive innovation partnerships do not significantly impact the ability of developed countries to improve their GVC position and that developed countries should frequently collaborate with countries with innovation power. Notably, the structural hole restriction is not significant for all three types of countries, indicating that the global innovation cooperation network is not network heterogeneous for the same type of countries. Furthermore, the same type of countries are subject to similar restrictions by other countries, and they do not have a controlling role. The previous benchmark analysis indicated that the less a country is affected by the degree of restriction of other countries, the greater its position in the GVC division of labor. Therefore, based on the logic of the exclusion method, since this restrictive effect is not reflected among the same type of countries, it is likely to occur in developed and underdeveloped countries. The previous empirical results also argue that a country’s GVC position can be enhanced by being more restrictive or less restricted by other countries’ innovation cooperation. This restriction is mainly reflected in the restrictions developed countries impose on emerging and developing countries, consistent with the view and facts presented below.

Table 5. Heterogeneity.

	Developed countries				Emerging countries				Developing countries
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
deg	0.002				−0.011				0.011*
	(0.003)				(0.007)				(0.001)
deg*		0.156***				−0.348*				0.506
		(0.052)				(0.178)				(0.918)
str			0.162***				−0.522***				0.452
			(0.051)				(0.185)				(0.818)
con				0.002				0.000				(0.005)
				(0.012)				−0.015				−0.015
Control	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Country FE	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
N	456	456	456	456	152	152	152	152	95	95	95	95
R2	0.638	0.645	0.642	0.638	0.773	0.776	0.783	0.769	0.884	0.879	0.879	0.879

Standard errors clustered at the country level are reported in parentheses. *, **, and *** denote significance at 10%, 5%, and 1% level, respectively. deg* represents the eigenvector centrality.

In the regressions for emerging countries, we obtain very different results from those for developed countries, suggesting that for emerging countries, relying on autonomous innovation rather than seeking cooperation with foreign countries seems to be a more reasonable choice. In recent years, emerging countries like China and India have undertaken a lot of international innovation cooperation; however, their GVC position has never improved. This study’s data show that as of 2018, China participated in 2,894 international patent collaborations, with a constraint system of 0.454; India participated in 1,123 international patent collaborations, with a constraint system of 0.532. In contrast, Denmark participated in only 433 international patent collaborations in 2018 but with a restriction system of 0.374, and France participated in 1,968 international patent collaborations with a restriction system of only 0.385. The reasons for this situation are manifold; the analysis in this study, combined with the historical literature, suggests that the main reason comes from the blockage of critical technologies by developed countries. Humphrey and Schmitz (2002) point out that GVC leaders often support product and process upgrades, but they may block functional upgrades to protect their core competencies. According to Romer’s (1990) north-south technology diffusion theory and Raup’s (1963) doctrine of the technological latecomer advantage, emerging countries tend to rely more on technology importation and imitation innovation. This reliance leads to less cooperation in high-tech innovation from foreign countries and a relatively weak innovation base for cutting-edge technologies. For example, Zhang and Yue (2019) find that the US is trying to prevent China from rising in the GVC using technological blocking, thereby containing China’s rise and putting the country in a “low-end lock” dilemma in the GVC. Second, as some developing countries gradually obtain low-cost advantages, emerging countries (i.e., China) no longer have the original labor advantage and must extensively conduct technological innovation for industrial upgrading to promote their GVC status. However, considering that technologies from developed countries may not apply to emerging countries, this situation can force emerging countries to make greater efforts to develop innovation and achieve complementarity between the introduction of foreign technologies and domestic technological innovation (Fu et al. 2011). Therefore, under the influence of both internal demand and external pressure, there is a growing call for independent innovation in emerging countries, especially in countries like China and India. Several empirical studies by scholars have reported similar views. Yuan and Sun (2017) found that the role of autonomous R&D in upgrading China’s manufacturing industry is facilitative, while cooperative R&D among firms has the opposite effect. Hu (2022) noted that the effect of autonomous innovation on the structural upgrading of China’s manufacturing industry is positive, while imitative innovation plays a suppressive role; this feature is particularly evident in high-technology industries. Moreover, Tang et al. (2021) found that for Chinese firms, autonomous innovation is more beneficial to firm innovation performance, and collaborative innovation only brings financial performance to Chinese firms.

Finally, the findings in Columns (9)–(11) show that the extensive development of international innovation partnerships in developing countries positively impacts enhancing their GVC position. Such results also show that international information exchange and cooperation become more critical for developing countries that are severely technologically backward and have weak innovation ecosystems. The extraterritorial impact of the innovation process is particularly salient given that these countries rarely achieve frontier innovation and must import most of their knowledge and technology. A stream of innovation systems literature analyzes the impact of foreign firms on innovation and learning processes in developing countries (Navaretti and Venables, 2004; Wagner, 2007). These scholars found that for firms in developing countries, international information exchange with more countries could provide new markets for their products and play a more critical role in acquiring knowledge and enhancing learning and innovation. Furthermore, international information exchange can also play a vital role in acquiring knowledge and enhancing learning and innovation. As previously mentioned, the innovation ecosystem in less developed countries is weak, and some key science and technology organizations in innovation cooperation, such as universities, R&D laboratories, and research institutes, are inadequate or even missing. Thus, developing countries require organizations that help with technology diffusion and diffusion (Pietrobelli and Rabellotti, 2009).

Conclusions and policy recommendations

With the development of economic globalization, GINs and GVCs have become increasingly inseparable. This study constructs a global innovation cooperation network based on patent international cooperation data and attempts to analyze the effects on the status of GVCs through network characteristics. The empirical results show a significant effect of GIN characteristics on GVC status enhancement. Among them, the effect of network degree centrality on GVC status enhancement is negative; that is, extensive international innovation partnerships are not conducive to GVC status enhancement. However, the effect of network Eigenvectors centrality on GVC status enhancement is positive. This result indicates that when considering the innovation capability of international innovation partnership partners, more extensive international innovation partnerships can significantly enhance a country’s GVC status. The effect of network linkage strength is positive, indicating that deeper international innovation cooperation relationships can contribute to improving GVC status. The empirical results of the network structure hole-limit regime indicate that the fewer restrictions a country receives in international innovation cooperation, the higher its GVC status will be. To further ensure the reliability of our findings, we conducted separate empirical studies for different types of countries. We found that our basic conclusions did not change in developed countries, where deeper integration and improved status in GINs contributed to the improvement of developed countries’ GVC position; however, the empirical study on emerging countries yields the exact opposite conclusion. The results seem to encourage emerging countries to strengthen their own innovation rather than continue engaging in extensive international innovation partnerships. The reasons for the emergence of this phenomenon are discussed in detail in the heterogeneity analysis in Section 7. Finally, an empirical study of developing countries indicates that if developing countries want to improve their position in GVCs, they should enhance innovation exchanges with other countries and integrate more into international R&D efforts.

Acknowledgements

The authors would like to acknowledge support from the Major Program of National Philosophy and Social Science Foundation of China [NO. 22&ZD162]; Major Social Science Foundation of Zhejiang, China [NO. 22QNYC14ZD]. This work also supported by the characteristic & preponderant discipline of key construction universities in Zhejiang province (Zhejiang Gongshang University- Statistics) and Collaborative Innovation Center of Statistical Data Engineering Technology & Application.

Author contributions

S.X did the main programming work as well as part of the paper writing; A.X was mainly responsible for most of the paper writing; G.L was mainly responsible for the improvement of the paper and data collection; and M.S was mainly responsible for the revision and embellishment of the paper.

Data availability

The datasets generated and analyzed in the current study are available from the corresponding author upon reasonable request.

Ethical approval

Ethical assessment is not required prior to conducting the research reported in this article, as the present study does not have experiments on human subjects and animals, and does not contain any sensitive and private information.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.

Competing interests

The authors declare no competing interests.

1

Website: https://www.wipo.int/global_innovation_index/en/2022/

2

The selected countries are Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovenia, Spain, Sweden, Switzerland, Turkey, United Kingdom, United States, Brazil, Bulgaria, China, Croatia, Cyprus, India, Malta, Philippines, Romania, Russian Federation, Saudi Arabia, Singapore, and Thailand.

3

Trade in Value Added - OECD: https://www.oecd.org/industry/ind/measuring-trade-in-value-added.htm

4

World Bank Open Data | Data: https://data.worldbank.org.cn/

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1057/s41599-024-03413-7.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Altenburg, T; Schmitz, H; Stamm, A. Breakthrough? China’s and India’s transition from production to innovation. World Dev.; 2008; 36, 2 pp. 325-344. [DOI: https://dx.doi.org/10.1016/j.worlddev.2007.06.011]

Amiti, M. Location of vertically linked industries: agglomeration versus comparative advantage. Eur. Economic Rev.; 2005; 49, 4 pp. 809-832. [DOI: https://dx.doi.org/10.1016/j.euroecorev.2002.12.001]

Balconi, M; Breschi, S; Lissoni, F. Networks of inventors and the role of academia: an exploration of Italian patent data. Res. policy; 2004; 33, 1 pp. 127-145. [DOI: https://dx.doi.org/10.1016/S0048-7333(03)00108-2]

Baldwin RE (2011) 21st Century Regionalism: Filling the gap between 21st century trade and 20th century trade rules. Available at SSRN 1869845

Cooke P (2003) Regional innovation and learning systems, clusters, and local and global value chains. Innov Clusters Interreg. Comp. 28–51

Crestanello, P; Tattara, G. Industrial clusters and the governance of the global value chain: The Romania–Veneto network in footwear and clothing. Regional Stud.; 2011; 45, 2 pp. 187-203. [DOI: https://dx.doi.org/10.1080/00343401003596299]

De Prato G, Nepelski D, & Simon J (2013) Asia in the global ICT innovation network: Dancing with the tigers. Elsevier

Dedrick, J; Kraemer, KL; Linden, G. Who profits from innovation in global value chains?: a study of the iPod and notebook PCs. Ind. Corp. change; 2010; 19, 1 pp. 81-116. [DOI: https://dx.doi.org/10.1093/icc/dtp032]

Ernst D (2009) A New Geography of Knowledge in the Electronics Industry? Asia’s Role in Global Innovation Networks. Policy Stud. 54. https://doi.org/10.2139/ssrn.2742923

Fransen, J; Helmsing, A. Absorptive capacity as a mediator: Innovation of handicraft exporters in Yogyakarta, Indonesia. Tijdschr. voor economische en. Soc. geografie; 2017; 108, 6 pp. 737-752. [DOI: https://dx.doi.org/10.1111/tesg.12212]

Fu, X; Pietrobelli, C; Soete, L. The role of foreign technology and indigenous innovation in the emerging economies: technological change and catching-up. World Dev.; 2011; 39, 7 pp. 1204-1212. [DOI: https://dx.doi.org/10.1016/j.worlddev.2010.05.009]

Gereffi G (1995) Global production systems and third world development. In B Stallings (Ed.), Global Change, Regional Response: The New International Context of Development (pp. 100–142). Cambridge University Press. https://doi.org/10.1017/CBO9781139174336.004

Gereffi, G; Kaplinsky, R. Introduction: Globalisation, value chains and development. IDS Bull.; 2001; 32, 3 pp. 1-8. [DOI: https://dx.doi.org/10.1111/j.1759-5436.2001.mp32003001.x]

Grossman, GM; Rossi-Hansberg, E. Trading tasks: A simple theory of offshoring. Am. Econ. Rev.; 2008; 98, 5 pp. 1978-1997. [DOI: https://dx.doi.org/10.1257/aer.98.5.1978]

He, SB; Yao, HL; Ji, Z. Direct and Indirect Effects of Business Environment on BRI Countries’ Global Value Chain Upgrading [Article]. Int. J. Environ. Res. Public Health; 2021; 18, 23 12492. [DOI: https://dx.doi.org/10.3390/ijerph182312492] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34886217][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8656886]

Hu JS (2022) Research on the Influence of Technological Progress Path on the Upgrading of Manufacturing Structure. https://doi.org/10.27209/d.cnki.glniu.2022.000474

Hummels, D; Ishii, J; Yi, K-M. The nature and growth of vertical specialization in world trade. J. Int. Econ.; 2001; 54, 1 pp. 75-96. [DOI: https://dx.doi.org/10.1016/S0022-1996(00)00093-3]

Humphrey, J; Schmitz, H. How does insertion in global value chains affect upgrading in industrial clusters?. Reg. Stud.; 2002; 36, 9 pp. 1017-1027. [DOI: https://dx.doi.org/10.1080/0034340022000022198]

Johnson, RC. Five facts about value-added exports and implications for macroeconomics and trade research. J. Econ. Perspect.; 2014; 28, 2 pp. 119-142. [DOI: https://dx.doi.org/10.1257/jep.28.2.119]

Johnson, RC; Noguera, G. Accounting for intermediates: Production sharing and trade in value added. J. Int. Econ.; 2012; 86, 2 pp. 224-236. [DOI: https://dx.doi.org/10.1016/j.jinteco.2011.10.003]

Liu J, Chaminade C, & Asheim B (2013) The Geography and Structure of Global Innovation Networks: A Knowledge Base Perspective. Eur. Planning Stud. 21. https://doi.org/10.1080/09654313.2012.755842

Navaretti GB, & Venables AJ (2004) Multinational firms in the world economy. Princeton University Press

Necoechea-Mondragón, H; Pineda-Domínguez, D; Pérez-Reveles, L; Soto-Flores, R. Critical factors for participation in global innovation networks. Empirical evidence from the Mexican nanotechnology sector. Technol. Forecast. Soc. Change; 2017; 114, pp. 293-312. [DOI: https://dx.doi.org/10.1016/j.techfore.2016.08.027]

Oecd (2017) The links between global value chains and global innovation networks: An exploration. OECD Science, Technology and Industry Policy Papers

Park, E; Gachukia, MK. The Role of the Local Innovation System for Inclusive Upgrading in the Global Value Chain: The Case of KenyaGAP in the Kenyan Horticultural Sector. Eur. J. Dev. Res.; 2021; 33, 3 pp. 578-603. [DOI: https://dx.doi.org/10.1057/s41287-020-00285-w]

Pietrobelli C, & Rabellotti R (2009) The global dimension of innovation systems: linking innovation systems and global value chains. In Handbook of innovation systems and developing countries. Edward Elgar Publishing

Pietrobelli, C; Rabellotti, R. Global Value Chains Meet Innovation Systems: Are There Learning Opportunities for Developing Countries?. World Dev.; 2011; 39, 7 pp. 1261-1269. [DOI: https://dx.doi.org/10.1016/j.worlddev.2010.05.013]

Porter, M. Competitive Advantage; 1985; New York, Free press:

Raup PM (1963) Economic Backwardness in Historical Perspective. In: JSTOR

Romer, PM. Endogenous technological change. J. Political Econ.; 1990; 98, 5, Part 2 pp. S71-S102. [DOI: https://dx.doi.org/10.1086/261725]

Sachwald F (2013) The development of global innovation networks. Global Innovation Networks, High Level Economic Expert Group Innovation for Growth i4g, European Commission, DG Research and Innovation, Luxembourg: Publications Office of the European Union, 143

Schmitz H, & Humphrey JF (2000) Governance and Upgrading: Linking Industrial Cluster and Global Value Chain Research

Schmitz H, & Strambach S (2008) The organisational decomposition of the innovation process: what does it mean for the global distribution of innovation activities?

Shin, N; Kraemer, KL; Dedrick, J. R&D, Value Chain Location and Firm Performance in the Global Electronics Industry [Article]. Ind. Innov.; 2009; 16, 3 pp. 315-330. [DOI: https://dx.doi.org/10.1080/13662710902923867]

Sinkovics N, Hoque S, & Sinkovics R (2018) Supplier Strategies and Routines for Capability Development: Implications for Upgrading. J. Int. Manag. 24. https://doi.org/10.1016/j.intman.2018.04.005

Su, F; Khan, Z; Lew, YK; Park, BI; Choksy, US. Internationalization of Chinese SMEs: The role of networks and global value chains. Brq-Bus. Res. Q.; 2020; 23, 2 pp. 141-158. [DOI: https://dx.doi.org/10.1177/2340944420916339]

Szalavetz, A. Digitalisation, automation and upgrading in global value chains - factory economy actors versus lead companies. Post-Communist Econ.; 2019; 31, pp. 1-25. [DOI: https://dx.doi.org/10.1080/14631377.2019.1578584]

Tang MP, Guo YZ, Xu QR, Mao SY (2021) A Comparative Study of the Impact of Innovation Model Choice on Firm Performance - Independent Innovation or Collaborative Innovation. Accounting Friends. (17):43–50

Wagner, J. Exports and productivity: A survey of the evidence from firm‐level data. World Econ.; 2007; 30, 1 pp. 60-82.MathSciNet ID: 618646[DOI: https://dx.doi.org/10.1111/j.1467-9701.2007.00872.x]

Wang Z, Wei SJ, Yu X, & Zhu K (2016) Characterizing global value chains. 19th Annual Conference on Global Economic Analysis, Washington DC

Wooldridge JM (2015) Introductory econometrics: A modern approach. Cengage learning

Yuan, YJ; Sun, DM. A Comparison of FDI Technology Spillover, Independent R&D and Cooperative R&D - Based on the Perspective of Manufacturing Technology Upgrading. Sci. Res.; 2017; 35, 9 14.

Xu, A; Wang, W; Zhu, Y. Does smart city pilot policy reduce CO2 emissions from industrial firms? Insights from China. J. Innov. Knowl.; 2023; 8, 3 [DOI: https://dx.doi.org/10.1016/j.jik.2023.100367]

Zhang, HN; Chen, ZY; Zhang, DL; Huo, JY; Li, YF. The Impact of a Complex Computer Numerical Model Based on Economic Cooperation on the Value Chain of International Trade Facilitation [Article]. Wirel. Commun. Mob. Comput.; 2022; 2022, [DOI: https://dx.doi.org/10.1155/2022/2997825]

Zhang, MZ; Yue, S. Perspective of U.S.-China Trade Friction Based on the Perspective of Global Value Chain. J. South China Normal Univ.; 2019; 51, 2 87.