Introduction

The Global Value Chain (GVC) has redefined the dynamics of international trade by shifting the basic unit of its trade from final products to value-added production stages. In this framework, the position of manufacturing industry has already become an important indicator for measuring a country’s competitiveness. Technological innovation is the primary catalyst for improving the GVC position. Therefore, it has emerged as a strategic priority for governments worldwide. However, as domestic innovation resources fail to meet the demands of increasingly sophisticated markets, accessing international innovation resources has emerged as an important strategy for economies to narrow technological inadequacy and accelerate R&D impact. In this context, multi-national corporations (MNCs) establish transnational R&D centers to co-create innovations with global partners. These centers utilize overseas innovation resources to create global innovation networks (GINs). While deepening integration into GINs offers developing economies opportunities to advance their participation in GVC and achieve technological progress, this process simultaneously exposes them to risks of “solidification” in GVC position and “immiserizing growth”. In the absence of strong indigenous innovation capacity and supportive institutional frameworks, these countries may remain trapped in the lower segments of GVCs, struggling to achieve structural upgrading. Therefore, this study systematically examines the impact of GINs embeddedness on manufacturing GVC and explores the underlying mechanisms in depth. This approach helps deepen our understanding of the technological development path of manufacturing and expands the research boundaries of GVC governance theory. Meanwhile, it provides theoretical support and practical references for developing economies to formulate policies that promote technological upgrading, optimize the allocation of innovative resources and achieve high-quality growth, thus possessing important practical significance.

Literature review

For many years, the academic community has investigated how technological innovation improves the GVC position of the manufacturing industries1. Zhang et al.2 pointed out that manufacturing enterprises could ascend from low-value-added production links to high-value-added innovation and design links through technological innovation. Bian and Fan1 also believed that innovation promotes the structural upgrading of the manufacturing industry and drives the development of high-tech industries and service industries. Given the insufficient supply of technological innovation resources within an economy, MNCs (Multi national companies) began to rely on overseas innovation resources. They carried out technological innovation with their partners and formed GINs gradually and organically3. This trend gradually captured the attention of scholars. They explored the motivations underlying the formation of the GINs, the network characteristics and how evolutionary processes worked from a network perspective4,5. Wolfe6 held the view that acquiring technological resources and talent were the primary driving force for MNCs to build the GINs. De Prato7 analyzed the structure and evolution of GINs by using patent-based data of international co-inventions and found that a country’s position in networks has a strong influence on the intensity of its collaboration with other countries.

The accelerating globalization of innovation collaboration shifted scholarly attention toward investigating the effect of GINs on the innovation capabilities of enterprises8,9,10. One perspective held that enterprises no longer rely on internal innovation resources solely for technological innovation, but enhanced their competitiveness and technological sophistication by assimilating external innovation resources through global collaboration. Wu Yingna11 investigated the influence of GINs on the innovation output of semiconductor firms and found that there was significant positive impact between them. MNCs transferred innovative technologies to various countries and regions around the world through global innovation collaboration and promoted technological innovation capabilities among local enterprises12. Channels by which this occurred included technological spillover, technology transfer and knowledge sharing of MNCs10. Innovation collaboration facilitates the aggregation of highly educated and skilled talents, providing more sufficient human resource reserves for the innovation of R&D institutions and accelerating the generation of innovative outcomes13. However, an alternative perspective challenged this viewpoint, arguing that innovation collaboration was also accompanied by certain challenges and risks. Developed economies might leverage their technological advantages in such partnerships to extract high price premiums, thereby retaining substantial value and profits14. Moreover, developing economies, constrained by their comparatively weaker technological innovation capacities, were prone to forming subordinate-type global innovation partnerships15. Such asymmetrical collaboration may result in detrimental or inequitable effects on enhancing corporate technological capabilities and core competitiveness16.

In the increasingly frequent global network structures shaped by cross-border R&D investments, national innovation performance is jointly influenced by external network linkages and absorptive capacity10. Chi proposed that collaborative innovation capability exerts a significant influence on firm’s collaborative innovation efforts and organizational performance17. Islam18 also believed that the heterogeneous impacts of global innovation collaboration between developed and developing economies are likely due to the fact that developing economies lack strong independent innovation capabilities and supportive institutional frameworks. Intellectual property rights (IPR) protection serves as the foundation for promoting global innovation collaboration. In Yang’s opinion19 robust IPR protection mechanisms incentivized enterprises and research institutions to pursue innovation while safeguarding their research outcomes, thereby fostering multinational collaboration and technology transfer. For developed economies, elevated IPR protection ensured their innovative outputs receive adequate economic returns while reinforcing the manufacturing GVC position20. For developing economies, the strengthening of IPR protection represented both an opportunity and a challenge21. A relatively high IPR protection helped to stimulate domestic innovation and attract foreign investment22. However, owing to the long-term dependence of developing economies on technology introduction from developed economies and the enhancement of their competitiveness through means such as imitation, absorption, and recreation, strengthened IPR protection might impose certain limitations on the technology absorption and innovation capabilities of developing economies23,24. Developing economies might find that stringent IPR protection levels could hinder the enhancement of their manufacturing position in the GVC25. Hwang argued that Once a country’s economic development level surpasses a certain threshold, the overall social welfare effect of patent protection transforms into an innovation incentive effect26. Therefore, it is necessary to examine whether the IPR protection exerts differential promoting effects on economies with varying levels of economic development and technological absorptive capacity.

It is evident that current literature primarily focuses on intermediate outcomes such as technological spillover effects, patent performance, and firms’ innovation outputs, while offering limited systematic analysis of how GINs, through the dimensions of embedding breadth and embedding depth, affect value capture capabilities in the GVCs of the manufacturing sector. Moreover, studies on the underlying mechanisms remain insufficient with most existing research remaining at the level of descriptive correlations. It remains unclear through which specific pathways manufacturing firms achieve an upgrade in their GVC position. There is also a lack of identification and empirical testing of heterogeneity across countries or regions in terms of institutional foundations and technological capabilities. Collectively, these unresolved issues provide a valuable research avenue for this study.

Accordingly, this study constructs GINs based on patent cooperation data of 50 economies from 2007 to 2021, characterizing the structural features and empirically examining the impact on manufacturing GVC positions. The innovations and marginal contributions of this paper are as follows: Firstly, this study investigates the impact of two distinct embeddedness metrics “embedding breadth” and “embedding depth” on manufacturing GVC positions from the perspective of multilateral collaborative. “Embedding breadth” measures the scope of global innovation collaboration undertaken by economies, whereas “embedding depth” indicates the total intensity of their innovation collaboration. Secondly, this study systematically analyzes the impact of both these GINs embeddedness metrics on manufacturing GVC positions with a particular focus on the heterogeneous effects between developed and developing economies and absorptive capacity. This approach not only expands the research frontier on GVC determinants but also provides insights into the trade effects of GINs embeddedness across economies at different absorptive capacity. Finally, based on the different research conclusions, this paper suggests the need to identify the effect of IPR protection in moderating heterogeneous absorptive capacity in the context of how the GINs embeddedness drives manufacturing GVC position. The findings contribute to theoretical explanations of the impact of global innovation collaboration among economies, and provide a reference for formulating national(regional) innovation collaboration strategies.

The remaining sections of this paper are structured as follows: section “Theoretical mechanisms and research hypothesis” conducts theoretical mechanism and research hypotheses; section “Research design” elaborates on the research design; section “Empirical results and analysis” analyzes the empirical results; section “Further analysis” performs heterogeneity tests and mechanism analysis; and section“ Conclusion and policy implications” concludes with core findings and provides corresponding policy recommendations.

Theoretical mechanisms and research hypothesis

The direct impact of GINs embeddedness on manufacturing value chains

As global competition intensifies and technological complexity increases, it becomes difficult for individual firms to independently accomplish high-end manufacturing or systemic innovation. GINs through cross-border technological collaboration, enable manufacturing enterprises to access diverse and heterogeneous technological resources, thereby effectively promoting their positions in GVCs11.

Higher embedding breadth indicates that an economy’s innovation partners are widely distributed, enabling to acquire innovation resources across broader technological domains. The acquisition of extensive and high-quality innovative resources enable economies with higher embedding breadth to achieve optimal allocation of scientific and technological innovation resources27. This process significantly unleashes technological productivity and innovation-driven forces, helping to enhance enterprises’ innovation capabilities and technological levels, thereby strengthening their competitiveness in GVCs28.

Higher embedding depth indicates that an economy engages in more frequent innovative collaborations and maintains closer cooperative relationships with other economies within the network. The closer innovation collaboration relationship breaks the constraint of low-risk resistance ability within the innovation entities, showing a stronger willingness to innovate29, which promotes the generation of innovation achievements more effectively30, helping enterprises establish industry leadership and consolidate their position in GVCs. Based on these insights, this study proposes the following hypothesis:

Hypothesis 1: Both embedding breadth and embedding depth in GINs can promote the manufacturing GVC position positively.

The indirect impact of GINs embeddedness on manufacturing value chains

Human capital structure upgrading

As a critical advanced production factor, human capital plays a key role in manufacturing GVC upgrading, with improvements in its quality and structure closely aligned with advancement to higher stages of the value chain.

Higher embedding breadth indicates a wide distribution of innovation collaboration partners, which facilitates the attraction of more high-level innovative talent, thereby contributing to the optimization of the human capital structure31. The high-end segments of the manufacturing value chain, such as R&D design and brand marketing, typically require intensive human capital input. The upgrading of the human capital structure helps enterprises break free from their dependence on low value-added activities and achieve a transformation from manufacturing to intelligent manufacturing, thereby enhancing their position in GVC32.

Higher embedding depth indicates a greater frequency of an economy’s participation in global innovation collaboration. Through deep collaboration, R&D entities gain access to cutting-edge technologies and advanced management practices33. This enhances their capabilities to explore frontier technologies and continuously improves managerial expertise, thereby facilitating the upgrading of the human capital structure. The upgrading of the human capital structure not only enhances manufacturing firms’ capabilities in technology absorption and innovation, but also drives the intelligent transformation of production methods and the refinement of management practices, thereby enabling a shift from low-end to high-end segments of the value chain.

Technological progress

The internationalization of innovation serves as a key channel for international technology spillovers and generates technology diffusion effects that far exceed those stemming from domestic innovation34. These diffusion effects not only enhance manufacturing firms’ independent innovation capabilities and lead to greater technological advancement, but also promote the diversification of intermediate goods, thereby exerting a significant positive impact on manufacturing GVC upgrading.

Economies with higher embedding breadth in GINs are able to access diverse and heterogeneous innovation resources and benefit from greater opportunities for international technology spillovers, thereby facilitating technological progress35. Technological progress enhances the flexibility of manufacturing processes and the capability for modular design, thereby providing a solid foundation for the diversification of intermediate goods. This diversity of intermediate products offers manufacturing firms more entry points into GVCs and greater opportunities for value creation.

Economies with higher embedding depth in GINs possess numerous strategic partners. Leading parties form R&D institutions through technological strategic alliances to jointly create and share new technological knowledge, which accelerates the formation of international technology spillovers, thereby enhancing the technological capabilities and value-capturing abilities of both parties36. Technological progress not only enhances the value added of manufacturing products and improves production efficiency, but also strengthens firms’ technological leadership and their ability to integrate value chains, thereby promoting a comprehensive upgrading of their GVC position. Based on these insights, this study proposes the following hypothesis:

Hypothesis 2: Embedded in GINs promotes manufacturing GVC upgrading through the human capital structure upgrading and technological progress.

Heterogeneous effects of GINs embeddedness on manufacturing value chains

Developed economies are predominantly assumed as the role of technological leaders within the GINs, whereas developing economies gradually increase their own technological levels and innovation capabilities by focusing on achieving technology introduction, imitation and re-innovation primarily37. Such role differentiation shapes distinct collaboration patterns directly in GVC across different nations and regions38.

Developed economies usually possess strong innovation capabilities and advanced technological endowments in innovation collaborative activities. They are always at the core of GINs and dominate technological R&D as well as innovation activities by acting as technology providers and knowledge exporters39. Through technological collaborations with other economies, developed economies leverage their advanced technologies to charge relatively high price premiums, retain substantial value and profits and continuously enhance their own competitive advantages and technological innovation capabilities. It enables them to further consolidate their technological superiority and promote manufacturing GVC position40.

Although developing economies have successfully embedded themselves into the GINs dominated by developed economies, it does not mean that they have the same status as developed economies41. Owing to their relatively weak technological innovation capabilities, developing economies are prone to form a “subordinate” global innovation cooperation relationship, which leads to the “solidification” of their positions in GVC42. They face numerous difficulties in climbing up to high-value-added segments and may even experience “low-end lock-in” phenomenon. Based on these insights, this study proposes the following hypothesis:

Hypothesis 3: The higher the level of economic development, the stronger the promoting effect of GINs embeddedness on manufacturing GVC position.

Hypothesis 4: The closer an economy to the frontier technology distance, the stronger the promoting effect of GINs embeddedness on manufacturing GVC position.

The moderating effect of IPR protection

IPR protection provides institutional support for the effective functioning of GINs by reducing the risks of technological collaboration, safeguarding returns on innovation and enhancing institutional attractiveness. These mechanisms facilitate deeper integration of manufacturing enterprises into the GVC and enhance their ability to capture value43.

In the global innovation collaboration, high level of IPR protection can effectively facilitate technology transfer from developed economies and ensure the global dissemination and application of their core technologies. Strong IPR protection safeguards the interests of developed economies during technology transfer processes44, thereby encouraging them to engage in technological exchange and knowledge sharing with developing economies and emerging markets45. It not only brings economic returns to developed economies but also promotes the global dissemination of technology. Meanwhile, it expands the technological influence of developed economies through technology spillovers and enhancing their manufacturing positions in GVC.

Excessive intellectual property protection may adversely affect the position of manufacturing industries in developing countries within the GVCs46. Strict IPR regimes can restrict technology transfer and local innovation, hindering firms in developing countries from enhancing their technological capabilities and competitiveness through imitation, absorption, and re-innovation. This may limit their effective participation in higher value-added segments of the GVC and constrain industrial upgrading and value capture47, thereby negatively impacting their position in the manufacturing value chain. Based on these insights, this study proposes the following hypothesis:

Hypothesis 5: Strengthened IPR protection promotes the positive effect of developed economies’ GINs embeddedness on manufacturing GVC position, while inhibits developing economies’ manufacturing GVC upgrading.

Research design

Model setting

Based on the preceding theoretical analysis and hypotheses, this study constructs the following empirical model in order to identify the impact of GINs embeddedness on manufacturing GVC position.

$${\text{GVC}}{{\text{P}}_{ik{\text{t}}}}={\alpha _0}+{\beta _0}{\text{Tne}}{{\text{t}}_{it}}+{\gamma _0}{\text{X}}+{\delta _i}+{\delta _k}+{\delta _t}+{\delta _{ikt}}$$
(1)

where \({\text{GVC}}{{\text{P}}_{ikt}}\) denotes a country’s GVC position, with subscript i denoting country, k denoting industry, and t denoting year. \({\text{Tne}}{{\text{t}}_{it}}\) indicates the network structure characteristics of economy i in year t, measuring by two indicators, embedding breadth and embedding depth respectively. \({\text{X}}\) is the control variable, \({\delta _i},~{\delta _{k,~}}~{\delta _t}~\)represents the country, industry and time fixed effects, \({\delta _{ikt}}\) represents the random error term.

Description of the data

Dependent variable

The dependent variable is represented by GVC position. It is proposed by Koopman48 based on GVC value-added decomposition and serves as an indicator of GVC position. This index is constructed by comparing a economies’ industry-level value-added exports of intermediate goods (used in other economies’ production and export of final products) with its value-added imports of intermediate goods (used in domestic production and export of final products), which reflects the industry’s value capture capability in GVC.

$${\text{GVC}}{{\text{P}}_{{\text{ik}}}}={\text{ln}}\left( {1+\frac{{{\text{I}}{{\text{V}}_{ik}}}}{{{{\text{E}}_{ik}}}}} \right) - {\text{ln}}\left( {1+\frac{{{\text{F}}{{\text{V}}_{ik}}}}{{{{\text{E}}_{ik}}}}} \right)$$
(2)

Among them, i represents the economy, k represents the industry. \({\text{I}}{{\text{V}}_{{\text{ik}}}}\) represents the domestic added value of k industry in country i when it is exported as an intermediate product, \({\text{F}}{{\text{V}}_{ik}}\) represents the foreign added value in the export of k industry in country i as a final product, and \({{\text{E}}_{ik}}\) represents the export value of k industry in country i counted by added value.

Independent variable

The explanatory variables in this paper are embedding breadth (degree centrality) and embedding depth (degree intensity). Two indicators are used to measure the GINs embeddedness, examining the impact of a country or region’s GINs embeddedness on manufacturing GVC from the perspective of “linkages”. It is common to base on patent information and use co-invented patents to explore issues related to innovation collaboration49. Since patents are critical indicators of innovation outcomes, this study uses the number of co-invention patents of 50 countries in PCT applications to reflect innovation collaboration relationships between economies, and constructs GINs to study the impact of GINs embeddedness on manufacturing GVC position.

In particular, GINs are constructed by taking economies as nodes and patent co-application relationships between different economies as edges. The embedding breadth is measured by“degree centrality”. It mainly measures the scope of global innovation collaboration carried out by an economy, and is expressed by the number of other nodes directly connected to it. It is divided into “out-degree centrality” (outd) and “in-degree centrality” (ind) with the following formula, where a is an element in the adjacency matrix, i and j represent different countries, and embedding breadth is the sum of “out-degree centrality” and “in-degree centrality”.

$${\text{out}}{{\text{d}}_i}=\mathop \sum \limits_{{\text{j}}} {{\text{a}}_{ij}}$$
(3)
$${\text{in}}{{\text{d}}_j}=\mathop \sum \limits_{{\text{i}}} {{\text{a}}_{i{\text{j}}}}$$
(4)

Network embedding depth is measured by “degree intensity”. It is mainly used to illustrate the total intensity of an economies’ innovation collaboration, and is expressed as the weighted total number of collaborations with other nodes connected directly to it. It is categorized into “out-degree intensity” (outs) and “in-degree intensity” (ins), as shown in the following formula, where w is an element in the weight matrix and embedding depth is the sum of the “out-degree intensity” and “in-degree intensity”.

$${\text{out}}{{\text{s}}_{\text{i}}}=\mathop \sum \limits_{{\text{j}}} {{\text{w}}_{{\text{ij}}}}$$
(5)
$${\text{in}}{{\text{s}}_{\text{j}}}=\mathop \sum \limits_{{\text{i}}} {{\text{w}}_{{\text{ij}}}}$$
(6)

Control variables

To account for the influence of other factors on manufacturing GVC and ensure the accuracy and reliability of empirical results, this study draws on existing scholarly research and selects the following control variables:

  1. (1)

    Level of financial development (Bank): It is measured by the proportion of domestic credit to the private sector in GDP. A well-developed financial system can reduce the financing constraints of enterprises, providing more financial support for them. It enables firms to invest in larger-scale R&D, make technological upgrades and improve production capacity, thereby facilitating their ascent to higher-value-added segments of the value chain.

  2. (2)

    Macroeconomic development environment (Mac): It is measured by GDP growth rate. The macroeconomic development environment reflects economic growth situation of a country or region. It plays an important role in shaping global capital flows and trade patterns, and thus has an impact on the participation patterns and positioning of countries or regions in GVC.

  3. (3)

    Level of technological innovation (Exp_tech): It is measured by the ratio of innovation expenditure to GDP. The level of technological innovation is a critical factor influencing a countries’ or region’s production efficiency and technological innovation capacity and directly determines its position in GVC.

  4. (4)

    Urbanization rate (Urban): It is measured by the ratio of the urban population to the total population in a country or region. The increase of urbanization rate can change the pattern of resource allocation and provide human resources support for the manufacturing and service industries, thus enabling more effective participation in GVC.

  5. (5)

    Relative scale of industrial added value (Ind_va): It is represented by the proportion of annual industrial added value in a country or region to its GDP. It reflects the contribution of the industrial sector to the national economy and is an important indicator for measuring the level of industrialization and economic development of a country or region.

  6. (6)

    Foreign direct investment (IFDI): It is measured by the ratio of inflow FDI to GDP. IFDI is an important channel for cross-border technology spillovers, helping domestic industries acquire advanced technologies, improve production efficiency, and further promote their GVC upgrading.

  7. (7)

    Trade openness (Trade): It is measured by the ratio of total exports and import trade to GDP. Trade openness directly reduces the import and export costs of intermediate and final goods, enabling firms to overcome geographical constraints and allocate resources on a global scale. Table 1 presents the definition of variables.

Table 1 The definition of variables.
Full size table

The GVC positions in this study are calculated based on the latest World Input-Output Table released by ABD and relevant data from GVC database (UIBE GVC Database) constructed by the University of International Business and Economics. The network characteristic indicators are primarily derived from PCT co-operation patent data in OECD database. Control variables are all sourced from the WDI database of World Bank. To ensure consistency and comparability across economies, this study focuses on national-level data, regional-level data are excluded. This approach enhances the reliability of comparative analyses and strengthens the validity of research conclusions. Specifically, we matched the data from the ABD database with the OECD database and excluded region-level observations to ensure the consistency and comparability of the results. As a result, we obtained a research sample consisting of 50 countries. These countries span a wide range of economic development levels and geographic regions, making them broadly representative of the global structure of innovation networks.

Based on the Hausman test results, this study employs a fixed effects model. To mitigate the impact of absolute numerical values, all data have undergone logarithmic transformation. Table 2 presents the descriptive statistical results of the main variables.

Table 2 Descriptive statistical results of main variables.
Full size table

Temporal evolution characteristics

Figure 1 shows the embedding breadth of 50 economies in 2007 and 2021. The more partners an economy has for innovation collaboration, the closer it is to the central position. The figure reveals that most economies engage in global innovation collaborations actively to pursue innovative development. In 2007, United States and other developed European economies ranked among the top of embedding breadth, which fully demonstrates that developed economies usually dominate R&D and innovation activities and often occupy a core position in GINs due to their strong innovation capabilities and technological levels. The embedding breadth in 2021 shows that the scope of global innovation collaboration continues to expand. Economies are increasingly participating in global innovation collaboration and actively seeking opportunities for collaborative innovation to maintain their competitive advantages. During this period, some developing economies’ embedding breadth has also continued to approach core positions, establishing innovation partnerships proactively with others, and pursuing technological exchanges and resource sharing on a global scale.

Fig. 1
figure 1

The GIN embedding Breadth of 50 Economies in 2007 and 2021.

Full size image
Fig. 2
figure 2

The GIN embedding Depth of 50 Economies in 2007 and 2021.

Full size image

Figure 2 shows the embedding depth of 50 economies from 2007 to 2021. The more frequently an economy engages in innovation collaboration, the more central its position becomes. In Fig. 2, United States occupied the core position in embedding depth with a significantly higher value than other economies both in 2007 and in 2021. This indicates that the U.S. maintains the closest connections and most active collaborations within GINs. By 2021, the U.S. retained its leading position with the highest number of collaborative engagements, far exceeding its 2007. Notably, China enhanced its embedding depth substantially by increasing cross-border innovation partnerships during this period, emerging as the third most frequent participant in global innovation collaborations, following only to the United States and Germany. The specific data on embedding breadth and embedding depth of different economies in 2007 and 2021 are shown in Table 3.

Table 3 Embedding breadth and embedding depth of 50 economies in 2007, 2014 and 2021.
Full size table

Empirical results and analysis

Benchmark regression analysis

The explanatory variables and control variables are included in the operation process of benchmark regression separately in order to effectively avoid possible multicollinearity issues during the regression process. Table 4 shows the benchmark regression results. The regression results of embedding breadth and embedding depth on manufacturing GVC position are presented in columns (1)–(2) and (3)–(4). The estimation coefficients of embedding breadth and depth are significantly positive at the 1% level, and the results are still significant after adding control variables, indicating that the enhancement of embedding breadth and depth can promote GVC position significantly during the sample observation period, which is consistent with hypothesis 1. Higher embedding breadth means that the innovation collaborative partners of an economy are widely distributed, providing more opportunities for them to absorb advanced overseas technologies and management experiences. Higher embedding depth signifies that the economy engages in innovation partnerships more frequently with other economies in the network. Through in-depth collaboration and exchanges, it is possible to boost the improvement of technological capabilities and upgrade the manufacturing GVC position.

Table 4 Benchmark regression results.
Full size table

Robustness checks

Replace the treatment of dependent variable processing method

This study conducts robustness tests around possible issues such as indicator selection and endogeneity in benchmark regression. Given that there are numerous indicators depicting GVC position, this study employs alternative indicator for dependent variable to conduct robustness checks, thereby mitigating potential biases stemming from indicator selection. In academic, the Export Sophistication Index (ESI) is usually used as an alternative indicator to measure an economies’ GVC position. Therefore, this paper calculates an economies’ TSI at industrial level from the perspective of value added to measure the manufacturing GVC position. The data is sourced from ADB database provided by the database of the University of International Business and Economics. As shown in Eqs. 7–9, \(\:{\text{X}}_{\text{i}}\:\text{a}\text{n}\text{d}\:{\text{Y}}_{\text{i}}\)represent the total exports and per capita income of economy respectively, while \(\:{\text{X}}_{\text{i}\text{k}}\)denotes the exports of Industry k in Economy i. The regression results are shown in column (1) and column (2) of Table 5. When using alternative indicators of the explained variables for regression, the results remain robust.

$${\text{E}}{{\text{S}}_i}=\mathop \sum \limits_{k} \frac{{{{\text{X}}_{ik}}}}{{{{\text{X}}_i}}}{\text{TS}}{{\text{I}}_k}$$
(7)
$${\text{TS}}{{\text{I}}_k}=\mathop \sum \limits_{i} \frac{{{\raise0.7ex\hbox{${{{\text{x}}_{ik}}}$} \!\mathord{\left/ {\vphantom {{{{\text{x}}_{ik}}} {{{\text{X}}_i}}}}\right.\kern-0pt}\!\lower0.7ex\hbox{${{{\text{X}}_i}}$}}}}{{\mathop \sum \nolimits_{i} \left( {{\raise0.7ex\hbox{${{{\text{x}}_{ik}}}$} \!\mathord{\left/ {\vphantom {{{{\text{x}}_{ik}}} {{{\text{X}}_i}}}}\right.\kern-0pt}\!\lower0.7ex\hbox{${{{\text{X}}_i}}$}}} \right)}}{{\text{Y}}_i}=\mathop \sum \limits_{i} {\text{RC}}{{\text{A}}_{ik}}{{\text{Y}}_i}$$
(8)
$${\text{TS}}{{\text{I}}_{ik}}=\frac{{{\raise0.7ex\hbox{${{\text{VA}}{{\text{X}}_{ik}}}$} \!\mathord{\left/ {\vphantom {{{\text{VA}}{{\text{X}}_{ik}}} {{\text{VA}}{{\text{X}}_i}}}}\right.\kern-0pt}\!\lower0.7ex\hbox{${{\text{VA}}{{\text{X}}_i}}$}}}}{{\mathop \sum \nolimits_{i} {\raise0.7ex\hbox{${{\text{VA}}{{\text{X}}_{ik}}}$} \!\mathord{\left/ {\vphantom {{{\text{VA}}{{\text{X}}_{ik}}} {\mathop \sum \nolimits_{i} {\text{VA}}{{\text{X}}_i}}}}\right.\kern-0pt}\!\lower0.7ex\hbox{${\mathop \sum \nolimits_{i} {\text{VA}}{{\text{X}}_i}}$}}}}{{\text{Y}}_i}={\text{RC}}{{\text{A}}_{ik}}{{\text{Y}}_i}$$
(9)

Censored treatment

To mitigate the influence of outliers on regression results, this paper trims the top and bottom 1% extreme values of the dependent and independent variables in the model and examines the changes in estimation outcomes. As shown in columns (3)-(4) of Table 5, the statistical significance of the GINs characteristic indicators remains robust.

Adjusting the sample range

In order to avoid the impact of the large-scale lockdowns due to the pandemic after 2020 on this study, this paper validates the results by using patent collaboration data of 50 economies from 2007 to 2019. The regression results in columns (5)-(6) of Table 5 show that the significance of the GINs characteristic indicators has not changed significantly.

Replace the treatment of independent variable processing method

This study constructs the GINs using patent cooperation data. However, this approach captures only formal innovation outcomes that result in patents. It might overlook a substantial amount of important informal information sharing and collaboration that does not lead to patent filings. Therefore, drawing on existing literature, this study conducts a robustness check by replacing the GIN embedding indicator with the number of cross-border patent citations (Citation) between economies. The data obtain from the Himmepat intelligent patent retrieval platform. The regression results in columns (7) of Table 5 show that the cross-border patent citations significantly promote manufacturing GVC upgrading.

Table 5 Robustness testing.
Full size table

Placebo test

Theoretically, the effect of embedded in GINs on manufacturing GVC upgrading may merely be a placebo effect caused by unrecognized limitations in the research design. To address the above concern and ensure the robustness of the conclusions, this study conducts a placebo test by randomly constructing pseudo-treatment groups. Specifically, this study performs 1,000 iterations of randomly matched placebo tests, constructs “new” core explanatory variables, and re-estimates the model accordingly. Figure 3 reports the distribution of the estimated coefficients of the core explanatory variable and their corresponding p-values based on 1,000 random draws. It is observed that the estimated coefficients are centered around zero and follow a normal distribution, with the vast majority of the results being statistically insignificant. Moreover, the actual coefficients of embedding breadth and embedding depth differ significantly from those obtained in the placebo tests, indicating that the manufacturing GVC upgrading is a result of embedded in GINs rather than random factors.

Fig. 3
figure 3

Random placebo test graph.

Full size image

Endogenous analysis

In general, reverse causality and omitted variables are the main sources of endogenous problems. This paper uses three approaches to examine potential endogenous issues in the model. Specifically, the core explanatory variable with one-period lagging, system GMM method and the instrumental variable (IV) approach50,51. This paper sets the mean values of the degree centrality (embedding breadth) and degree intensity (embedding depth) of other economies except itself as instrumental variables. The rationale for constructing the instrumental variable lies primarily in the fact that an economy’s relative position in GINs is often correlated with other economies’ positions within the same network, while the network centrality of other economies does not influence the economy’s global innovation collaboration directly. Specifically, GINs are highly interconnected and the embedding characteristics of other economies reflect common network dynamics that influence a given country’s position in GVCs, even though these dynamics are not directly driven by that country. As a result, network indicators from other countries possess exogeneity and can serve as valid instrumental variables. The regression results are presented in Table 6. Columns (1)-(2) in Table 6 present regression results of the core explanatory variable with one-period lagging, Columns (3)-(4) demonstrate regression outcomes addressing endogeneity by using GMM method, and Columns (5)-(6) exhibit estimations employing instrumental variables. All three approaches consistently validate that GINs embeddedness facilitates manufacturing GVC position.

Table 6 Endogeneity tests.
Full size table

This study conducts the following tests to ensure the validity of instrumental variables: Firstly, regarding instrument under-identification, the Kleibergen-Paap rk LM statistic test rejects the null hypothesis of under-identification, indicating no under-identification issues. Secondly, for weak instrument detection, the first-stage F-statistic confirms the instruments pass weak identification tests. These results demonstrate that both embedding breadth and embedding strength exert positive and robust effects significantly on GVC position.

Mechanism test

Embedded GINs are conducive to the human capital structure upgrading and technological progress. By collaborating with leading international research teams and participating in global innovation projects, firms and research institutions can attract high-end overseas talent, fostering the development of a more specialized and advanced human capital structure. Meanwhile, the accelerated flow and sharing of knowledge, technology and innovation resources enable firms to achieve technological breakthroughs and iterative improvements through the introduction, absorption and reinvention of advanced technologies. Based on the precious analysis52,53, it is necessary to further examine the significant roles of human capital structure upgrading and technological progress in facilitating manufacturing GVC upgrading through embedded GINs. Specifically, human capital structure upgrading (Labor force) is measured by the proportion of the labor force with higher education among the total labor population. Technological progress (TFP) is represented by the total factor productivity of each economy.

Table 7 Mechanism test.
Full size table

Columns (1) and (2) of Table 7 present the mechanism test results with human capital structure upgrading as the mediating variable. The results show that when human capital structure upgrading is used as the mediating variable, the coefficients of both the embedding breadth and embedding depth are significantly positive at the 1% level. Columns (3) and (4) report the results with technological progress as the mediating variable. The regression coefficients are also significantly positive at the 1% level, indicating that the GINs embedding breadth and depth promote technological progress of economies, thereby facilitating manufacturing GVC upgrading. These findings provide support for Hypothesis 2.

Further analysis

Heterogeneity analysis

Heterogeneity of economic development level

To examine the heterogeneous effects of GINs embeddedness on manufacturing GVC across economies at different economic development levels, this paper categorizes 50 economies into advanced and developing groups based on OECD classification criteria and conduct regression separately. Columns (1) and (2) of Table 8 present the regression results of advanced economies’ embedding breadth and depth. The results indicate that both embedding breadth and embedding depth promote manufacturing GVC position significantly in advanced economies. Columns (3)-(4) of Table 8 present the regression results for embedding breadth and embedding depth in developing economies. The findings reveal that embedding breadth in developing economies is insignificant statistically, while embedding depth exerts a significant inhibitory effect on manufacturing GVC position at the 1% level. This suggests that establishing broad innovation partnerships with other economies fails to enhance the manufacturing GVC position of developing economies. Moreover, deeper embedding depth may lead to a “low-end lock-in” phenomenon, trapping these economies in low-value-added segments of GVC.

Table 8 Heterogeneity of economic development level.
Full size table

Heterogeneity of technological distance from frontier

To examine the heterogeneous effects of GINs embeddedness on manufacturing GVC upgrading across economies with different technological levels, this study uses the total factor productivity (TFP) of the United States as the global technological frontier. The technological distance is measured by the gap between each economy’s TFP and that of the United States. Based on this measure, this study analyzes the heterogeneous impacts of GINs embeddedness on manufacturing GVC upgrading for economies at varying technological distances from the frontier. A larger technological distance to the frontier indicates that a country is closer to the global technological frontier. Using the median of the technological distance from frontier, the 50 economies are divided into “high” and “low” groups. The detailed regression results are presented in Table 9. Columns (1) and (2) of Table 9 present the regression results of embedding breadth and embedding depth for economies with a high technological distance from the frontier. The results indicate that both embedding breadth and embedding depth significantly promote the upgrading of manufacturing GVC positions. Columns (3) and (4) of Table 9 present the regression results of embedding breadth and embedding depth for economies with a low technological distance from the frontier. The study finds that GINs embedding breadth and embedding depth are statistically insignificant for economies with low technological distance from the frontier. This indicates that low technological levels hinder these economies from improving their manufacturing GVC position through innovation collaboration.

Table 9 Heterogeneity of technological distance from frontier.
Full size table

The moderating IPR protection

Intellectual property rights (IPR) protection serves as the cornerstone for fostering cross-border cooperation. By reducing the risks of technological collaboration, safeguarding returns on innovation and enhancing institutional attractiveness, IPR protection provides institutional support for the effective functioning of GINs. Robust IPR mechanisms incentivize multinational collaboration and technology transfer among enterprises and research institutions worldwide. However, excessive or inappropriate IPR protection may create technological barriers, restricting technology absorption and indigenous innovation capabilities of firms in developing countries, thereby hindering their upgrading and value capture in the GVCs. To investigate the heterogeneous effects of GINs embeddedness on manufacturing GVC position across economies at different development levels, this study introduces interaction terms between intellectual property rights (IPR) protection and GINs embeddedness to assess their impact on manufacturing GVC empirically. Data on IPR protection are sourced from the World Bank’s World Development Indicators (WDI) database. Columns (1) and (2) of Table 10 present the regression results for the interaction effects between GINs characteristic indicators and intellectual property rights (IPR) protection in developed economies. The regression results show that the interaction terms for both embedding breadth and embedding depth with IPR protection are significant positive at 1% level statistically. It indicates that higher intellectual property rights (IPR) protection ensure advanced economies receive necessary incentives and safeguards during innovation collaboration, allowing innovation outcomes to achieve adequate economic returns. Consequently, this mechanism further drives domestic manufacturing sectors to ascend to higher-end positions in GVC. Columns (3) and (4) of Table 10 report regression results for the interaction effects between GINs characteristic indicators and intellectual property rights (IPR) protection in developing economies. The regression results reveal that the interaction terms for both embedding breadth and embedding depth with IPR protection are significant negative at the 1% level statistically. It indicates that strengthened intellectual property rights (IPR) protection in developing economies significantly inhibits the role of GINs embeddedness in promoting manufacturing GVC position. A plausible explanation is that excessively stringent IPR protection may create technological barriers, limiting firms’ free access to critical technologies and their capacity for incremental innovation, thereby weakening their potential to upgrade within the global value chain through embedding.

Table 10 The moderating effect of IPR protection.
Full size table

Conclusion and policy implications

In the conclusion, we compared our findings with existing studies in the field of GINs. Consistent with earlier research, this study confirms that GINs embeddedness has a significant positive impact on the upgrading of manufacturing GVC position. This finding has a great significance for economies formulating strategies to embedded in global innovation collaborations and enhance national competitiveness. In addition, this study adopts cross-border patent collaboration data as a proxy for constructing GINs. While this is a commonly used indicator, it has certain limitations in fully capturing the multidimensional nature of GINs. Future research could build upon this study by incorporating broader indicators of innovation factor flows to enrich the understanding of GINs.

Specifically, This study examines the impact of GINs embeddness on manufacturing GVC by employing cross-border patent collaboration data from PCT applications across 50 economies from 2007 to 2021.The research results are as follows: Firstly, both embedding breadth and embedding promote manufacturing GVC position significantly. These findings remain robust after a series of sensitivity checks, including replacing core dependent variables, censored treatment, adjusting sample range and addressing potential endogeneity issues. Secondly, after categorizing economies into developed and developing economies based on income levels, it is found that both embedding breadth and embedding depth enhance GVC position in advanced economies significantly, while these effects are insignificant for developing economies. More precisely, a relatively high embedding depth may lead to the phenomenon of “low-end lock-in” in developing economies. Thirdly, to further explore the reasons behind these results, this study divided the embedding indicators into two groups based on the level of frontier technology distance. The analysis shows that in the group with a higher frontier technology distance, the GINs embeddedness significantly promotes GVC upgrading. In contrast, the group with a lower technology frontier distance, the effect of GIN embeddedness on GVC upgrading is not statistically significant. It suggests that absorptive capacity is a key factor contributing to the observed differences. Finally, by introducing the indicator of intellectual property right (IPR) protection and conducting interaction regressions with embedding breadth and embedding depth, it is found that stronger IPR protection amplifies the positive effect of GINs embeddedness on manufacturing GVC position in advanced economies significantly. Conversely, it suppresses the positive effect of GINs embeddedness on manufacturing GVC position in developing economies.

To enhance the core competence of developing economies in GINs comprehensively and promote their manufacturing GVC position, this paper proposes the following policy recommendations.

Firstly, given the positive impact of embedded in GINs on the upgrading of manufacturing GVCs, both developed and developing economies should continuously expand international scientific and technological exchanges and cooperation, and absorb advanced knowledge, technologies, and management experience through high-level international innovation collaborations.

Secondly, since the level of economic development and the frontier technology distance affect the impact of GINs embeddedness on manufacturing GVC upgrading, developing economies should especially focus on improving their economic development, strengthening independent innovation capabilities, and achieving a high degree of technological self-reliance.

Thirdly, because institutional frameworks have the impact of embedding indicators on manufacturing GVC upgrading in developing economies, these economies should choose appropriate institutional frameworks. In particular, regarding the level of IPR protection, developing economies should establish reasonable IPR protection standards to create a favorable innovation environment.

This paper investigates the impact of embedded GINs on the manufacturing GVC. The findings are robust and reliable, supported by robustness analyses and discussions addressing endogeneity. Furthermore, the study delves into the heterogeneous effects of GIN participation on GVC upgrading between developing and developed economies, highlighting that absorptive capacity and institutional frameworks are key components of global innovation cooperation strategies for developing countries. While the research has both theoretical and practical significance, there are undeniable limitations in the selection of GINs indicators, as the construction of GINs is not comprehensively captured from multiple perspectives. This also provides directions for future research.