Firms’ Position in the Supply Chain Network, R&D Input, and Innovation Output: Striving for the Top or Settling in the Corner? Implications for Sustainable Growth and Adaptive Capacity
Department of Financial Accounting, School of Business Administration, Quanzhou Campus, Huaqiao University, Quanzhou 362021, China
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(4), 1729; https://doi.org/10.3390/su17041729
Submission received: 18 January 2025
/
Revised: 14 February 2025
/
Accepted: 17 February 2025
/
Published: 19 February 2025
(This article belongs to the Special Issue Entrepreneurship, Innovation and Sustainable Business Development in a Changing Economic Environment)
Abstract
:This study examines the potential constraints that firms may face when occupying central positions within supply chain networks, particularly in terms of innovation. While prior research highlights the benefits of centrality for resource acquisition and knowledge flow, our findings suggest that such positioning can, under certain conditions, hinder innovation. Using unbalanced panel data from Chinese A-share listed firms in Shanghai and Shenzhen (2009–2021), we conduct an empirical investigation into this effect, incorporating the mediating role of R&D investment and the moderating influence of ownership structure. The analysis reveals that supply chain network centrality has a significantly negative impact on total innovation output, invention patents, and low-end patents, with all effects statistically significant at the 0.001 level. This adverse impact is particularly pronounced in state-owned enterprises, where dependence on established networks further restrains innovation. These results suggest that supply chain centrality may hinder firms’ long-term innovation capacity, which could, in turn, weaken their sustainability by limiting their ability to adapt to technological change and evolving industrial environments. These findings suggest that policymakers could implement targeted incentives, such as R&D subsidies, to mitigate the innovation constraints faced by central firms. Meanwhile, corporate managers should adopt strategies like open innovation and supply chain diversification to sustain long-term innovation.
1. Introduction
The 20th report of the Communist Party of China explicitly emphasizes the need to uphold innovation as the core engine of China’s comprehensive modernization drive, accelerate the implementation of an innovation-driven development strategy, improve the scientific and technological innovation system, and deepen reforms within the scientific and technological framework. This commitment is further reiterated in both the 14th Five-Year Plan and the Outline of Vision Goals for 2035, which highlight innovation as central to China’s overall modernization, with technological self-reliance serving as a strategic pillar for national development. As key players in this context, enterprises play a crucial role in contributing to national scientific and technological progress. Only by continuously enhancing the innovation capabilities of enterprises can we successfully achieve China’s ambitious goal of high-quality economic development.
Research on the determinants of firms’ innovation performance has attracted considerable interest from both scholars and practitioners. Existing studies have shown that firms’ innovation output is influenced by internal factors such as corporate governance [1], firm size [2,3], internal control [4], corporate social responsibility [5], and employee stock ownership plans [6], as well as external factors like government R&D subsidies [7], tax incentives [8], bank connections [9], and market competition [10]. It is important to note that a company is not an isolated entity; it operates in a complex business environment and interacts closely with various organizations and institutions. Through the social network established with these organizations, companies can share information and resources, making their social networks a crucial factor influencing their innovation performance. Scholars have studied the impact of director networks [11], international trade networks [12], and local networks [13] on firm innovation, while other studies have highlighted the impact of supply chain networks on corporate social responsibility [14,15], market competitive position [16], risk-taking behavior [17,18], and firm R&D investment [19]. Some of the literature has explored the influence of supply chain network position on firm innovation from the perspective of supply chain network embedding, such as the impact of supply chain network position on innovation performance [20,21,22] and innovation diversification [23].
While existing studies generally view the supply chain network position as beneficial, it is important to recognize that there may also be potential negative effects in theory. Additionally, the measures used in these studies may not be appropriate. For instance, closeness centrality is only effective for fully connected networks, where any node can reach all other nodes through a few intermediaries. However, a two- or three-tier supply chain networks built using database information may fail to satisfy this requirement. Moreover, existing research presents a contradiction when inferring the impact of a firm’s position within the supply chain network on innovation output. For example, closer proximity to the center of the supply chain network enhances innovation performance [21], while being near the center is associated with lower R&D investment [19]. It is generally believed that R&D investment does not always lead to innovation output, yet without R&D, innovation output is almost impossible to achieve. Being centrally positioned in the supply chain network reduces financing constraints, ensuring the stable progress of firms’ diversified innovation projects [23]. However, innovation output results from various factors, and R&D investment is just one of them. Firms may achieve more innovation output by acquiring other resources that contribute to innovation, even with reduced R&D input. Therefore, there are multiple theoretical possibilities.
This paper examines firms’ positions in the supply chain network, R&D investment, and innovation output, addressing three key issues: (1) how a firm’s network position influences innovation output, (2) the mediating role of R&D investment in this relationship, and (3) the moderating effect of ownership on this mediation.
Using a sample of non-financial A-share listed companies from Shanghai and Shenzhen between 2009 and 2021, the paper finds that a firm’s supply chain network centrality significantly and negatively affects its innovation output. This conclusion holds true even after robustness tests are conducted using instrumental variable methods, PSM, Heckman two-stage methods, and alternative variables. The mediating mechanism test reveals that a central position in the supply chain network leads to lower R&D investment, resulting in reduced innovation output. Heterogeneity analysis shows that the relationship between supply chain network position, R&D input, and innovation output is more pronounced in state-owned enterprises.
The potential contributions of this paper are as follows. First, it empirically demonstrates the negative impact of a firm’s supply chain network position on innovation output, enriching the literature on its economic consequences and addressing inconsistencies in existing research. Second, the analysis of the mediating mechanism suggests that firms in more central positions in the supply chain network may exhibit “inertia” in R&D investment due to their status advantages, leading to innovation delays. This provides theoretical guidance for firms in central positions to examine their behavior and recognize that “advantages may also become disadvantages”. Third, the heterogeneity analysis reveals that state-owned enterprises exhibit stronger “inertia” in R&D input and more pronounced negative consequences for innovation output, which offers a theoretical foundation for advancing state-owned enterprise reform.
2. Theoretical Analysis and Research Hypothesis
2.1. The Impact of Enterprises’ Supply Chain Network Position on Innovation Output
Innovation includes both product and process innovation. Product innovation refers to the development of new or enhanced tangible products and intangible services, while process innovation involves introducing new production methods for goods and services, whether technical or organizational [24]. The ultimate goal of innovation is to better meet market demand. The task involves accurately capturing and analyzing market information, which requires acquiring information and using resources. Enterprises can improve collaboration in knowledge sharing, transfer, and creation with other external enterprises through the supply chain network [23]. The number of enterprises an enterprise can contact is determined by its position in the supply chain network. An enterprise’s position in the supply chain network determines its dominance in terms of resource and information acquisition speed and the degree of control exercised by other enterprises decreases as the position becomes more central [20]. Resources and information are essential for innovation. Therefore, enterprises that are more central in the supply chain network are more likely to increase their innovation output.
However, while occupying a central position in the supply chain network provides firms with resource advantages, it may also introduce constraints that inhibit innovation.
First, firms face cognitive limitations in processing and integrating external knowledge. A central position in the supply chain exposes firms to a high volume of information from diverse stakeholders. However, rather than facilitating innovation, excessive knowledge inflows may lead to cognitive overload, reducing a firm’s ability to effectively filter, assimilate, and leverage new insights [25].
Second, from a corporate strategy perspective, firms in dominant supply chain positions often exhibit conservative innovation behaviors. Their market strength [16] and extensive bargaining power with supply chain partners reduce the immediate pressure to pursue radical innovations. Instead, they may prioritize incremental improvements over disruptive breakthroughs, as the latter involves greater uncertainty and potential risk.
Third, from an organizational behavior perspective, central firms tend to reinforce established routines and hierarchical decision-making processes, which can lead to structural rigidity. Over time, these firms may develop a preference for stability and predictability, limiting their willingness to explore unconventional innovation pathways.
Fourth, from a corporate culture perspective, firms embedded at the core of supply chain networks often prioritize operational consistency over exploratory initiatives. Their emphasis on maintaining seamless coordination within the supply chain can lead to an over-reliance on standardized practices, reducing their openness to novel and disruptive innovations.
Taken together, these factors suggest that while supply chain centrality offers firms competitive advantages, it may simultaneously impose structural, strategic, and cultural constraints that hinder their innovation potential. Accordingly, we put forward the following hypothesis.
H1.
Greater centrality within the supply chain network negatively affects an enterprise’s innovation output.
2.2. The Mediating Role of R&D Investment
From the perspective of the entire supply chain network, enterprises can achieve better performance, higher operational efficiency, and sustainable competitiveness [26,27]. The position of an enterprise relative to others in the supply chain network can affect its strategy and behavior [28]. Enterprises in a central position within the supply chain network can access more information and technical resources, improve their investment levels and efficiency, enhance their innovation abilities, and reduce transaction costs. This, in turn, promotes the improvement of their competitive position [16]. Enterprises in a central position within the supply chain network can obtain higher returns and have an advantageous position in negotiations with market capital providers. This can ease the constraints of internal and external sources of financing. Innovation input ability is negatively correlated with both internal and external financing constraints [29]. Thus, a higher degree of centrality in the supply chain network leads to a greater amount of funds available for research and development projects.
However, companies at the core of the supply chain network tend to face higher risks, such as greater volatility in stock returns [17], earnings [18], and operational risks [30]. Investing in R&D activities can be risky for enterprises, as it may worsen their cash flow instability and create a higher degree of conflicts of interest between management and shareholders. This can also negatively impact their position in the financing market and increase constraints on internal and external financing, ultimately leading to a decline in innovation output. Moreover, enterprises that leverage their central position in the supply chain network to gain advantages in negotiations with upstream and downstream firms may experience reduced pressure to innovate, ultimately leading to decreased innovation input and output. Accordingly, we present the following hypothesis.
H2.
Greater centrality in the supply chain network constrains R&D investment, ultimately hindering innovation output.
2.3. The Moderating Effect of Ownership
State-owned and non-state-owned enterprises differ significantly in resource control and access to policy support. State-owned enterprises are typically under government control, which provides them with resource advantages through resource allocation and other means. Secondly, state-owned enterprises typically have established relationships with larger upstream and downstream cooperative enterprises [31]. These stable cooperative relationships can better promote collaborative innovation among enterprises [32]. Third, state-owned enterprises can obtain large-scale and long-term credit funds from banks at a lower credit cost [33]. They also face lower financing constraints and have relatively abundant funds for R&D investment. Therefore, the centrality of the supply chain network position of state-owned enterprises may lead to more R&D investment, resulting in higher innovation output compared to non-state-owned enterprises.
However, enterprise ownership may negatively influence the link between supply chain network position and innovation output. First, state-owned enterprises often bear greater social responsibilities [34] and government tasks which results in the limited resources of state-owned enterprises being allocated to other tasks, such as infrastructure construction and employee welfare, thus limiting the amount of R&D investment. Secondly, state-owned enterprises may hold a monopoly position or receive government support in certain fields. As a result, their market share and profits may not rely on technological innovation, which can weaken their motivation to invest in R&D and reduce the efficiency of innovation output generated from such investments. Third, state-owned enterprises have a unique management system with administrative characteristics. They typically have a large management hierarchy and a more redundant organizational structure, resulting in a cumbersome decision-making process and low efficiency. This system may impose more restrictions and obstacles on the decision-making and promotion of high-risk R&D projects. Therefore, compared to non-state-owned enterprises, state-owned enterprises bear greater social responsibilities and government mandates, face less competitive pressure, and operate under an administrative management system, all of which collectively result in lower R&D investment and reduced innovation output. Based on this, we propose the following hypothesis.
H3.
Compared to non-state-owned enterprises, state-owned enterprises’ centrality in the supply chain network more strongly suppresses R&D input, further reducing innovation output.
Figure 1 shows the framework of this study.
3. Research Design
3.1. Sample Selection and Data Sources
This paper examines A-share listed companies on the Shanghai and Shenzhen stock exchanges from 2009 to 2021. The initial research sample was screened as follows: (1) financial and insurance companies were excluded following common research practices; (2) companies under special treatment, including ST, ST*, PT, and delisted firms, were excluded; (3) observations with missing key data were removed. The final sample comprised 11,207 observations. Data on the top five suppliers and customers of publicly listed firms, as well as patent application information, were obtained from the CNRDS database, while financial indicators, corporate governance characteristics, and other relevant data were sourced from the CSMAR database. To reduce outlier effects, all continuous variables were Winsorized at the 1% level on both ends.
3.2. Definition of Variables
3.2.1. Dependent Variables
Innovation output (Output). Two types of indicators measure enterprise innovation output: absolute and relative indices. The absolute index measures the total patent application count or grants count, while the relative index includes metrics like new product market share and the share of intangible assets in total assets at the end of the period. Compared to relative indices, absolute patent-based measures may better reflect a firm’s technological innovation capacity. According to the studies [31,35,36], the influence of supply chain network position on different innovation outputs is examined.
For total innovation output, the combined count of invention, utility model, and design patent applications is used (Output1). Invention patent applications represent high-end innovation output (Output2), while utility model and design patent applications combined represent low-end innovation output (Output3). To control for scale effects, the logarithm of patent applications (divided by 10,000 and incremented by 1) is applied.
We apply a logarithmic transformation to the innovation output variables to account for their skewed distribution, scale variations, and multiplicative effects. This transformation aids in normalizing the data, reducing heteroskedasticity, and aligning the model more closely with economic principles and statistical robustness.
3.2.2. Independent Variable
Degree centrality (Degree). This key measure evaluates a participant’s importance in a network based on direct connections. Established literature provides systematic measures of network centrality, including degree, betweenness, closeness, and eigenvector centrality.
Degree centrality, which counts the count of direct connections a firm has in the network, is more directly linked to the amount of information and resources available than betweenness (a bridge measure), eigenvector (which accounts for neighbors’ centrality), or closeness centrality (characteristic of fully connected networks). Thus, degree centrality is used to measure a firm’s position within the supply chain network [16,30,37].
Construction steps:
- Identify top five suppliers and customers for each listed company using the CNRDS database;
- Build a ‘listed company–main supplier (customer)’ relationship table by year in Excel;
- Convert the relationship table to a network format using creatpajek 1.0 software, referencing Luo Jiede’s lecture notes on social network analysis;
- Import the network into Pajek and compute degree centrality. A node’s degree is calculated as its direct connections divided by the total nodes minus one.
3.2.3. Mediating Variable
3.2.4. Moderating Variable
Ownership (State). Ownership (State) is indicated by a binary value: assigned 1 for state-owned enterprises and 0 for non-state-owned enterprises.
3.2.5. Control Variables
3.3. Main Effect Modeling
To test hypothesis H1, a firm-level panel data model was constructed. The model uses enterprise innovation output as the explained variable and enterprises’ supply chain network position (degree centrality) as the explanatory variable.
In the model (1), represents the enterprise, represents the year, is a group of control variables, and represent the industry and year fixed effects, respectively, and is the random error term.
4. Results and Analysis
4.1. Descriptive Statistics and Correlation Analysis
4.1.1. Descriptive Statistical Results
This chapter uses Stata 17.0 statistical software to conduct empirical analysis on the collected research data. Table 2 shows the descriptive statistics for each variable. Output1 has a mean value of 0.005 and a median value of 0.001. Since the mean is greater than the median, it suggests that most enterprises have lower innovation output, while a few have a significantly higher innovation output. The standard deviation is 0.012, indicating considerable variation in the total innovation output of listed enterprises.
The comparison between the mean and median for invention patent output (Output2) and low-end patent output (Output3) follows a similar pattern as Output1. Additionally, the mean value of invention patent output (Output2) is smaller than that of low-end patent output (Output3), suggesting that more innovative outputs tend to be relatively smaller in number.
Enterprises’ position in the supply chain network (Degree) range from 0.000 to 0.003, with a mean of 0.001, a median of 0.001, and a standard deviation of 0.001. These results are consistent with the findings [16,23], reflecting variations in the positions of different enterprises in the supply chain network. The descriptive statistics for the control variables are consistent with those of previous studies.
4.1.2. Correlation Analysis
In order to prevent multicollinearity issues between variables and to make a preliminary assessment of the proposed research hypothesis, Pearson and Spearman correlation analysis methods are used to examine the correlations within the total sample of enterprise innovation output. The correlation analysis results for all variables are reported in Table 3.
First, the position in the supply chain network shows a significant negative correlation with total innovation output at the 1% level. This provides preliminary evidence that an enterprise’s position in the supply chain network negatively impacts its innovation output. However, further regression analysis is needed to explore the underlying mechanism of this influence.
Second, the majority of correlation coefficients between supply chain network position and control variables are significant at the 1% level. This supports the validity of the selected control variables, indicating that they effectively account for the effects of other factors.
Finally, the pairwise correlation coefficients among the control variables are generally below 0.7, reducing the likelihood of multicollinearity significantly affecting the subsequent regression analysis.
Further observation of Table 4 reveals that the variance inflation factor (VIF) for each major variable remains low. This suggests no significant multicollinearity among the model’s variables, thereby improving the reliability of coefficient estimates and ensuring the model’s suitability for subsequent regression analysis.
4.2. Analysis of Regression Results
4.2.1. Main Effects Tests
Table 5 presents the regression results for the impact of enterprises’ supply chain network position on their innovation output. Columns (1) to (3) indicate a significant negative correlation between the Degree and the Output, meaning that less innovation output is observed when the enterprise holds a more central position in the supply chain network. The coefficients of Degree in columns (2) and (3) are −1.165 and −1.104, respectively, both significant at the 1% level. This indicates that position centrality has a more substantial negative impact on invention patent output. This result aligns with the following reasoning: compared to low-end innovation output, invention patents require more time and capital and have a lower probability of success. As a result, firms are more likely to opt for utility model and design patents, which not only efficiently meet their immediate development needs but also allow them to respond quickly to market changes.
4.2.2. Robustness Tests
1. Substitution of dependent variables
In addition to the commonly used number of patent applications, some researchers use the number of patent grants [41,42] as a proxy variable to measure firms’ innovation output. The differences between the two indicators are as follows. (1) The time it takes for a patent to be granted to an enterprise depends on the type of patent and the progress of the application, so the number of patents applied for more accurately reflects the timing of the enterprise’s innovation output. (2) Not all patent applications are successful, and only granted patents have been certified by the National Patent Office. Thus, the number of patents granted better reflects the actual value of the enterprise’s effective innovation output. The processing of the index remains consistent with the approach described above, and the results remain robust after re-regression, as shown in columns (4) to (6) of Table 5. The coefficients of Degree in columns (5) and (6) are both significantly negative at the 1% level. These values differ from those in columns (2) and (3), which confirms the difference between the two indicators.
2. Propensity score matching (PSM)
To mitigate the impact of differences in characteristics beyond an enterprise’s position in the supply chain network, the propensity score matching (PSM) method was applied. Enterprises were divided into four groups based on the Degree index, ranked from largest to smallest. The highest group was designated as the experimental group (assigned a value of 1), while the remaining three groups were combined as the control group (assigned a value of 0). The same control variables as in model (1) were used as covariates, and 1:1 matching with replacement was conducted based on the propensity score (Pscore). After matching, 2975 enterprises were included in the experimental group and 1419 in the matched control group.
The covariate balance test indicates that, after matching, mean differences in all control variables between the treatment and control groups dropped below 5% and became statistically insignificant. Moreover, the value of LR chi2 decreased significantly, confirming that the matching effect met expectations.
The matched subsamples were subsequently re-estimated using regression analysis, with the results presented in Table 6. The regression results show that the coefficient on Degree remains negative and significant at the 1% level. This finding suggests that even after controlling for differences in other characteristics, greater centrality in the supply chain network is still associated with lower innovation output.
3. Instrumental variable method (2SLS)
To examine whether endogeneity is present in the model, we conducted the Hausman test, which compares the consistency of OLS estimates with those of the two-stage least squares (2SLS) method. The test results show that the p-values of the Chi-squared statistic are all below 0.01, leading us to reject the null hypothesis. This indicates that the OLS estimates are inconsistent due to endogeneity, making 2SLS the more appropriate estimation method. To address potential endogeneity issues, we conducted regression analysis using the instrumental variable method. Regarding the selection of instrumental variables, it is important to note that the supply chain network position of a single enterprise is typically associated with the average position of other enterprises in the industry. However, its own innovation output is unlikely to be affected by the supply chain network position of other enterprises in the same industry. Thus, this paper selects the mean degree centrality (IV Degree) of other firms in the same industry and year as the first instrumental variable [16,43]. Additionally, the quality of transportation infrastructure in the region where the enterprise is located is a crucial factor for suppliers and customers when selecting potential partners. If the transportation infrastructure in the region is of higher quality, suppliers and customers are more likely to collaborate with the enterprise, thereby enhancing its position in the supply chain network. However, there is no direct theoretical connection between the quality of transportation infrastructure and the innovation output of enterprises. Therefore, we use the quality of transportation infrastructure (IV Traffic) in the region where listed enterprises are located as the second instrumental variable. The level of transportation infrastructure is measured by dividing the total highway mileage in the region by the population at the end of the year, with data derived from the China Statistical Yearbook of each year [44].
The results of the two-stage least squares (2SLS) regression are shown in Table 7 and Table 8. As can be seen from the first-stage estimation results in column (1) of Table 7 and Table 8, the coefficients of IV Degree and IV Traffic are both significantly positive at the 1% level, which is consistent with the expectations of this paper. The validity test results of the two instrumental variables show that the F-values of the first stage are 128.51 and 109.12, respectively, both significantly higher than the critical value of 10. The Kleibergen–Paap rk LM values are 319.050 and 23.740, respectively, and both are significant at the 1% level, rejecting the null hypothesis of “insufficient identification”. The Kleibergen–Paap rk Wald F values are 380.082 and 23.757, respectively, both higher than the critical value of 16.38 (Stock-Yogo) at the 10% significance level, strongly rejecting the null hypothesis of “weak instrument”. These results indicate that both the mean centrality of other enterprises in the same industry (IV Degree) and the quality of transportation infrastructure (IV Traffic) are effective instrumental variables for the position of firms in the network of supply chains. The second-stage estimation results in columns (2) to (4) of Table 7 and Table 8 show that the coefficient on Degree is negative at the 1% significance level. This study confirms that there is an inverse relationship between the centrality of an enterprise’s position in the supply chain network and its innovation output, even after correcting for endogeneity.
4. Heckman’s two-stage method
Some enterprises did not disclose their suppliers and customers in their annual reports, potentially leading to sample selection bias. To address this issue, the Heckman two-stage method was employed. In the first stage, data from all A-share listed enterprises during the sample period were used to estimate the likelihood of disclosing supplier and customer information. A dummy variable was created to indicate whether an enterprise disclosed this information: a value of 1 was assigned if the enterprise disclosed complete supplier and customer information, and 0 otherwise.
The disclosure practices of other firms in the same industry are assumed to influence a firm’s decision to disclose, while being unlikely to directly affect the firm’s characteristics. Therefore, the mean disclosure rate of other firms in the same industry and year (Ev_dis) was included as an exclusion restriction in the first-stage regression model. Additionally, the control variables from model (1), which are related to whether a firm discloses detailed information, were included as control variables in the first-stage regression.
Using the regression results from the first stage, the Inverse Mills Ratio (IMR) was calculated and included as a control variable in the second-stage regression model. The regression results are presented in Table 9. Columns (2) to (4) of Table 9 show that the coefficient on Degree remains negative and significant at the 1% level in the second stage. This confirms that the centrality of an enterprise’s position in the supply chain network negatively impacts its innovation output, even after controlling for potential sample selection bias.
5. Mechanism Test
5.1. Mediating Effect of R&D Investment
This paper selects the relative indicators (R&D expenditure/operating revenue) widely used by scholars [38,39,40] to measure R&D investment (RD). The following model is designed using the mediating factor test method [45], with control variables consistent with those in the previous section.
Model (2) is identical to Model (1). Model (3) introduces R&D as a mediating variable, and Model (4) adds this mediating variable to Model (2). The control variables are consistent with those in the previous section.
Table 10 shows the results of the test examining the relationship between supply chain network position, R&D input, and innovation output. In Table 10, column (1) uses enterprise R&D input (RD) as the dependent variable and supply chain network position (Degree) as the independent variable. The coefficient of Degree is −1.941, significant at the 1% level, indicating a negative correlation between supply chain network position and enterprise R&D input. The coefficients of Degree and RD in columns (2) through (4) are −2.265 and 0.025, −1.132 and 0.017, and −1.088 and 0.008, respectively, all significant at the 1% level.
Additionally, the coefficient of Degree increased from −2.315, −1.165, and −1.104 in columns (1) to (3) in Table 5 to −2.265, −1.132, and −1.088 in columns (2) to (4) in Table 10, respectively. The sign of ζ1η2 is negative, consistent with , indicating that R&D input acts as a partial mediating factor. This demonstrates that R&D input plays a significant role in mediating the relationship between enterprises’ supply chain network position and innovation output. The mediating effect accounts for 2.10%, 2.83%, and 1.41% (ζ1η2/α1) of the total effect in the respective models.
Furthermore, we conducted a re-test using the Bootstrap method (based on 500 resampled datasets) and found that the p-values for the Ind_eff test were 0.001, 0.000, and 0.002, respectively. These results further support the role of R&D input as a partial mediating factor, providing additional evidence to confirm Hypothesis 2.
5.2. Moderating Effect of Ownership
To investigate the impact of ownership heterogeneity (State), we use model [46] to test the moderated mediating effects. The model is constructed as follows:
Model (5) tests whether the direct effect of a firm’s position in the supply chain network on innovation output is influenced by moderating variables when mediating factors are not considered.
If b1 in Model (6) and c4 in Model (7) are significant, it indicates a moderated mediating effect through the second half-path. If b3 in Model (6) and c3 in Model (7) are significant, it indicates a moderated mediating effect through the first half-path. The control variables remain consistent across all models.
Table 11 and Table 12 present the moderating effect of ownership (State) on the mediating effect of R&D input. First, the regression results in Table 11 show that the coefficient of DegreeState in column (1) is −0.126, but it is not significant. This suggests that ownership (State) does not moderate the direct effect of enterprises’ supply chain network position on innovation output. Second, the coefficient of DegreeState in column (2) is −1.607, significant at the 10% level, and the coefficient of RD in column (3) is 0.014, significant at the 10% level, indicating that the moderated mediating effect is established and the first half-path is moderated. Specifically, in comparison to non-state-owned enterprises, state-owned enterprises’ position in the supply chain network has a greater decreasing effect on R&D investment. Finally, the coefficient of Degree in column (2) is −1.243, significant at the 10% level, and the coefficient of RD*State in column (3) is 0.057, significant at the 1% level, indicating that the moderated mediating effect is established and the second half-path is moderated. In other words, under the premise that the enterprises‘ position in the supply chain network negatively affects their R&D input, the reduction in R&D input has a stronger mitigating effect on state-owned enterprises’ innovation output compared to non-state-owned enterprises. This result is consistent with the regression results using Output2 and Output3 as indicators of innovation output in Table 12 (the regression results of Model (6) are consistent with column (2) in Table 11 and are therefore omitted).
6. Conclusions and Implication
6.1. Conclusions
This study examines the impact of supply chain network centrality on firms’ innovation output, with a further analysis of the mediating role of R&D investment and the moderating effect of ownership structure.
The findings indicate that firms with higher centrality in the supply chain network tend to have lower innovation output, challenging the mainstream view that central positioning within the supply chain fosters innovation [18,21,22]. Although central firms typically have greater access to resources and information flow [47,48], our analysis reveals that their lower R&D investment constrains innovation. The negative effect of network centrality on R&D investment aligns with previous findings [19]. This adverse impact is more pronounced in invention patents, which generally represent a higher level of innovation, suggesting that firms in central positions may be less inclined to engage in high-risk, high-barrier technological advancements. Further analysis highlights the critical moderating role of ownership structure. Compared to non-state-owned enterprises, state-owned enterprises (SOEs) experience a steeper decline in R&D investment when occupying central positions in the supply chain, leading to an even greater reduction in innovation output. This pattern may stem from the dominant position of SOEs in supply chains, making them more reliant on existing resources rather than proactively pursuing innovation. The stronger negative impact of supply chain centrality on innovation among SOEs not only reflects their strategic behavior in market competition but also underscores the broader influence of supply chain structure on corporate decision-making.
The findings of this study not only reveal the inhibiting effect of supply chain centrality on innovation but also highlight its potential drawbacks for firms’ long-term sustainable development. Firms in central positions within the supply chain may become more reliant on existing business models, reducing investment in technological upgrades and industry transformation. This, in turn, could weaken their long-term adaptability and competitive edge in the market. Particularly in the context of global economic transitions and the growing emphasis on sustainability, corporate innovation plays a crucial role in environmental, social, and economic sustainability. Therefore, the suppressive effect of supply chain centrality on innovation may not only hinder firms’ short-term technological competitiveness but also limit their potential for sustained growth over longer time horizons.
6.2. Theoretical Contributions
This study offers a new perspective on the relationship between supply chain centrality and corporate innovation, contributing in several key ways. First, it challenges the prevailing view that supply chain centrality enhances innovation by providing empirical evidence that central positioning may, in fact, suppress innovation. It also identifies the mediating role of R&D investment in this process. While prior research has largely focused on how supply chain centrality fosters innovation through resource sharing and knowledge flow, our findings suggest that occupying a central position in the supply chain network may reduce firms’ innovation incentives, enriching discussions on the strategic impact of network structures. Furthermore, by uncovering differences in innovation outcomes across ownership structures, this study extends research on the intersection of supply chain networks and corporate governance. The results indicate that the easier access to resources enjoyed by state-owned enterprises may reinforce innovation path dependence, ultimately shaping their strategic innovation choices.
From a managerial perspective, the findings of this study serve as a cautionary note for firms on how to navigate the innovation challenges associated with supply chain centrality. Huawei, for instance, has fostered an “internal entrepreneurship mechanism” to encourage autonomous innovation, successfully incubating projects like HiSilicon, which led to breakthroughs in chip design. Gree has leveraged a diversified supply chain strategy to overcome path dependence, drive technological advancements, mitigate risks associated with key component supplies, and maintain its leadership in core air-conditioning technologies. Toyota has adopted an “open innovation model”, collaborating with supply chain partners to develop new technologies, securing a global edge in hybrid and hydrogen fuel cell innovations. Meanwhile, Tesla has enhanced innovation capabilities through supply chain collaboration, utilizing external technological resources and working closely with suppliers such as Panasonic and CATL to accelerate battery technology advancements and cost reductions. These cases illustrate that firms in central positions in the supply chain can mitigate the adverse effects of path dependence on innovation by refining innovation management models, strengthening external partnerships, and optimizing organizational incentives, ultimately enhancing long-term competitiveness and sustainability.
For state-owned enterprises, given their strong position in the supply chain and greater resource accessibility, fostering innovation can be achieved through strategies such as establishing independent R&D subsidiaries, enhancing the integration of global innovation networks, and optimizing internal incentive mechanisms. For example, China Aerospace Science and Technology Corporation has strengthened its internal innovation capacity by setting up independent R&D centers and introducing market-driven competition mechanisms, leading to significant technological breakthroughs in areas such as manned spaceflight, lunar exploration, and the Beidou navigation system.
6.3. Policy Implications
To mitigate the constraints of supply chain centrality on innovation, policymakers should implement more targeted reforms and incentive measures.
First, the government should refine the performance evaluation system for state-owned enterprises, reducing the pressure of short-term performance targets that may discourage innovation investment. Instead, firms should be encouraged to integrate technological breakthroughs and long-term competitiveness into their core assessment criteria. For example, Germany’s Industrie 4.0 strategy has promoted digital and intelligent transformation in manufacturing through policy frameworks and technological support. While it does not mandate enterprises to disclose innovation performance or incorporate it into a national evaluation system, its policy-driven approach to fostering innovation offers valuable insights. China could adopt a similar strategy by developing an innovation-centered evaluation framework that incentivizes enterprises to focus on long-term technological advancement.
Second, while the government has already implemented various financial measures to support corporate innovation, more targeted support for firms in central positions in the supply chain could encourage greater investment in frontier technologies, green innovation, and digital transformation. For example, South Korea’s Semiconductor Industry Support Program has helped companies like Samsung maintain a competitive edge in the global semiconductor market through tax incentives, R&D subsidies, and infrastructure support. China could adopt a similar approach by enhancing support mechanisms, such as offering R&D tax benefits and dedicated technology funds, to reduce firms’ reliance on existing industry models and strengthen their innovation drive.
Third, while China has made progress in industry–academia–research collaboration, there is still room to deepen cooperation and improve technology commercialization. Strengthening policy guidance to encourage firms in central positions in the supply chain to establish joint laboratories with research institutions and universities for long-term collaboration could be beneficial. For example, the China Railway Rolling Stock Corporation’s close partnerships with universities have led to breakthroughs in high-speed rail technology, demonstrating that fostering knowledge sharing and collaborative R&D can effectively enhance the innovation capacity of firms in central positions in the supply chain while strengthening their competitiveness in sustainability and industrial upgrading.
6.4. Limitations and Future Research Directions
This study has certain limitations, and future research could further expand in the following directions. (1) Investigations could be conducted on the specific mechanisms through which supply chain centrality affects innovation, such as information processing capacity, bargaining power, and strategic inertia, using more refined empirical analysis methods. (2) Cross-country comparisons could be conducted to examine how institutional environments influence the relationship between supply chain centrality and innovation, enhancing the study’s global relevance. (3) An analysis of industry differences could be conducted to explore how supply chain centrality impacts innovation across various sectors. (4) Dynamic data could be incorporated to examine how the effects of supply chain centrality on corporate innovation evolve over time, providing insights into its long-term impact.
Author Contributions
Conceptualization, L.C. and L.W.; methodology, L.C.; formal analysis, L.C. and L.W.; writing—original draft preparation, L.C.; writing—review and editing, L.W.; funding acquisition, L.W. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by Fujian Province Innovation Strategy Research Plan Project, grant number 2023R0044.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data presented in this study are available in Baidu Netdisk at https://pan.baidu.com/s/1_dkb5_RKdOf-ohyyXPFAYg, accessed on 17 January 2025. Access Code: 5tpp. These data were derived from the following resources available in the public domain: [CSMAR, CNRDs and China Statistical Yearbook of each year].
Conflicts of Interest
The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
Abbreviations
The following abbreviations are used in this manuscript:
| MDPI | Multidisciplinary Digital Publishing Institute |
| DOAJ | Directory of open access journals |
| TLA | Three-letter acronym |
| LD | Linear dichroism |
References
- Belloc, F. Corporate governance and innovation: A survey. J. Econ. Surv. 2012, 26, 835–864. [Google Scholar] [CrossRef]
- Shefer, D.; Frenkel, A. R&D, firm size and innovation: An empirical analysis. Technovation 2005, 25, 25–32. [Google Scholar] [CrossRef]
- Rogers, M. Networks, Firm size and innovation. Small Bus. Econ. 2004, 22, 141–153. [Google Scholar] [CrossRef]
- Chan, K.C.; Chen, Y.; Liu, B. The linear and non-linear effects of internal control and its five components on corporate innovation: Evidence from Chinese firms using the COSO framework. Eur. Account. Rev. 2021, 30, 733–765. [Google Scholar] [CrossRef]
- Ma, Y.C.; Luo, Z.X. Corporate social responsibility, financing constraints and innovation performance of manufacturing industry: The "inverted u-shaped" regulating effect of R&D innovation investment. J. Shandong Univ. Sci. Technol. Soc. Sci. 2023, 25, 101–112. [Google Scholar] [CrossRef]
- Wang, W.H.; He, H.J.; Zhong, H.L. ESOP and enterprise innovation performance—Based on the mediating effect of innovation culture. Rev. Econ. Manag. 2023, 39, 97–108. [Google Scholar] [CrossRef]
- Xiong, K.J. The impact of government R&D subsidies and non-R&D subsidies on enterprises’ innovation output. J. Northeast. Univ. Soc. Sci. 2023, 25, 26–35. [Google Scholar] [CrossRef]
- Xiao, P.; Dai, L.T. Financial subsidies, tax incentives and innovation incentives: Effect analysis and policy implications. Rev. Econ. Manag. 2023, 39, 21–35. [Google Scholar] [CrossRef]
- Zhai, S.B.; Xu, H.R.; Tang, W.; Gao, K.; Cao, L. Bank connection and corporate innovation: Evidence from Chinese listed firms in manufacturing industries. Account. Res. 2018, 7, 50–56. [Google Scholar]
- Aghion, P.; Bloom, N.; Blundell, R.; Griffith, R.; Howitt, P. Competition and innovation: An inverted-u relationship. Q. J. Econ. 2005, 120, 701–728. [Google Scholar] [CrossRef]
- Zhang, D.; Hao, R. Can interlocking directorate network promote technological innovation performance? based on the mediating effect of R&D investment. Sci. Technol. Manag. Res. 2018, 38, 183–191. [Google Scholar]
- Shang, H.; Chen, Y.; Lu, A. International trade networks and corporate innovation. J. Int. Trade 2022, 11, 18–34. [Google Scholar] [CrossRef]
- Wang, L.; Chi, Q.Z. Relationship between localized network tie, globalized network tie and innovation performance of firms in cluster: Mediating effect of absorptive capacity. J. Technol. Econ. 2014, 33, 1–9. [Google Scholar]
- Park-Poaps, H.; Rees, K. Stakeholder forces of socially responsible supply chain management orientation. J. Bus. Ethics 2010, 92, 305–322. [Google Scholar] [CrossRef]
- Zhang, J.P.; Li, Z.F. Heavy is the head that wears the crown: Responsibility and fulfillment of supply-chain network center enterprises—From the perspective of ESG. Foreign Econ. Manag. 2024, 46, 86–101. [Google Scholar] [CrossRef]
- Yu, M.Y.; Lv, K.F.; Ruan, Y.P. Supply chain network position and enterprise competitive position. Syst. Eng. Theory Pract. 2022, 42, 1796–1810. [Google Scholar]
- Mihov, A.; Naranjo, A. Customer-base concentration and the transmission of idiosyncratic volatility along the vertical chain. J. Empir. Financ. 2017, 40, 73–100. [Google Scholar] [CrossRef]
- Lv, K.F.; Yu, M.Y.; Ruan, Y.P. If you listen to both sides, you will be wise—Supply chain network centrality and enterprise risk. Manag. Rev. 2023, 35, 266–280. [Google Scholar] [CrossRef]
- Zhang, J.P. The impact of closeness centrality in supply chain network on the R&D investment of enterprises. J. Xi’an Aeronaut. Inst. 2022, 40, 18–28. [Google Scholar]
- Kim, M.K.; Narayanan, S.; Narasimhan, R. Supply network architecture and its contingent impact on innovation Performance: A field study. Int. J. Prod. Econ. 2020, 224, 107551. [Google Scholar] [CrossRef]
- Lu, D.Y. The innovation effect and the mechanism analysis of supply chain network structure: Theory and empirical evidence. J. Yunnan Univ. Financ. Econ. 2023, 39, 40–58. [Google Scholar] [CrossRef]
- Zhang, J.P. Firms’ closeness centrality in supply chain network and innovation performance: Evidence from Chinese listed firms. J. Hubei Univ. Arts Sci. 2022, 43, 34–42. [Google Scholar]
- Zhao, X.Y.; Yi, C.J.; Liao, J. Can the supply chain network position Improve the level of innovation diversity of enterprises? Chin. J. Manag. Sci. 2023, 31, 340–349. [Google Scholar] [CrossRef]
- Edquist, C.; Chaminade, C. Industrial policy from a systems-of-innovation perspective. EIB Pap. 2006, 11, 108–132. [Google Scholar]
- Gilsing, V.; Nooteboom, B.; Vanhaverbeke, W.; Duysters, G.; Oord, A.V.D. Network embeddedness and the exploration of novel technologies: Technological distance, betweenness centrality and density. Res. Policy 2008, 37, 1717–1731. [Google Scholar] [CrossRef]
- Corbett, C.J.; Blackburn, J.D.; Wassenhove, L.N.V. Case study: Partnerships to improve supply Chains. MIT Sloan Manag. Rev. 1999, 40, 71–82. [Google Scholar]
- Dyer, J.H.; Nobeoka, K. Creating and managing a high-performance knowledge-sharing network: The Toyota case. Strateg. Manag. J. 2000, 21, 345–367. [Google Scholar] [CrossRef]
- Borgatti, S.P.; Li, X. On social network analysis in a supply chain context. J. Supply Chain Manag. 2009, 45, 5–22. [Google Scholar] [CrossRef]
- Chen, J.; Qu, L.Y.; Song, M.Y.; Ning, J.H. Research on the impact of internal and external financing constraints on innovation investment capability of the enterprises under different property rights. J. Tianjin Chengjian Univ. 2023, 29, 443–448. [Google Scholar] [CrossRef]
- Shi, J.Y.; Yang, J.H.; Li, Y.X.; Zhang, Q.W. A slight move in one part may affect the situation as a whole: Supply network position, operational risk and corporate performance. China Ind. Econ. 2019, 9, 136–154. [Google Scholar] [CrossRef]
- Mo, Y.L.; Wu, G.R. How can supply chain finance promote the innovation performance of SMEs?—Based on the perspective of corporate financialization and supply chain integration. Wuhan Financ. 2023, 7, 33–42. [Google Scholar]
- Li, Z.; Chen, J. Supplier innovativeness, supplier relationships, and enterprises’ finance constraints. Manag. Decis. Econ. 2024, 45, 652–663. [Google Scholar] [CrossRef]
- Yang, Z.; Liu, F.; Li, M.M. Interest rate marketization, inefficient investment and capital allocation: Based on the natural experiment of cancellation of upper and lower limits for loan interest rates set by PBoC. J. Financ. Res. 2017, 5, 81–96. [Google Scholar]
- Zhang, Y.; Lian, R.J. Analysis on the influence of corporate social responsibility group effect on innovation. Stat. Decis. 2023, 39, 174–178. [Google Scholar] [CrossRef]
- Kong, D.M.; Xu, M.L.; Kong, G.W. Pay gap and firm innovation in China. Econ. Res. J. 2017, 52, 144–157. [Google Scholar]
- Tang, S.; Wu, X.C.; Zhu, J. Digital finance and enterprise technology innovation: Structural feature, mechanism identification and effect difference under financial supervision. J. Manag. World 2020, 36, 52–66+9. [Google Scholar] [CrossRef]
- El-Khatib, R.; Fogel, K.; Jandik, T. CEO network centrality and merger performance. J. Financ. Econ. 2015, 116, 349–382. [Google Scholar] [CrossRef]
- Makri, M.; Lane, P.J.; Gomez-Mejia, L.R. CEO incentives, innovation, and performance in technology-intensive firms: A reconciliation of outcome and behavior-based incentive schemes. Strateg. Manag. J. 2006, 27, 1057–1080. [Google Scholar] [CrossRef]
- Gentry, R.J.; Shen, W. The impacts of performance relative to analyst forecasts and analyst coverage on firm R&D intensity. Strateg. Manag. J. 2013, 34, 121–130. [Google Scholar] [CrossRef]
- Wu, F.; Hu, H.Z.; Lin, H.Y.; Ren, X.Y. Enterprise digital transformation and capital market performance: Empirical evidence from stock liquidity. J. Manag. World 2021, 37, 130–144+10. [Google Scholar] [CrossRef]
- Yi, J.T.; Zhang, X.P.; Wang, H.C. Firm heterogeneity, top executives’ overconfidence, and corporate innovation performance. Nankai Bus. Rev. 2015, 18, 101–112. [Google Scholar]
- Meng, Q.B.; Li, X.Y.; Zhang, P. Can employee stock ownership plans promote innovation?—Empirical evidence based on the perspective of employees. J. Manag. World 2019, 35, 209–228. [Google Scholar] [CrossRef]
- Wang, H.C.; Cao, F.; Ye, K.T. Monitoring or tunneling? The proportion of the proportion held by the big shareholders and the risk of the crash of the stock price. J. Manag. World 2015, 2, 45–57+187. [Google Scholar] [CrossRef]
- Fang, J.; Yao, S.J.; Feng, G.F.; An, L. Can transportation infrastructure promote consumption market development?—Evidence from China’s high-speed rail projects. Nankai Econ. Stud. 2023, 9, 55–74. [Google Scholar] [CrossRef]
- Wen, Z.L.; Ye, B.J. Analyses of mediating effects: The development of methods and models. Adv. Psychol. Sci. 2014, 22, 731–745. [Google Scholar] [CrossRef]
- Wen, Z.L.; Ye, B.J. Different methods for testing moderated mediation models: Competitors or backups. Acta Psychol. sin. 2014, 46, 714–726. [Google Scholar] [CrossRef]
- Granovetter, M. Economic Action and Social Structure: The Problem of Embeddedness. Am. J. Sociol. 1985, 91, 481–510. [Google Scholar] [CrossRef]
- Salman, N.; Anne-Laure, S. Indirect Networks: An Intangible Resource for Biotechnology Innovation. RD Manag. 2005, 35, 203–215. [Google Scholar] [CrossRef]
Figure 1.
Enterprises’ supply chain network position and innovation output.
Table 1.
Variable definitions.
| Type | Name | Symbol | Definition |
|---|---|---|---|
| Dependent variables | Total innovation output | Output1 | Ln [(Invention patent + utility model patent + design patent)/10,000 + 1] |
| Invention patent output | Output2 | Ln (Invention patent/10,000 + 1) | |
| Low-end patent output | Output3 | Ln [(utility model patent + design patent)/10,000 + 1] | |
| Independent variables | Degree centrality | Degree | Number of network nodes with direct contact to the company/(total number of nodes in the entire supply chain network − 1) |
| Intermediary variables | R&D investment | RD | R&D expenditure/operating income |
| Moderating variables | Ownership | State | 1 for state-owned enterprises and 0 for non-state-owned enterprises |
| Control variables | Firm size | Size | Ln (total assets) |
| Firm growth ability | Growth | (Current year operating income—Prior year operating income)/Prior year operating income | |
| Firm age | Age | Ln (the total number of years from the company’s IPO to the end of the sample year) | |
| Return on assets | ROA | Net profit/total assets of the company | |
| Asset-liability ratio | Lev | Total liability/total assets | |
| Ownership concentration | TOP | Number of shares held by the largest shareholder/total number of shares | |
| Independent director ratio | Indep | Number of independent directors/Total number of board members | |
| Board size | Board | Ln (the number of board member) | |
| Whether the Board chairman and general manager are the same | Dual | If the President and the CEO are the same, take 1, otherwise take 0. | |
| TobinQ | TobinQ | Market value of the enterprise/total assets at the end of the period |
Table 2.
Descriptive statistics of the variables.
| Name | Obs | Mean | Median | SD | Min | P25 | P75 | Max |
|---|---|---|---|---|---|---|---|---|
| Output1 | 11,207 | 0.005 | 0.001 | 0.012 | 0.000 | 0.000 | 0.003 | 0.080 |
| Output2 | 11,207 | 0.002 | 0.000 | 0.006 | 0.000 | 0.000 | 0.001 | 0.039 |
| Output3 | 11,207 | 0.003 | 0.000 | 0.007 | 0.000 | 0.000 | 0.002 | 0.044 |
| Degree | 11,207 | 0.001 | 0.001 | 0.001 | 0.000 | 0.001 | 0.002 | 0.003 |
| RD | 11,207 | 0.029 | 0.019 | 0.040 | 0.000 | 0.000 | 0.040 | 0.247 |
| Age | 11,207 | 2.046 | 2.303 | 0.959 | 0.000 | 1.386 | 2.833 | 3.332 |
| Growth | 11,207 | 0.179 | 0.119 | 0.384 | −0.546 | −0.015 | 0.285 | 2.210 |
| ROA | 11,207 | 0.042 | 0.039 | 0.059 | −0.225 | 0.014 | 0.071 | 0.224 |
| Lev | 11,207 | 0.442 | 0.442 | 0.216 | 0.051 | 0.266 | 0.613 | 0.886 |
| Top | 11,207 | 0.357 | 0.333 | 0.152 | 0.088 | 0.235 | 0.464 | 0.746 |
| Size | 11,207 | 22.170 | 21.940 | 1.382 | 19.820 | 21.120 | 22.960 | 26.110 |
| Indep | 11,207 | 0.371 | 0.333 | 0.052 | 0.300 | 0.333 | 0.400 | 0.571 |
| TobinQ | 11,207 | 2.459 | 1.897 | 1.755 | 0.852 | 1.324 | 2.955 | 11.070 |
| Board | 11,207 | 2.164 | 2.197 | 0.200 | 1.609 | 2.079 | 2.197 | 2.708 |
| Dual | 11,207 | 0.232 | 0.000 | 0.422 | 0.000 | 0.000 | 0.000 | 1.000 |
| State | 11,207 | 0.447 | 0.000 | 0.497 | 0.000 | 0.000 | 1.000 | 1.000 |
Table 3.
Correlation analysis of the total sample of innovation output.
| Output1 | Degree | RD | Age | Growth | ROA | Lev | Top | Size | Indep | TobinQ | Board | Dual | State | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Output1 | 1 | −0.051 *** | 0.457 *** | −0.036 *** | 0.047 *** | 0.061 *** | 0.054 *** | 0.026 *** | 0.274 *** | 0.037 *** | −0.115 *** | 0.016 * | 0.041 *** | −0.032 *** |
| Degree | −0.130 *** | 1 | 0.012 | 0.076 *** | −0.078 *** | −0.094 *** | −0.047 *** | −0.074 *** | −0.070 *** | −0.034 *** | 0.037 *** | −0.063 *** | −0.001 | −0.040 *** |
| RD | 0.082 *** | 0.031 *** | 1 | −0.400 *** | 0.042 *** | 0.176 *** | −0.379 *** | −0.134 *** | −0.264 *** | 0.043 *** | 0.289 *** | −0.140 *** | 0.211 *** | −0.347 *** |
| Age | 0.110 *** | 0.076 *** | −0.320 *** | 1 | −0.163 *** | −0.331 *** | 0.423 *** | −0.059 *** | 0.441 *** | −0.004 | −0.382 *** | 0.107 *** | −0.240 *** | 0.431 *** |
| Growth | 0.005 | −0.054 *** | −0.017 * | −0.070 *** | 1 | 0.343 *** | −0.023 ** | 0.018 * | 0.018 * | 0.003 | 0.132 *** | −0.016 * | 0.057 *** | −0.093 *** |
| ROA | 0.037 *** | −0.117 *** | 0.068 *** | −0.298 *** | 0.252 *** | 1 | −0.458 *** | 0.082 *** | −0.132 *** | −0.026 *** | 0.388 *** | −0.016 * | 0.106 *** | −0.199 *** |
| Lev | 0.155 *** | −0.039 *** | −0.364 *** | 0.458 *** | 0.011 | −0.412 *** | 1 | 0.081 *** | 0.557 *** | −0.006 | −0.492 *** | 0.177 *** | −0.186 *** | 0.364 *** |
| Top | 0.120 *** | −0.080 *** | −0.163 *** | −0.055 *** | 0.018 * | 0.100 *** | 0.086 *** | 1 | 0.209 *** | 0.023 ** | −0.130 *** | 0.014 | −0.065 *** | 0.241 *** |
| Size | 0.432 *** | −0.066 *** | −0.244 *** | 0.412 *** | 0.027 *** | −0.053 *** | 0.539 *** | 0.262 *** | 1 | −0.000 | −0.648 *** | 0.268 *** | −0.198 *** | 0.396 *** |
| Indep | 0.082 *** | −0.036 *** | 0.053 *** | −0.013 | 0.015 | −0.018 * | 0.001 | 0.051 *** | 0.034 *** | 1 | 0.015 | −0.482 *** | 0.094 *** | −0.072 *** |
| TobinQ | −0.116 *** | 0.039 *** | 0.289 *** | −0.273 *** | 0.083 *** | 0.257 *** | −0.380 *** | −0.114 *** | −0.485 *** | 0.043 *** | 1 | −0.188 *** | 0.178 *** | −0.352 *** |
| Board | 0.062 *** | −0.059 *** | −0.125 *** | 0.116 *** | −0.026 *** | −0.000 | 0.178 *** | 0.029 *** | 0.277 *** | −0.476 *** | −0.175 *** | 1 | −0.176 *** | 0.269 *** |
| Dual | −0.013 | 0.000 | 0.186 *** | −0.252 *** | 0.028 *** | 0.081 *** | −0.185 *** | −0.075 *** | −0.184 *** | 0.105 *** | 0.137 *** | −0.164 *** | 1 | −0.297 *** |
| State | 0.103 *** | −0.041 *** | −0.286 *** | 0.429 *** | −0.074 *** | −0.150 *** | 0.362 *** | 0.243 *** | 0.393 *** | −0.070 *** | −0.272 *** | 0.271 *** | −0.297 *** | 1 |
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. The Pearson correlation coefficient appears in the lower left corner, while the Spearman correlation coefficient is in the upper right.
Table 4.
Variance inflation factor analysis.
| Variable | VIF | 1/VIF |
|---|---|---|
| Degree | 1.370 | 0.729 |
| RD | 1.790 | 0.559 |
| Age | 1.860 | 0.538 |
| Growth | 1.150 | 0.870 |
| ROA | 1.600 | 0.626 |
| Lev | 2.260 | 0.442 |
| Top | 1.260 | 0.796 |
| Size | 2.630 | 0.380 |
| Indep | 1.380 | 0.727 |
| TobinQ | 1.830 | 0.547 |
| Board | 1.590 | 0.631 |
| Dual | 1.150 | 0.870 |
| State | 1.640 | 0.608 |
| Mean VIF | 2.930 |
Table 5.
Supply chain network position and innovation output.
| Dependent Variables | Substitution of Dependent Variables | |||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | (6) | |
| Output1 | Output2 | Output3 | Output1 | Output2 | Output3 | |
| Degree | −2.315 *** | −1.165 *** | −1.104 *** | −1.832 *** | −0.488 *** | −1.157 *** |
| (−11.662) | (−12.543) | (−9.759) | (−11.032) | (−12.431) | (−9.430) | |
| Age | 0.000 *** | 0.000 *** | 0.000 *** | 0.000 *** | 0.000 ** | 0.000 *** |
| (3.279) | (2.970) | (2.755) | (2.974) | (1.986) | (2.849) | |
| Growth | −0.001 ** | −0.000 * | −0.000 *** | −0.001 *** | −0.000 *** | −0.001 *** |
| (−2.529) | (−1.693) | (−3.062) | (−4.049) | (−3.590) | (−4.252) | |
| ROA | 0.002 | −0.001 | 0.003 *** | 0.000 | −0.002 *** | 0.002 * |
| (1.109) | (−1.284) | (2.804) | (0.062) | (−4.661) | (1.725) | |
| Lev | −0.003 *** | −0.002 *** | −0.001 *** | −0.003 *** | −0.001 *** | −0.002 *** |
| (−6.498) | (−7.872) | (−4.611) | (−6.107) | (−8.395) | (−4.641) | |
| Top | 0.002 ** | 0.000 | 0.001 *** | 0.002 *** | 0.000 * | 0.002 *** |
| (2.470) | (1.300) | (2.916) | (2.768) | (1.695) | (3.011) | |
| Size | 0.005 *** | 0.002 *** | 0.003 *** | 0.004 *** | 0.001 *** | 0.003 *** |
| (25.739) | (24.179) | (25.132) | (22.878) | (21.267) | (23.076) | |
| Indep | 0.008 *** | 0.005 *** | 0.004 ** | 0.008 *** | 0.003 *** | 0.004 ** |
| (2.968) | (3.582) | (2.282) | (3.343) | (4.979) | (2.367) | |
| TobinQ | 0.001 *** | 0.000 *** | 0.000 *** | 0.001 *** | 0.000 *** | 0.000 *** |
| (10.190) | (11.661) | (8.416) | (9.446) | (10.776) | (8.261) | |
| Board | −0.000 | 0.000 | −0.000 | 0.000 | 0.000 | −0.000 |
| (−0.107) | (0.478) | (−1.069) | (0.150) | (1.262) | (−0.777) | |
| Dual | 0.001 *** | 0.000 * | 0.000 *** | 0.001 *** | 0.000 * | 0.001 *** |
| (2.950) | (1.688) | (3.692) | (3.518) | (1.745) | (3.955) | |
| _cons | −0.104 *** | −0.047 *** | −0.055 *** | −0.084 *** | −0.019 *** | −0.058 *** |
| (−23.872) | (−23.006) | (−23.226) | (−21.522) | (−20.975) | (−21.396) | |
| Industry fixed | YES | YES | YES | YES | YES | YES |
| Year fixed | YES | YES | YES | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.323 | 0.294 | 0.308 | 0.291 | 0.263 | 0.282 |
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 6.
Propensity score matching method (PSM).
| (1) | (2) | (3) | |
|---|---|---|---|
| Output1 | Output2 | Output3 | |
| Degree | −2.217 *** | −1.116 *** | −1.106 *** |
| (−5.338) | (−5.706) | (−4.546) | |
| Age | −0.000 | −0.000 * | −0.000 |
| (−1.379) | (−1.764) | (−0.895) | |
| Growth | −0.001 ** | −0.000 ** | −0.001 *** |
| (−2.548) | (−2.046) | (−2.820) | |
| ROA | −0.001 | −0.002 | 0.001 |
| (−0.212) | (−1.222) | (0.531) | |
| Lev | −0.003 *** | −0.002 *** | −0.001 |
| (−2.938) | (−4.231) | (−1.520) | |
| Top | 0.004 *** | 0.001 ** | 0.003 *** |
| (3.125) | (2.283) | (3.271) | |
| Size | 0.004 *** | 0.002 *** | 0.002 *** |
| (10.222) | (9.086) | (9.926) | |
| Indep | 0.008 * | 0.004 | 0.004 |
| (1.698) | (1.460) | (1.550) | |
| TobinQ | 0.001 *** | 0.000 *** | 0.000 *** |
| (4.808) | (5.024) | (4.205) | |
| Board | 0.001 | 0.000 | 0.001 |
| (0.726) | (0.686) | (0.672) | |
| Dual | 0.000 | −0.000 | 0.000 |
| (0.779) | (−0.498) | (1.556) | |
| _cons | −0.088 *** | −0.040 *** | −0.049 *** |
| (−8.407) | (−7.651) | (−8.145) | |
| Industry fixed | YES | YES | YES |
| Year fixed | YES | YES | YES |
| Obs | 4394 | 4394 | 4394 |
| Adjusted R2 | 0.281 | 0.250 | 0.269 |
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 7.
Instrumental variable method (2SLS): IV Degree.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Degree | Output1 | Output2 | Output3 | |
| IV Degree | 0.970 *** | |||
| (19.496) | ||||
| Degree | −5.969 *** | −3.181 *** | −2.644 *** | |
| (−7.335) | (−8.142) | (−5.760) | ||
| Age | 0.000 | 0.000 *** | 0.000 *** | 0.000 *** |
| (1.093) | (3.436) | (3.145) | (2.911) | |
| Growth | 0.000 | −0.001 ** | −0.000 * | −0.000 *** |
| (0.563) | (−2.489) | (−1.651) | (−3.041) | |
| ROA | −0.001 *** | −0.000 | −0.002 ** | 0.002 * |
| (−4.220) | (−0.096) | (−2.560) | (1.884) | |
| Lev | 0.000 | −0.004 *** | −0.002 *** | −0.001 *** |
| (0.283) | (−6.197) | (−7.405) | (−4.488) | |
| Top | 0.000 | 0.002 *** | 0.001 | 0.001 *** |
| (0.999) | (2.691) | (1.574) | (3.088) | |
| Size | −0.000 *** | 0.004 *** | 0.002 *** | 0.002 *** |
| (−16.638) | (21.515) | (20.101) | (20.961) | |
| Indep | −0.001 *** | 0.006 ** | 0.003 ** | 0.002 |
| (−5.329) | (1.994) | (2.442) | (1.536) | |
| TobinQ | −0.000 *** | 0.001 *** | 0.000 *** | 0.000 *** |
| (−6.202) | (8.160) | (9.078) | (6.922) | |
| Board | −0.000 | −0.000 | 0.000 | −0.001 |
| (−0.738) | (−0.292) | (0.256) | (−1.206) | |
| Dual | −0.000 | 0.001 *** | 0.000 | 0.000 *** |
| (−1.269) | (2.655) | (1.344) | (3.494) | |
| _cons | 0.003 *** | −0.090 *** | −0.039 *** | −0.049 *** |
| (16.692) | (−16.020) | (−15.181) | (−15.462) | |
| Industry fixed | YES | YES | YES | YES |
| Year fixed | YES | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.293 | 0.293 | 0.249 | 0.291 |
| Chi-squared value of Hausman | 18.24 *** | 25.82 *** | 9.90 *** | |
| F-value of the first stage | 128.510 | |||
| LM value of Kleibergen–Paap rk | 319.050 *** | |||
| Wald F value of Kleibergen–Paap rk | 380.082 | |||
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 8.
Instrumental variable method (2SLS): IV Traffic.
| (1) | (2) | (3) | (4) | |
| Degree | Output1 | Output2 | Output3 | |
| IV Traffic | 0.000 *** a | |||
| (4.874) | ||||
| Degree | −22.418 *** | −11.614 *** | −10.627 *** | |
| (−4.213) | (−4.334) | (−3.909) | ||
| Age | 0.000 | 0.001 *** | 0.000 ** | 0.000 *** |
| (1.038) | (2.855) | (2.572) | (2.718) | |
| Growth | −0.000 | −0.001 * | −0.000 | −0.000 ** |
| (−0.078) | (−1.670) | (−1.043) | (−2.210) | |
| ROA | −0.001 *** | −0.010 ** | −0.007 *** | −0.003 |
| (−4.933) | (−2.177) | (−3.146) | (−1.163) | |
| Lev | −0.000 | −0.004 *** | −0.002 *** | −0.001 *** |
| (−0.631) | (−3.701) | (−4.236) | (−2.923) | |
| Top | 0.000 | 0.003 *** | 0.001 * | 0.002 *** |
| (1.543) | (2.711) | (1.859) | (3.118) | |
| Size | −0.000 *** | 0.003 *** | 0.001 *** | 0.002 *** |
| (−15.841) | (4.286) | (3.267) | (4.960) | |
| Indep | −0.001 *** | −0.006 | −0.002 | −0.003 |
| (−5.744) | (−1.099) | (−0.971) | (−1.157) | |
| TobinQ | −0.000 *** | 0.000 | 0.000 | 0.000 |
| (−6.188) | (0.841) | (0.870) | (0.720) | |
| Board | −0.000 | −0.001 | −0.000 | −0.001 |
| (−1.473) | (−0.853) | (−0.473) | (−1.560) | |
| Dual | −0.000 | 0.000 | 0.000 | 0.000 * |
| (−0.754) | (1.002) | (0.088) | (1.829) | |
| _cons | 0.004 *** | −0.029 | −0.008 | −0.019 * |
| (23.150) | (−1.476) | (−0.822) | (−1.910) | |
| Industry fixed effect | YES | YES | YES | YES |
| Year fixed effect | YES | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.269 | −0.597 | −0.912 | −0.337 |
| Chi-squared value of Hausman | 35.27 *** | 44.30 *** | 24.18 *** | |
| F-value of the first stage | 109.120 | |||
| LM value of Kleibergen–Paap rk | 23.740 *** | |||
| Wald F value of Kleibergen–Paap rk | 23.757 | |||
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses. a the coefficient value is 0.000002 due to the small calculated value of degree centrality.
Table 9.
Heckman’s two-stage method.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Disclosure | Output1 | Output2 | Output3 | |
| Ev_dis | 2.865 *** | |||
| (17.617) | ||||
| Degree | −2.312 *** | −1.162 *** | −1.105 *** | |
| (−11.661) | (−12.524) | (−9.773) | ||
| IMR | 0.001 | 0.001 ** | −0.000 | |
| (0.934) | (2.016) | (−0.243) | ||
| Age | 0.060 *** | 0.000 *** | 0.000 *** | 0.000 *** |
| (6.023) | (3.397) | (3.411) | (2.592) | |
| Growth | 0.037 * | −0.001 ** | −0.000 | −0.000 *** |
| (1.761) | (−2.480) | (−1.591) | (−3.074) | |
| ROA | −0.630 *** | 0.002 | −0.001 | 0.003 *** |
| (−4.285) | (0.950) | (−1.635) | (2.857) | |
| Lev | 0.099 * | −0.003 *** | −0.002 *** | −0.001 *** |
| (1.954) | (−6.407) | (−7.707) | (−4.609) | |
| Top | 0.024 | 0.002 ** | 0.001 | 0.001 *** |
| (0.457) | (2.490) | (1.348) | (2.909) | |
| Size | 0.004 | 0.005 *** | 0.002 *** | 0.003 *** |
| (0.479) | (25.749) | (24.192) | (25.138) | |
| Indep | −0.074 | 0.008 *** | 0.005 *** | 0.004 ** |
| (−0.438) | (2.950) | (3.548) | (2.283) | |
| TobinQ | −0.002 | 0.001 *** | 0.000 *** | 0.000 *** |
| (−0.382) | (10.162) | (11.604) | (8.422) | |
| Board | 0.349 *** | 0.000 | 0.000 | −0.000 |
| (7.282) | (0.082) | (0.878) | (−1.101) | |
| Dual | −0.088 *** | 0.001 *** | 0.000 | 0.001 *** |
| (−4.942) | (2.766) | (1.292) | (3.728) | |
| _cons | −2.407 *** | −0.105 *** | −0.048 *** | −0.055 *** |
| (−10.538) | (−23.633) | (−22.843) | (−22.767) | |
| Industry fixed effect | YES | YES | YES | YES |
| Year fixed effect | YES | YES | YES | YES |
| Obs | 34,493 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.133 | 0.323 | 0.294 | 0.308 |
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 10.
Mechanism of mediating test: R&D investment.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| RD | Output1 | Output2 | Output3 | |
| Degree | −1.941 *** | −2.265 *** | −1.132 *** | −1.088 *** |
| (−3.849) | (−11.484) | (−12.327) | (−9.640) | |
| RD | 0.025 *** | 0.017 *** | 0.008 *** | |
| (8.694) | (10.527) | (5.769) | ||
| Age | −0.007 *** | 0.001 *** | 0.000 *** | 0.000 *** |
| (−16.094) | (4.654) | (4.925) | (3.530) | |
| Growth | −0.001 | −0.001 ** | −0.000 | −0.000 *** |
| (−1.636) | (−2.402) | (−1.518) | (−2.986) | |
| ROA | −0.075 *** | 0.004 ** | 0.000 | 0.004 *** |
| (−8.502) | (2.124) | (0.210) | (3.357) | |
| Lev | −0.042 *** | −0.002 *** | −0.001 *** | −0.001 *** |
| (−18.469) | (−4.369) | (−4.999) | (−3.347) | |
| Top | −0.019 *** | 0.003 *** | 0.001 ** | 0.002 *** |
| (−9.711) | (3.065) | (2.153) | (3.259) | |
| Size | 0.003 *** | 0.005 *** | 0.002 *** | 0.003 *** |
| (7.402) | (25.413) | (23.720) | (24.904) | |
| Indep | 0.015 ** | 0.008 *** | 0.004 *** | 0.003 ** |
| (2.391) | (2.839) | (3.409) | (2.205) | |
| TobinQ | 0.004 *** | 0.001 *** | 0.000 *** | 0.000 *** |
| (10.200) | (8.814) | (9.698) | (7.574) | |
| Board | 0.004 ** | −0.000 | 0.000 | −0.000 |
| (2.372) | (−0.244) | (0.291) | (−1.152) | |
| Dual | 0.004 *** | 0.001 ** | 0.000 | 0.000 *** |
| (4.883) | (2.499) | (1.041) | (3.436) | |
| _cons | −0.042 *** | −0.103 *** | −0.046 *** | −0.054 *** |
| (−5.158) | (−23.684) | (−22.751) | (−23.099) | |
| Ind_eff test (p-val) | 0.001 | 0.000 | 0.002 | |
| Industry fixed effect | YES | YES | YES | YES |
| Year fixed effect | YES | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.439 | 0.327 | 0.302 | 0.309 |
** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 11.
Moderated mediating effect of ownership. The dependent variable is Output1.
| (1) | (2) | (3) | |
|---|---|---|---|
| Output1 | RD | Output1 | |
| Degree | −2.258 *** | −1.243 * | −2.211 *** |
| (−10.959) | (−1.704) | (−11.252) | |
| State | 0.000 | 0.001 | −0.001 *** |
| (0.854) | (0.773) | (−4.003) | |
| Degree × State | −0.126 | −1.607 * | |
| (−0.395) | (−1.908) | ||
| RD | 0.014 *** | ||
| (5.115) | |||
| RD × State | 0.057 *** | ||
| (7.981) | |||
| Age | 0.000 *** | −0.006 *** | 0.000 *** |
| (2.612) | (−14.587) | (3.945) | |
| Growth | −0.001 ** | −0.001 * | −0.001 ** |
| (−2.431) | (−1.735) | (−2.219) | |
| ROA | 0.002 | −0.075 *** | 0.003 |
| (1.152) | (−8.471) | (1.601) | |
| Lev | −0.004 *** | −0.041 *** | −0.003 *** |
| (−6.532) | (−18.427) | (−4.670) | |
| Top | 0.002 ** | −0.018 *** | 0.002 *** |
| (2.196) | (−9.046) | (2.841) | |
| Size | 0.005 *** | 0.003 *** | 0.005 *** |
| (25.641) | (7.456) | (25.582) | |
| Indep | 0.008 *** | 0.014 ** | 0.009 *** |
| (2.954) | (2.315) | (3.067) | |
| TobinQ | 0.001 *** | 0.004 *** | 0.001 *** |
| (10.124) | (10.116) | (8.960) | |
| Board | −0.000 | 0.004 ** | −0.000 |
| (−0.235) | (2.460) | (−0.342) | |
| Dual | 0.001 *** | 0.004 *** | 0.001 *** |
| (3.099) | (4.709) | (3.229) | |
| _cons | −0.104 *** | −0.043 *** | −0.103 *** |
| (−23.693) | (−5.246) | (−23.633) | |
| Industry fixed effect | YES | YES | YES |
| Year fixed effect | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.323 | 0.439 | 0.332 |
* significant at 10% level; ** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Table 12.
Moderated mediating effect of ownership. The dependent variables are Output2 and Output3, respectively.
Table 12.
Moderated mediating effect of ownership. The dependent variables are Output2 and Output3, respectively.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Output2 | Output2 | Output3 | Output3 | |
| Degree | −1.075 *** | −1.097 *** | −1.125 *** | −1.070 *** |
| (−11.683) | (−12.067) | (−9.372) | (−9.485) | |
| State | 0.001 *** | −0.000 *** | −0.000 | −0.001 *** |
| (2.591) | (−3.688) | (−0.322) | (−3.542) | |
| Degree × State | −0.200 | 0.046 | ||
| (−1.334) | (0.254) | |||
| RD | 0.010 *** | 0.004 *** | ||
| (6.930) | (3.059) | |||
| RD × State | 0.036 *** | 0.020 *** | ||
| (8.914) | (5.657) | |||
| Age | 0.000 | 0.000 *** | 0.000 *** | 0.000 *** |
| (1.545) | (3.527) | (2.669) | (3.418) | |
| Growth | −0.000 | −0.000 | −0.000 *** | −0.000 *** |
| (−1.463) | (−1.154) | (−3.058) | (−2.933) | |
| ROA | −0.001 | −0.000 | 0.003 *** | 0.003 *** |
| (−1.164) | (−0.512) | (2.803) | (3.025) | |
| Lev | −0.002 *** | −0.002 *** | −0.001 *** | −0.001 *** |
| (−8.113) | (−5.571) | (−4.525) | (−3.457) | |
| Top | 0.000 | 0.001 | 0.001 *** | 0.002 *** |
| (0.693) | (1.603) | (2.849) | (3.239) | |
| Size | 0.002 *** | 0.002 *** | 0.003 *** | 0.003 *** |
| (23.948) | (23.847) | (25.155) | (25.099) | |
| Indep | 0.005 *** | 0.005 *** | 0.004 ** | 0.004 ** |
| (3.509) | (3.706) | (2.304) | (2.360) | |
| TobinQ | 0.000 *** | 0.000 *** | 0.000 *** | 0.000 *** |
| (11.585) | (9.936) | (8.361) | (7.634) | |
| Board | 0.000 | −0.000 | −0.000 | −0.000 |
| (0.096) | (−0.041) | (−1.023) | (−1.093) | |
| Dual | 0.000 ** | 0.000 ** | 0.000 *** | 0.001 *** |
| (2.284) | (2.451) | (3.583) | (3.692) | |
| _cons | −0.047 *** | −0.046 *** | −0.055 *** | −0.054 *** |
| (−22.643) | (−22.566) | (−23.224) | (−23.170) | |
| Industry fixed effect | YES | YES | YES | YES |
| Year fixed effect | YES | YES | YES | YES |
| Obs | 11,207 | 11,207 | 11,207 | 11,207 |
| Adjusted R2 | 0.295 | 0.312 | 0.308 | 0.311 |
** significant at 5% level; *** significant at 1% level. White-robust adjusted T values are shown in parentheses.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Cheng, L.; Wu, L. Firms’ Position in the Supply Chain Network, R&D Input, and Innovation Output: Striving for the Top or Settling in the Corner? Implications for Sustainable Growth and Adaptive Capacity. Sustainability 2025, 17, 1729. https://doi.org/10.3390/su17041729
AMA Style
Cheng L, Wu L. Firms’ Position in the Supply Chain Network, R&D Input, and Innovation Output: Striving for the Top or Settling in the Corner? Implications for Sustainable Growth and Adaptive Capacity. Sustainability. 2025; 17(4):1729. https://doi.org/10.3390/su17041729
Chicago/Turabian StyleCheng, Le, and Liyuan Wu. 2025. "Firms’ Position in the Supply Chain Network, R&D Input, and Innovation Output: Striving for the Top or Settling in the Corner? Implications for Sustainable Growth and Adaptive Capacity" Sustainability 17, no. 4: 1729. https://doi.org/10.3390/su17041729
APA StyleCheng, L., & Wu, L. (2025). Firms’ Position in the Supply Chain Network, R&D Input, and Innovation Output: Striving for the Top or Settling in the Corner? Implications for Sustainable Growth and Adaptive Capacity. Sustainability, 17(4), 1729. https://doi.org/10.3390/su17041729
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.