Next Article in Journal
Regional Features and National Differences in Population Distribution in China’s Border Regions (2000–2015)
Previous Article in Journal
Visions and Expectations for the Norwegian Bioeconomy
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

External Knowledge Sourcing and Green Innovation Growth with Environmental and Energy Regulations: Evidence from Manufacturing in China

School of Economics and Management, Harbin Engineering University, Harbin 150001, China
*
Author to whom correspondence should be addressed.
Sustainability 2017, 9(3), 342; https://doi.org/10.3390/su9030342
Submission received: 15 December 2016 / Revised: 25 January 2017 / Accepted: 16 February 2017 / Published: 25 February 2017
(This article belongs to the Section Economic and Business Aspects of Sustainability)

Abstract

:
This paper adopts the slacks-based measure-directional distance function (SBM-DDF, 2009) method for deriving the “Green Innovation Growth” rates of 28 manufacturing industries in China. The results indicate that the overall level of green innovation growth in China’s manufacturing is relatively low, with a declining trend. The tradeoffs among energy, environment and economy are rather sharp, and the “Porter Effect (1995)” (environmental regulation will promote green technology innovation) is not currently realized quickly in manufacturing. These evaluations imply an unsustainable development model in China, with significant differences among industries. By using a dynamic panel threshold model and employing an industry-level panel dataset for 2008–2014, we show that external knowledge sourcing has a significant negative impact on green innovation growth but with different constraints on R&D levels among industries. With the strengthening of R&D levels, gradually surpassing “critical mass”, the negative role of external knowledge sourcing in driving this mechanism becomes smaller and smaller; it has a non-linear relationship with the “threshold effect”. Consequently, we provide insights into the relationship among energy consumption, environmental pollution and technology innovation, and show how the heterogeneity of the R&D threshold affects differences in external knowledge sourcing and green innovation growth. These insights lead to a better understanding of the driving force, realizing path and policy design for green innovation growth.

1. Introduction

Economic growth and industrial development in developing countries have long relied mainly on a resource path that entails high investment, high energy consumption, high emissions, low quality, low efficiency and low output [1,2]. Although this extensive mode of industrial growth contributes significantly to economic growth, it has also increased energy consumption, environmental pollution, and pollution emissions. Moreover, energy consumption has been approaching the limit of environmental carrying capacity. After more than 30 years of rapid development, China has become the world’s second largest economy. However, it has paid a huge price in terms of energy, environment and social welfare, and has surpassed the United States as the world’s largest pollution emitter. As the core of China’s economy, manufacturing plays a significant role in industrial economic growth, energy consumption and environmental pollution, and its negative impact on both energy and the environment cannot be ignored. For instance, the problem of haze pollution, which is mainly composed of PM10 and PM2.5, is particularly severe in China. Areas with a high concentration of manufacturing are shrouded in “the Gray Great Wall”, stretching for thousands of kilometers. The manufacturing industry is responsible for 469,366 billion cubic meters of total volume of industrial waste gas emissions, accounting for 67.61% of China’s total emissions; manufacturing consumes 2450.51 million tons of standard coal, accounting for 57.55% of China’s total energy consumption [3,4]. The high energy consumption and high emissions of Chinese manufacturing have come under great pressure and have become the focus of global attention.
Therefore, it is urgent that China transition to a mode of manufacturing that prioritizes energy saving and environmental protection, and this transition must be driven by the integration of green growth and innovation. “Green growth” is a mode of pursuing economic growth and development while preventing environmental deterioration, biodiversity loss and unsustainable use of natural resources [5,6]. Due to the need for balanced consideration of the relationship between growth and the environment, it has become the new preference of many countries and international organizations (United Nations Environment Programme, Organization for Economic Co-operation and Development and European Union) and is regarded as an important strategy in achieving sustainable development [7,8]. Meanwhile, if innovation is the driving force behind development, green innovation is the driving force behind sustainable development. Green innovation is the innovation pathway to reducing pollution, saving resources and improving the environment; it has both ecological and economic benefits, fundamentally promotes sustainable development and enhances the competitiveness of enterprises [9]. To achieve sustainable development, it is increasingly important that industries develop and strengthen green technology innovation [10,11].
In this context, Green Innovation Growth is a complex developmental process based on technological innovation and knowledge accumulation. Due to the uneven distribution of industrial innovation resources in China, manufacturing is limited by its own resources: the Research and Development (R&D) foundation is very weak; and R&D intensity is only approximately 0.91%, which is well below the average level of 2.5%–4% in developed countries. Thus, the corresponding innovation-driving effect is limited. With the development of open innovation, Chinese manufacturing enterprises are creating and acquiring knowledge through external sourcing to produce new knowledge flows to build and update their technology stocks [12]. Furthermore, by cooperating with enterprises, universities and research institutes with technical advantages, including cooperative R&D, R&D outsourcing and commissioned technology development, can achieve innovation transformation [13]. Therefore, although a great deal of innovation comes from entrepreneurs and intrapreneurs independent of external knowledge, external knowledge sourcing has become one of the main paths by which the Chinese manufacturing industry pursues green innovation growth. Chinese manufacturing needs a clear understanding of the role (and adverse effects) of external knowledge sources in the growth of green innovation; it must explore new systems of technological innovation and environmental regulation, establish a model of green and eco-industrial development, and contribute to global green development.
This paper is organized as follows. Section 2 provides an overview of the related literature. Section 3 presents the methodology and empirical specifications, as well as the estimation variables and our database. Section 4 discusses the evaluations of Green Innovation Growth under environmental and energy regulations. The empirical results and discussions of dynamic threshold effects between external knowledge sourcing and green innovation growth are presented in Section 5. Finally, Section 6 provides conclusions and recommendations.

2. Literature Review

Scholars have extensively studied and discussed open innovation, external knowledge sourcing and diversified innovation paths. It is worth noting that more and more scholars are beginning to examine the impact of external knowledge sourcing on innovation performance. For example, Reagans and McEvily [14] proposed that external R&D institutions could promote the flow of knowledge, especially implicit knowledge. Meanwhile, external connections can bring all types of valuable innovation resources to the firm, reduce environmental uncertainty, promote the development of innovation activities and ultimately improve innovation performance. Hewitt-Dundas [15] noted that the resources and abilities obtained from an external innovation partner could encourage enterprises’ innovation behavior and improve innovation ability. Cooperative partners may also benefit from complementary and shared innovative resources to make up for their individual lack of enterprise resources and limited capacity [16]. Technical cooperation can help enterprises cope with global competition, financial and economic uncertainty, and rapid and fundamental technological change, and cooperation can reduce investment risk [17,18]. Based on the data of Austria enterprises, Todtling et al. [19] showed that external cooperation innovation in Austria has a significant effect on innovation performance. Furthermore, Liu [20] found that foreign technical cooperation has positive impacts on China’s energy industry performance, which are more obvious in developed regions of China.
However, unlike most studies, De Man and Duysters [21] found that an enterprise’s external cooperation could have positive, negative and neutral effects on innovation performance. Vega-Jurado et al. [22] observed that the influence of external sourcing on innovation performance is very weak in manufacturing of Spain. This also happens in Korea; interestingly, there is no significant impact of external knowledge-oriented approach while there is positive impact of internal knowledge-oriented approach [23]. The high transaction costs of searching for and selecting partners, configuring additional resources, and coordinating and managing R&D activities of cooperative members reduce the benefits of external innovation activities and challenge organizational structure. Indeed, external cooperation with competitors or institutions has a negative impact on innovation performance [24], and the failure rate of cooperative R&D is higher [25]. Other studies suggest that there is an inverted U-shaped relationship between external knowledge resources and innovation performance in Europe, meaning that a threshold is created by a critical point [26,27]. Beyond this threshold, the enterprise’s innovation performance is reduced, and with the increase in the stock of technology, external knowledge sourcing will lead to increased opportunity costs, and the negative effect will be more significant. In other words, internal and external R&D may be complementary to varying degrees, but they may also substitute for each other [28,29].
In fact, no matter what type of innovation path an enterprise follows, it must adapt to the base of technology and resources in its industry, that is, the concept of development threshold [30]. Technological innovation depends on the accumulation of knowledge, and the level of R&D determines the ability to innovate [31,32]. In turn, a strong technological base allows the extension of a firm’s technological capabilities and increases the odds of developing and realizing new products [33]. In short, the foundation of R&D is support for technological innovation, which has a significant impact on innovation output. Thus, is innovation in external knowledge sourcing really effective at different levels of R&D? How does the heterogeneity of R&D thresholds affect the mechanism by which external knowledge sourcing shapes green innovation growth, and what type of heterogeneity has the greatest impact?
This study attempts to fill some of the abovementioned gaps. First, we clarify the relationship among energy consumption, environmental pollution and technological innovation by developing the slacks-based measure-directional distance function (SBM-DDF) to calculate the green innovation growth index of China’s manufacturing in 2008–2014; we further analyze its structure across industries with environmental and energy regulations. Second, any innovation path needs to correspond to the industry’s level of technological development; only with a minimal foundation of R&D can the industry effectively follow a certain innovation path. We introduce the threshold factor of R&D level into the complex mechanism linking external knowledge sourcing and green innovation growth to examine the dynamic threshold effect and its heterogeneity, and we try to explore the role of driving factor, realizing path and policy design in green innovation growth. Accordingly, this research provides key insights on the implications of green innovation growth.

3. Methodology and Model

3.1. The SBM-DDF Model

Many scholars have studied green technology innovation [34,35,36]. However, prior studies have the following fundamental problems: Most adopt the non-parametric method of Malmquist–Luenberger, based on the directional distance function proposed by Chung et al. [37], which presents environmental pollution as an undesirable output. However, due to the existence of radial and directional deviation, production efficiency is overestimated and cannot be non-proportionally adjusted simultaneously [38]. The non-radial and non-oriented method of Slacks-Based Measure (SBM), based on the slack variable [39], corrected the problem to some extent, but it is still unable to adjust the input and output radially, and the efficiency is overvalued. Consequently, Fukuyama and Weber [38] proposed a new method that combines the SBM with the direction vector model (SBM-DDF); this can solve the above problems and better fit the input–output relationship under the constraints of environment and energy. The methodological contributions of SBM-DDF are as follows. (i) Undesirable outputs: SBM-DDF examines the decrease in undesirable outputs while examining the increase in output; (ii) Redundancy: This method can get input redundancy, and output deficiency has a strong policy implications; (iii) Weighting: SBM-DDF can be used to set the weight of the input and output variables, which is conducive to the analysis framework of the policy factors. It can reveal the government’s motivation and tendency in the green development and its different preference structure; (iv) Validity and authenticity: Green efficiency is essentially to examine the contribution of environmental and energy input to economic output, and it needs to be measured and decomposed in the framework of total factor. In other words, SBM-DDF can get a real sense of green efficiency [40]. Therefore, this paper uses the new development model of the SBM-DDF to measure green innovation growth in manufacturing (Details are provided in Appendix A).

3.2. Indicators for Evaluating Green Innovation Growth

3.2.1. Input Indicators

In this paper, the investment of green innovation in manufacturing mainly includes R&D capital stock (because early investment in R&D will affect the current input, we estimate the R&D capital stock using the perpetual inventory method), full-time R&D personnel and energy consumption [41,42]. It should be noted that for green innovation growth based on energy efficiency, we adopt the indicators of comprehensive energy consumption (million tons of standard coal) in manufacturing to measure the energy input of green innovation growth.

3.2.2. Output Indicators

Green innovation outputs include desirable outputs and undesirable outputs. Desirable outputs reflect the willingness of industrial technology to innovate. We choose the number of patent applications to illustrate the level of direct innovation in manufacturing. Meanwhile, we take new product sales revenue as the indirect output indicator, representing the commercial value of technology innovation [43].
Undesirable outputs, as the “by-products” of the desirable outputs, directly or indirectly damage performance during the process of technological innovation (pollution emissions, environmental effects, etc.). We select undesirable outputs from the main pollution sources of manufacturing, including the amount of discharged pollutants (industrial wastewater, industrial waste gas, and industrial solid waste), as well as SO2, soot and dust emissions [44,45]. Thus, we fit these indicators into a comprehensive index of environmental pollution using the entropy method (The indexes are provided in Supplementary Materials):
First, data standardization:
P i j = X i j i = 1 n X i j , ( i = 1 , 2 , n , j = 1 , 2 , m )
where n is the number of industries, m is the number of indicators, X i j is the pollution value of indicator j industry i , and i = 1 n X i j is the sum of the pollution value.
Second, calculating the entropy of indicator j :
e j = k i = 1 n p i j ln ( p i j ) , k > 0 , k = 1 / ln ( n ) , e j   0
Then, calculating the difference coefficient of indicator j :
g j = 1 e j m E e , E e = j = 1 m e j , 0 g j 1 , j = 1 m g j = 1
The weight values of the pollution indicators are obtained:
w i = g j j = 1 m g j ( 1 j m )
Finally, we can obtain the comprehensive index of environmental pollution each year:
s i = j = 1 m w j × p i j ( i = 1 , 2 , n )
The evaluation indicators system of green innovation growth is shown in Table 1.

3.3. Dynamic Panel Threshold Model

According to previous studies [30], the innovation path has its own development threshold, which is compatible with the level of technological development. To examine the non-linear relationship of external knowledge sourcing and green innovation growth, we employ the dynamic panel threshold model based on the method proposed by Hansen [46]. (Two statistical tests are needed with the panel threshold model: (1) To test for a threshold with the null hypothesis of no threshold effect: H0: β11 = β12. (2) When there is a threshold effect (β11β12), we need to use the likelihood ratio statistic to test whether γ ^ is consistent for γ0 (the true value of γ) with the null hypothesis: H 0 : γ = γ 0 . More details can be found in Hansen [46].) It is worth noting that the Hansen [46] method is only applicable to the non-dynamic panel model; it cannot reflect dynamic change or the lag effect of the sample object, and it also ignores the processing of endogenous variables. Thus, we add the lag variable (we find that the instrumental variables of the Second-order lags are significant and the fitting degree is the best, so we choose the first-order and the Second-order lags) to control the lag effect and include dynamic factors based on dynamic panel estimation. We first estimate the dynamic panel threshold and then use the “first-order difference GMM” [47] to estimate the parameters between different threshold regimes. Using this method, we investigate the effect of external knowledge sourcing on green innovation growth with different thresholds on R&D level between industries. Taking a single-threshold model as an example, the specification is as follows:
G I G i t = θ + α 1 G I G i t 1 + α 2 G I G i t 2 + α 3 R E G i t + α 4 F D I i t + α 5 A G G i t + α 6 S C A i t + α 7 E N D i t + β 1 E K S i t 1 I ( L E V i t γ ) + β 2 E K S i t 1 I ( L E V i t > γ ) + μ i + v t + ε i t
where I ( ) is the indicator function, and γ is the threshold value. The observations are divided into two regimes depending on whether the threshold variable L E V i t is lower or higher than the threshold value γ . The regimes are distinguished by differing regression slopes, β1 and β2. The subscripts i and t denote province and year. μ i is a specific effect of the individual; ν t is a specific effect of time; and ε i t is a random disturbance. In addition, we allow a one-year lag [48] in the EKS variable (as green innovation growth requires changes in technological processes that require time to become effective, it appears reasonable that there exists a time lag between innovation path and changes in green innovation growth).
We take the index of green innovation growth (GIG) as the dependent variable and external knowledge sourcing (EKS) as the independent variable. With open innovation, the source of innovation is diversified, and external knowledge sourcing is one way for industries to acquire knowledge and develop new technologies through cooperation with external organizations [12,28]. External knowledge sourcing primarily measures R&D costs, including cooperation expenses paid by the industry to external organizations such as enterprises, universities and research institutes [22,49]. Thus, we use external R&D stock (we estimate the external R&D capital stock using the perpetual inventory method) to represent the degree of external knowledge sourcing. The threshold variable is the R&D level (LEV). “Financial Times” reporter Peter Marsh has suggested that “China has won the manufacturing, but lost the R&D” [50]. Fortunately, the Chinese government has a clear understanding of this and believes that China’s manufacturing industry is large but not strong, and the foundation of R&D is still weak. Different levels of R&D have heterogeneous impacts on green innovation growth. We use the ratio of the R&D stock value to the main business income value as a proxy indicator.
In addition, we add a series of control variables. Based on the Porter hypothesis, the “compensation effect” and “learning effect” of environmental regulation will promote green technology innovation [51]. However, the development of the industry bears high environmental costs, and environmental regulation may also be inversely proportional to green innovation growth [48,52,53]. We use investment in the treatment of environmental pollution by industry as a proxy for environmental regulation (REG). Foreign direct investment (FDI) will increase the host country’s environmental pressure (Pollution Haven Hypothesis) [54] but also promote industrial green production and reduce environmental pollution (Pollution Halo Hypothesis) [55] by introducing the technologies of green energy and clean production. We use the ratio of industrial assets from foreign investment to total industrial assets to represent the FDI effect. Industrial agglomeration (AGG) is measured by “the number of manufacturing enterprises”. Industry scale (SCA) is also one of the important factors used by scholars [56]. We define the industry scale as the ratio of its total asset value to the number of enterprises. The share of production factors, such as capital and labor, has a significant impact on efficiency. We use the ratio of total asset value to average annual number of employees to reflect the structure of industrial factor endowment (END).

3.4. Data Sources and Processing

The database of this paper is industry-level panel data for 2008–2014 obtained from the National Bureau of Statistics of China. We selected 28 manufacturing industries for the sample. Given the influence of price factors on R&D expenditures, we convert the nominal value of R&D expenditures to the actual value using a “price index for R&D”—the weighted average of the consumer price index and the fixed-asset investment price index, with the weights estimated to be 0.55 and 0.45, respectively. Investments in the treatment of environmental pollution and total asset value are weighted with the fixed-asset price index, and the main business income values are weighted with the PPI; the base period is 2008. Table 2 summarizes the descriptive statistics of variables.

4. Measuring Green Innovation Growth

We measure green innovation growth in relation to environmental and energy regulations for manufacturing based on the SBM-DDF model, and the results are shown in Table 3.
The overall level of green innovation growth in China’s manufacturing sector is relatively low, with an average of 0.6137. There is ample room for the development of green innovation growth in China, and the heterogeneity among industries is significant. Based on the gap in green innovation growth among manufacturing industries, we create three groups: the high green innovation group, middle green innovation group and low green innovation group. The industries in the high green innovation group, with a mean of 0.9775, can reach the forefront of green innovation growth, while the mean of the low green innovation industries is only 0.4381, which is significantly lower than the high green innovation group. In the high green innovation group, which includes industries that manufacture Tobacco; Articles for Culture, Education, Arts and Crafts; Sport and Entertainment Activities; Electrical Machinery and Apparatus; Computers, Communication and Other Electronic Equipment; Measuring Instruments; and Machinery and Transport Equipment, green practices, are the strongest, and the driving effect of green technology innovation is remarkable. Most of these industries are technology-intensive, their input–output efficiency (in terms of human and capital resources in technological innovation) is higher, and their energy consumption and environmental pollution are relatively low. Light industries, such as the Manufacture of Printing and Textiles, are at the middle level of green innovation, which also includes a number of high-tech industries and equipment manufacturing, such as the Manufacture of Metal Products, General Purpose Machinery and Medicines. To realize a transformation in green technology innovation, these industries can be further upgraded to achieve green growth, but they must pay attention to energy saving and improve the efficiency of their pollution emissions through more green technology innovations. Meanwhile, some high-tech manufacturing sectors have not yet realized the effects of low pollution and low energy consumption, indicating a lack of deep integration between technological innovation and green growth. The manufacture of Chemical Raw Materials and Chemical Products; the Processing of Petroleum; Coking; and the Processing of Nuclear Fuel have the lowest level of green innovation growth. The driving force behind technological innovation in these industries is insufficient, and they consume a large amount of energy and produce high emissions, causing serious damage to the ecological environment.
Regarding development trends (Figure 1), most manufacturing is not showing growth in green innovation, and there is even a downward trend to a certain extent, with a large decline from 2008 to 2011. In 2012, the Chinese government vigorously implemented an innovation-driven development strategy, and continued to strengthen its green development policy, leading to some improvements to green innovation in manufacturing. However, due to the weak foundation of endogenous innovation and higher operating costs of energy and environmental technology, manufacturing cannot realize the “Porter effect” in a short time. Therefore, manufacturing does not show sustainable green innovation growth in 2013–2014. The concrete distribution of best practice manufacturing in Green innovation growth (2008–2014) is presented in Supplementary Materials.

5. Empirical Results and Discussions

To examine how external knowledge sourcing affects green innovation growth given the heterogeneity of R&D thresholds, we first estimate the number of thresholds to determine the form of the model. Then, the F statistic and p-value are obtained by the bootstrapping method (Table 4 and Table 5). We find that a single threshold is significant at the 5% level and the double threshold is significant at the 1% level, while the triple threshold is not significant because the bootstrap p-value is 0.954. Therefore, further analysis is based on the double threshold model.
According to threshold theory (details for estimation techniques of panel threshold models are provided in Hansen [46]), the effects of external knowledge sourcing on green innovation growth include a significant double threshold for R&D level: the threshold values are 0.001 and 0.007, which are within the 95% confidence intervals [0.001, 0.001] and [0.007, 0.007], respectively. Thus, we can divide the sample into different regimes: low level of R&D (LEV ≤ 0.001), middle level of R&D (0.001 < LEV ≤ 0.007) and high level of R&D (LEV > 0.007). The function chart of the threshold variable “likelihood ratio” sequence, with the well-defined change in threshold value, shows the structure of the estimate and the confidence interval (Figure 2). (We use the likelihood ratio L R n ( γ ) to construct the “non-rejection region” showing the valid confidence interval of γ. The “non-rejection region” at the confidence level of 1 α is a series of γ values that meet L R n ( γ ) c ( α ) . That is, when L R n ( γ ) c ( α ) = 2 l n ( 1 1 α ) ( α is a significant level, c ( α ) = 7.35 at 95% confidence level), we cannot reject the original hypothesis that the threshold estimate is equal to the true value ( H 0 : γ = γ 0 ). This corresponds to the horizontal line in the Figure 2.)
Then, based on different regimes, we further use the dynamic estimation method of the first-order difference GMM to estimate the partition coefficient between different regimes.
Table 6 reports the estimates of the effects of external knowledge sourcing on green innovation growth. Overall, external knowledge sourcing has a significant negative effect on green innovation growth in manufacturing. With improvements in R&D level, there are differences in the mechanism of external knowledge sourcing: When the R&D level is lower than the threshold value 0.001, external knowledge sourcing has a significant negative impact on green innovation growth at the 1% level; and between 0.001 and 0.007, this negative impact is significantly weakened, but there is still a certain degree of limitation. When the threshold value is higher than 0.007, the negative effect is further weakened, and external knowledge sourcing is the minimum limit to green innovation growth. The distribution of manufacturing in these different threshold intervals each year is provided in Supplementary Materials. Statistically, this presents a threshold characteristic: when the R&D level is higher than the critical value, the negative effect of external knowledge sourcing on green innovation becomes smaller and smaller.
Normally, to improve knowledge accumulation, external sourcing is necessary. However, we found that this is clearly not true for developing countries, where there is a large “gap”. This finding demonstrates that there may be a substitutional relationship between R&D and external sourcing under the influence of R&D level. When developing countries accumulate knowledge, internal R&D becomes more efficient, and the enterprise can use more R&D to replace external sourcing. At a certain level of R&D, and with an increase in external investment, R&D begins to see diminishing marginal returns; at this point, R&D and external sourcing still essentially have a simple substitutional relationship.
In China, the distribution of innovation resources is significantly uneven across industries. Industries with high R&D in manufacturing focus on R&D investment and endogenous innovation, external knowledge sourcing has long been neglected, and the diversified innovation effect is not fully developed. In contrast, industries with low R&D levels exhibit over-reliance on external cooperation and technology acquisition and will damage their industry’s R&D initiatives and encounter problems with absorption and intellectual property [57], which will eventually reduce the positive spillover effect of external knowledge sourcing (Substitution effect). R&D is the intrinsic basis for the process of green innovation in industry, and corresponding knowledge and technical support are needed for every link in R&D, organization, design and manufacture, sales and other aspects of green innovation. However, most of manufacturing’s R&D foundation is weak due to resource constraints. R&D intensity in China is only approximately 0.91% (2014), and enterprises with R&D institutions accounted for only 13.34% (2014) of all enterprises. Coupled with the urgent problems of energy shortage, expensive resources and environmental protection, there are no strong foundational conditions for manufacturing to expand its external activities, and the corresponding driving effects are limited.
When the industry has accumulated a higher R&D base, industrial knowledge helps external sourcing identify the most worthwhile market issues and then optimize innovation performance, making up for the diminishing marginal returns of R&D investment [58]. Meanwhile, knowledge stock has absolutely independent property rights and can be applied to internal R&D at a lower cost, thereby promoting the commercialization of internal R&D and innovation output [59,60]. Consequently, the industry will promote green innovation growth and gradually realize the transformation to green innovation.
Finally, the lag variable is significant at the 1% level, which indicates that the dynamic panel threshold model constructed in this paper is reasonable. The Sargan tests show Prob > χ2 = 0.0539, which does not reject the original hypothesis that the instrumental variables are reasonable at the 5% level. The AR(1) and AR(2) test (Table 7) also shows that the model setting and the use of a first-order difference GMM are also more reasonable.

6. Conclusions

This study attempts to measure green innovation growth with environmental and energy regulations based on the SBM-DDF method and then takes the heterogeneity of the R&D threshold as the breakthrough point to examine the nonlinear effects of external knowledge sourcing on green innovation growth by developing a dynamic panel threshold model. Accordingly, the research provides interesting insights into the implications of green innovation.
The average level of green innovation growth in China’s manufacturing is relatively low. There is considerable room for improvement in green innovation growth in China, and the heterogeneity among industries is significant. The growth effect of green innovation is strongest in the manufacture of Tobacco; Articles for Culture, Education, Arts and Crafts; Sport and Entertainment Activities; Electrical Machinery and Apparatus; Computers, Communication and Other Electronic Equipment; Measuring Instruments; and Machinery and Transport Equipment. Light industry, such as the Manufacture of Printing and Textiles, is at the middle level of green innovation; it also includes a number of high-tech industries and equipment manufacturing, such as the Manufacture of Metal Products; General Purpose Machinery; and Medicines. The manufacture of Chemical Raw Materials and Chemical Products, Processing of Petroleum, Coking, and Processing of Nuclear Fuel has the lowest level of green innovation growth. These industries consume a large amount of energy and have high pollution emissions; they urgently need to be transformed by green innovation. Moreover, there is very little growth in green innovation in most manufacturing, and there is even a downward trend. The “Porter effect” cannot currently be realized in China. Finally, external knowledge sourcing has a significant negative effect on green innovation growth but with a non-linear relationship to the “threshold effect”: when the R&D level is higher than the critical value, the negative role of external knowledge sourcing in driving this mechanism becomes smaller and smaller.
From the above analysis, this study has several policy implications. First, as the main body that can promote the process of industrialization and modernization in China, manufacturing exhibits a low level of green innovation; it obviously has serious problems of environmental pollution and energy consumption. The manufacturing industry must strengthen its drive toward innovation integration and green development to achieve sustained growth in green innovation. Only by strengthening the integration of innovation-driven and green development can manufacturing achieve sustainable growth. Second, Chinese government needs to give full consideration to the gaps among industries and then implement a system of green innovation. On the one hand, the government should focus on supporting and developing strong green innovation industries and increase incentive policies such as tax, financial and other preferential incentives for these industries to further enhance the level of green innovation growth; it should also encourage these industries to play a leading role in the development of green innovation. On the other hand, for industries with low green innovation, the government should actively coordinate development planning and flexibly choose diversified environmental regulation tools, such as environmental taxes, sewage charges, clean development mechanisms (CDM), and carbon trading and planning, to encourage technological exchanges and cooperation between low and high green innovation industries, further promote the diffusion and spillover of knowledge and technology, and realize the balance of green innovation growth across these industries. Third, China must be aware of the adverse effects of external knowledge sourcing on green innovation growth. Therefore, manufacturing should properly control the input costs of external sourcing and establish a “bridge” between R&D and external sourcing by using the government’s preferential policies to promote cooperation on green innovation, improving the industrial green innovation system to absorb external knowledge of green technology, and optimizing green production and technical conditions, which will in turn expand R&D for manufacturing to effectively improve external green innovation. Finally, considering the threshold effect of R&D level, when choosing an innovation path for green innovation growth, developing countries need to select their optimal innovation paths depending on the distribution of the level of R&D across different industries. The government should further establish and improve the efficient implementation platform and the monitoring mechanism for the rational screening, introduction, transformation and accumulation. When the country reached a developed basis of R&D, the government can gradually turn to external knowledge sourcing to effectively drive the development of green innovation through the rational allocation of diversified innovation path.
There are several limitations of this study relating to: (i) the limited number of observations; (ii) the effective decision-making units (DMU) of green innovation growth cannot be further differentiated; (iii) presence of other unknown threshold factors in the relationship between external knowledge sourcing and green innovation growth; and (iv) there may also be more driving paths in the regression model as well as potential interactions between them. The limited number of observations is offset by the fact that multiple adjustments in statistical caliber of manufacturing enterprises by National Bureau of Statistics of China, especially in the classification of manufacturing industry, there are differences in the statistical aspects of science and technology and environmental energy. The data of external knowledge activities were published from 2008; hence, our data starts from 2008. Second, regarding the DMU of green innovation growth, such as industries in high green innovation group, their green innovation evaluation value reached the highest value 1.0000. However, it is important to note that this does not mean that no further improvement necessary for these industries, rather it is due to the limitations of our model being unable to distinguish differences among these effective DMU of green innovation growth. The last two limitations indicating that the influence of the external knowledge sourcing on green innovation growth may have other threshold effects beyond R&D level in the complex green innovation system. There is a threshold created by a critical point in open innovation system [26,27], and this includes threshold factors such as GDP level, education level, pollution events, command-and-control regulation and market-based regulation, which may have significant influence on green innovation [48]. Moreover, in addition to external knowledge sourcing, the diversified innovation paths also include internal knowledge sourcing, non-R&D and others, and possible interactions between these driving paths may have been overlooked. For instance, internal and external R&D may be complementary to varying degrees but may also substitute for each other [29]. Green innovation may be affected by interactions between these diversified driving paths. However, in the regression model, only uncorrelated explanatory variables were included in order to avoid biased regression results due to multi-collinearity [61]. Adding more driving paths in the model is not feasible because of the low number of observations.
For future work, there is a need for an improved evaluation system to subdivide and compare the structure of effective decision-making units, and further identify and measure the difference in green innovation growth among high group industries. Meanwhile, it would be necessary to empirically study the underlying threshold factors and driving paths. Future research may analyze interactions between diversified driving paths on green innovation with various threshold constraints.

Supplementary Materials

The following are available online at www.mdpi.com/2071-1050/9/3/342/s1, Table S1: the comprehensive index of environmental pollution, Table S2: Distribution of best practice manufacturing in Green innovation growth (2008–2014), Table S3: Distribution of manufacturing in different threshold intervals each year.

Acknowledgments

The authors thank the editors and reviewers. We also gratefully acknowledge the help of Thompson S.H. TEO from National University of Singapore. This work is support by grants from the National Natural Science Foundation of China (71273072), the National Soft Science Foundation of China (2013GXS5B190).

Author Contributions

Jian Hou and Heng Chen conceived and designed the experiments; Jian Hou performed the experiments; Jian Hou and Heng Chen analyzed the data and contributed reagents/materials/analysis tools; and Jian Hou and Jianzhong Xu wrote the paper.

Conflicts of Interest

The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Appendix A

Specifications of the SBM-DDF Model

We regard each industry in manufacturing as a production decision unit and construct the best practice frontier in each period. Each industry has n inputs: x = ( x 1 , , x n ) , and m outputs: y = ( y 1 , , y m ) ; then, the input–output value of industry k = 1 , , K is ( x k , t , y k , t ) in the period t = 1 , , T . When the production possibility set meets the condition that the closed set, bounded set, output and input can be freely disposed, the production technology model can be expressed as follows using data envelopment analysis:
P t = { ( x t , y t ) : x n t k = 1 K λ k t x n k t , n ; y m t k = 1 K λ k t x m k t , m ; k = 1 K λ k t = 1 ; λ k t 0 , k } x t = ( x 1 t , , x n t ) ; y t = ( y 1 t , , y m t )
where λ k t is the weight variable, the sum of the weight is 1, and the weight is non-negative, indicating variable returns to scale ( V R S ). If this condition is removed, it becomes constant returns to scale ( C R S ).
Then, we set the directional distance function based on SBM:
S V t ( x t , k , y t , k ; g x , g y ) = M a x s x , s y 1 N n = 1 N S n x g n x + 1 M m = 1 M S m y g m y 2 s . t . k = 1 K λ k t x k n t + s n x = x k n t , n k = 1 K λ k t y k m t + s m y = y k m t , m k = 1 K λ k t = 1 , k s n x 0 , s m y 0
where ( x t , k , y t , k ) is the input and output vector of industry k in the period t . ( g x , g y ) is the positive direction vector of the value of the output expansion and input reduction. ( s n x , s m y ) is the slack vector of inputs and outputs. s n x , s m y is positive when the actual input is greater than the frontier input and the actual output is less than the frontier potential output. When the measurement unit of the direction vector ( g x , g y ) and the slack vector ( s n x , s m y ) is consistent, we can make the slack vector standardized, sum the standard input and output slack separately, and then calculate its average value. Thus, the objective function is to maximize the sum of the average values of the inputs and outputs. When the direction vectors ( g x , g y ) are x n max x n min and y m max y m min , 0 S V t ( x t , k , y t , k ; g x , g y ) 1 . Finally, we construct the index of green innovation growth ( G I G ) as follows:
G I G = 1 S V t ( x t , k , y t , k ; g x , g y )
The G I G value is within the interval [0, 1]; and the lower S V t (inefficiency value), the higher G I G .

References

  1. Dai, J.; Chen, B.; Hayat, T.; Alsaedi, A.; Ahmad, B. Sustainability-based economic and ecological evaluation of a rural biogas-linked agro-ecosystem. Renew. Sust. Energy Rev. 2015, 41, 347–355. [Google Scholar] [CrossRef]
  2. Sun, X.; Xiong, S.; Zhu, X.J.; Zhu, X.D.; Li, Y.F.; Li, B.L. A new indices system for evaluating ecological-economic-social performances of wetland restorations and its application to Taihu lake basin, China. Ecol. Model. 2015, 295, 216–226. [Google Scholar] [CrossRef]
  3. National Bureau of Statistics of China. China Statistical Yearbook on Environment; National Bureau of Statistics of China: Beijing, China, 2015.
  4. National Bureau of Statistics of China. China Energy Statistical Yearbook; National Bureau of Statistics of China: Beijing, China, 2015.
  5. Antal, M. Green goals and full employment: Are they compatible? Ecol. Econ. 2014, 107, 276–286. [Google Scholar] [CrossRef]
  6. Organization for Economic Co-operation and Development. Towards Green Growth: A Summary for PolicyMakers; OECD: Paris, France, 2011. [Google Scholar]
  7. Kim, S.E.; Kim, H.; Chae, Y. A new approach to measuring green growth: Application to the OECD and Korea. Futures 2014, 63, 37–48. [Google Scholar] [CrossRef]
  8. Martin, J. “Green growth”: From a growing eco-industry to economic sustainability. Energy Policy 2012, 48, 13–21. [Google Scholar]
  9. Sun, L.Y.; Miao, C.L.; Li, Y. Ecological-economic efficiency evaluation of green technology innovation in strategic emerging industries based on entropy weighted topsis method. Ecol. Indic. 2017, 73, 554–558. [Google Scholar] [CrossRef]
  10. Byron, C.J.; Jin, D.; Dalton, T.M. An Integrated ecological–economic modeling framework for the sustainable management of oyster farming. Aquaculture 2015, 447, 15–22. [Google Scholar] [CrossRef]
  11. Ooba, M.; Hayashi, K.; Fujii, M.; Fujita, T.; Machimura, T.; Matsui, T. A long-term assessment of ecological-economic sustainability of woody biomass production in Japan. J. Clean. Prod. 2015, 88, 318–325. [Google Scholar] [CrossRef]
  12. Lin, B.W.; Wu, C.H. How does knowledge depth moderate the performance of internal and external knowledge sourcing strategies? Technovation 2010, 30, 582–589. [Google Scholar] [CrossRef]
  13. Minin, A.D.; Frattini, F.; Piccaluga, A. Fiat: Open Innovation in a Downturn (1993–2003). Calif. Manag. Rev. 2010, 52, 132–159. [Google Scholar] [CrossRef]
  14. Reagans, R.; McEvily, B. Network Structure and Knowledge Transfer: The Effects of Cohesion and Range. Adm. Sci. Q. 2003, 48, 240–267. [Google Scholar] [CrossRef]
  15. Hewitt-Dundas, N. Resource and capability constraints to innovation in small and large plants. Small Bus. Econ. 2006, 26, 257–277. [Google Scholar] [CrossRef]
  16. Ang, S.H. Competitive intensity and collaboration: Impact of firm growth across technological environments. Strateg. Manag. J. 2008, 29, 1057–1075. [Google Scholar] [CrossRef]
  17. Sampson, R.C. R&D alliances and firm performance: The impact of technological diversity and alliance organization on innovation. Acad. Manag. J. 2007, 50, 364–386. [Google Scholar]
  18. Wu, J. Technological collaboration in product innovation: The role of market competition and sectoral technological intensity. Res. Policy 2012, 41, 489–496. [Google Scholar] [CrossRef]
  19. Todtling, F.; Lehner, P.; Kaufmann, A. Do different types of innovation rely on specific kinds of knowledge interactions? Technovation 2009, 29, 59–71. [Google Scholar]
  20. Liu, W.W.; Xu, X.D.; Yang, A.L.; Zhao, J.Y.; Xing, J. Impacts of FDI Renewable Energy Technology Spillover on China’s Energy Industry Performance. Sustainability 2016, 8, 846. [Google Scholar] [CrossRef]
  21. Man, A.P.D.; Duysters, G. Collaboration and innovation: A review of the effects of mergers, acquisitions and alliances on innovation. Technovation 2005, 25, 1377–1387. [Google Scholar] [CrossRef]
  22. Vega-Jurado, J.; Gutiérrez-Gracia, A.; Fernández-de-Lucio, I. Does external knowledge sourcing matter for innovation? Evidence from the Spanish manufacturing industry. Ind. Corp. Chang. 2009, 18, 637–670. [Google Scholar] [CrossRef]
  23. Choi, B.; Lee, J.N.; Ham, J. Assessing the Impact of Open and Closed Knowledge Sourcing Approach on Innovation in Small and Medium Enterprises. Procedia Comput. Sci. 2016, 91, 314–323. [Google Scholar] [CrossRef]
  24. Caloghirou, Y.; Kastelli, I.; Tsakanikas, A. Internal capability and external knowledge sources: Complements or substitutes for innovative performance? Technovation 2004, 24, 29–39. [Google Scholar] [CrossRef]
  25. Nieto, M.J.; Santamara, L. The importance of diverse collaborative networks for the novelty of product innovation. Technovation 2007, 6, 367–377. [Google Scholar] [CrossRef]
  26. Berchicci, L. Towards an open R&D system: Internal R&D investment, external knowledge acquisition and innovative performance. Res. Policy 2013, 42, 117–127. [Google Scholar]
  27. Laursen, K.; Salter, A. Open for innovation: The role of openness in explaining Innovation performance among UK manufacturing firms. Strateg. Manag. J. 2006, 27, 131–150. [Google Scholar] [CrossRef]
  28. Cassiman, B.; Veugelers, R. In search of complementarity in innovation strategy: Internal R&D and external knowledge acquisition. Manag. Sci. 2006, 52, 68–82. [Google Scholar]
  29. Hagedoorn, J.; Wang, N. Is there complementarity or substitutability between internal and external R&D strategies? Res. Policy 2012, 41, 1072–1083. [Google Scholar]
  30. Blomstrom, M.; Sjoholm, F. Technology Transfer and Spillovers: Does Local Participation with Multination is Matter? Eur. Econ. Rev. 1999, 43, 915–923. [Google Scholar] [CrossRef]
  31. Glass, A.J.; Saggi, K. Intellectual property rights and foreign direct investment. J. Int. Econ. 2002, 56, 387–410. [Google Scholar] [CrossRef]
  32. Putranto, K.; Stewart, D.; Moore, G. International Technology Transfer and Distribution of Technology Capabilities: The Case of Railway Development in Indonesia. Technol. Soc. 2003, 121, 43–53. [Google Scholar] [CrossRef]
  33. Becker, W.; Dietz, J. R&D cooperation and innovation activities of firms--evidence for the German manufacturing industry. Res. Policy 2004, 33, 209–223. [Google Scholar]
  34. Gosens, J.; Lu, Y.L. From lagging to leading? Technological innovation systems in emerging economies and the case of Chinese wind power. Energy Policy 2013, 60, 234–250. [Google Scholar] [CrossRef]
  35. Miao, C.L.; Sun, L.Y.; Yang, L. The studies of ecological environmental quality assessment in anhui province based on ecological footprint. Ecol. Indic. 2016, 60, 879–883. [Google Scholar] [CrossRef]
  36. Tseng, M.L.; Wang, R.; Chiu, A.S.F.; Geng, Y.; Lin, Y.H. Improving performance of green innovation practices under uncertainty. J. Clean. Prod. 2013, 40, 71–82. [Google Scholar] [CrossRef]
  37. Chung, Y.H.; Fare, R.; Grosskopf, S. Productivity and undesirable outputs: A directional distance function approach. J. Environ. Manag. 1997, 51, 229–240. [Google Scholar] [CrossRef]
  38. Fukuyama, H.; Weber, W.L. A directional slacks-based measure of technical inefficiency. Socio-Econ. Plan. Sci. 2009, 43, 274–287. [Google Scholar] [CrossRef]
  39. Tone, K. Dealing with Undesirable Outputs in DEA: A Slacks Based Measure (SBM) Approach. GRIPS Research Report Series. 2003. Available online: http://ci.nii.ac.jp/els/110003481580.pdf?id=ART0004103187&type=pdf&lang=en&host=cinii&order_no=&ppv_type=0&lang_sw=&no=1487749452&cp= (accessed on 22 February 2017).
  40. Zhang, S.H.; Jiang, W.J. Energy Efficiency Measures: Comparative Analysis. J. Quant. Tech. Econ. 2016, 7, 3–24. [Google Scholar]
  41. Guan, J.C.; Chen, K.H. Measuring the innovation production process: A cross-region empirical study of China’s high-tech innovations. Technovation 2010, 30, 348–358. [Google Scholar] [CrossRef]
  42. Sharma, S.; Thomas, V.J. Inter-country R&D efficiency analysis: An application of data envelopment analysis. Scientometrics 2008, 76, 483–501. [Google Scholar]
  43. Liu, X.; Buck, T. Innovation performance and channels for international technology spillovers: Evidence from Chinese high-tech industries. Res. Policy 2007, 36, 355–366. [Google Scholar] [CrossRef]
  44. Long, X.L.; Zhao, X.C.; Cheng, F.X. The comparison analysis of total productivity and eco-efficiency in China’s cement manufactures. Energy Policy 2015, 81, 61–66. [Google Scholar] [CrossRef]
  45. Ren, S.G.; Li, X.L.; Yuan, B.L.; Li, D.Y.; Chen, X.H. The effects of three types of environmental regulation on eco-efficiency: A cross-region analysis in China. J. Clean. Prod. 2016. [Google Scholar] [CrossRef]
  46. Hansen, B.E. Threshold Effects in Non-Dynamic Panels: Estimation, Testing, and Inference. J. Econ. 1999, 93, 345–368. [Google Scholar] [CrossRef]
  47. Arellano, M.; Bond, S.R. Some Tests of Specification for panel Data: Monte Carlo Evidence and an Application to Employment Equations. Rev. Econ. Stud. 1991, 58, 277–297. [Google Scholar] [CrossRef]
  48. Xie, R.H.; Yuan, Y.J.; Huang, J.J. Different types of environmental regulations and heterogeneous influence on “green” productivity: Evidence from china. Ecol. Econ. 2017, 132, 104–112. [Google Scholar] [CrossRef]
  49. Carayannopoulos, S.; Auster, E.R. External knowledge sourcing in biotechnology through acquisition versus alliance: A KBV approach. Res. Policy 2010, 39, 254–267. [Google Scholar] [CrossRef]
  50. Marsh, P. The New Industrial Revolution: Consumers, Globalization and the End of Mass Production; Yale University Press: New Haven, CT, USA, 2012. [Google Scholar]
  51. Porter, M.E.; Van, D.L.C. Toward a new conception of the environment-competitiveness relationship. J. Econ. Perspect. 1995, 9, 97–118. [Google Scholar] [CrossRef]
  52. Ramanathan, R.; He, Q.L.; Black, A.; Ghobadian, A.; Gallear, D. Environmental regulations, innovation and firm performance: A revisit of the Porter hypothesis. J. Clean. Prod. 2016. [Google Scholar] [CrossRef]
  53. Zhao, X.; Sun, B.W. The influence of Chinese environmental regulation on corporation innovation and competitiveness. J. Clean. Prod. 2016, 112, 1528–1536. [Google Scholar] [CrossRef]
  54. Lau, L.; Choong, C.; Eng, Y. Investigation of the Environmental Kuznets Curve for Carbon Emissions in Malaysia: Do Foreign Direct Investment and Trade Matter? Energy Policy 2014, 68, 490–497. [Google Scholar] [CrossRef]
  55. Tang, C.F.; Tan, B.W. The Impact of Energy Consumption, Income and Foreign Direct Investment on Carbon Dioxide Emissions in Vietnam. Energy 2015, 79, 447–454. [Google Scholar] [CrossRef]
  56. Horta, I.M.; Camanho, A.S.; Costa, J.M.D. Performance assessment of construction companies: A study of factors promoting financial soundness and innovation in the industry. Int. J. Prod. Econ. 2012, 137, 84–93. [Google Scholar] [CrossRef]
  57. Keupp, M.; Gassmann, O. Determinants and archetype users of open innovation. R&D Manag. 2009, 39, 331–341. [Google Scholar]
  58. Cassidy, M.; Görg, H.; Strobl, E. Knowledge accumulation and productivity Evidence from plant level data from Ireland. Scott. J. Political Econ. 2005, 52, 344–358. [Google Scholar]
  59. Thornhill, S. Knowledge, innovation and firm performance in high- and low-technology regime. J. Bus. Ventur. 2006, 21, 687–703. [Google Scholar] [CrossRef]
  60. Zahra, S.A.; Nielsen, A.P. Sources of capabilities, integration and technology commercialization. Strateg. Manag. J. 2002, 23, 377–398. [Google Scholar] [CrossRef]
  61. Phan, T.H.D.; Brouwer, R.; Davidson, M.D. A global survey and review of the determinants of transaction costs of forestry carbon projects. Ecol. Econ. 2017, 133, 1–10. [Google Scholar] [CrossRef]
Figure 1. Trends of green innovation growth in manufacturing (2008–2014).
Figure 1. Trends of green innovation growth in manufacturing (2008–2014).
Sustainability 09 00342 g001
Figure 2. The construction of confidence intervals: (a) the first threshold value; and (b) the second threshold value.
Figure 2. The construction of confidence intervals: (a) the first threshold value; and (b) the second threshold value.
Sustainability 09 00342 g002
Table 1. The indicators system of green innovation growth.
Table 1. The indicators system of green innovation growth.
TypeIndicatorDefinition
InputR&D input
Human input
Energy input
R&D capital stock
Full-time R&D personnel
Energy consumption (million tons of standard coal)
Desirable outputsDirect output
Indirect output
The number of patent applications
New product sales revenue
Undesirable outputsThe comprehensive index of environmental pollutionThe amount of discharged pollutants (industrial wastewater, industrial waste gas, industrial solid waste), as well as SO2, soot and dust emissions
Table 2. Descriptive statistics of variables.
Table 2. Descriptive statistics of variables.
VariableMeanP50S.D.MinimumMaximum
Green Innovation Growth (GIG)0.6530.6190.2310.2941.000
External Knowledge Sourcing (EKS)12.78512.9151.33410.40815.228
R&D Level (LEV)0.0050.0030.0030.0010.012
Environmental Regulation (REG)10.59310.6981.7426.95713.825
Foreign Direct Investment (FDI)0.3020.2860.1300.0010.663
Industrial Agglomeration (AGG)9.0669.0701.1164.86810.519
Industry Scale (SCA)231437870598402147372333
Industrial Factor Endowment (END)50.11036.13543.9377.824242.566
Table 3. Index of green innovation growth in manufacturing (2008–2014).
Table 3. Index of green innovation growth in manufacturing (2008–2014).
Manufacturing2008200920102011201220132014Mean
High green innovation group1.00001.00001.00000.97760.98680.93120.94920.9775
Tobacco1.00001.00001.00001.00001.00001.00001.00001.0000
Artworks, and Articles for Culture, Education, Sports and Recreation1.00001.00001.00001.00001.00001.00001.00001.0000
Electrical Machinery and Equipment1.00001.00001.00001.00001.00001.00001.00001.0000
Computer, Communication and Other Electronic Equipment1.00001.00001.00001.00001.00001.00001.00001.0000
Measuring Instrument and Meters1.00001.00001.00001.00001.00001.00001.00001.0000
Transport Equipment1.00001.00001.00000.85320.91140.83560.90370.9268
Furniture1.00001.00001.00001.00001.00000.72660.76840.9201
Middle green innovation group0.69190.68850.63360.57980.59040.57850.58850.6199
Leather, Fur, Feathers and Related Products and Shoes1.00001.00000.70080.63810.64680.62900.62010.7322
Textile, Apparel and Accessories0.72400.75110.73810.64420.76750.73440.74020.7275
Printing, Reproduction of Recording Media0.73130.79500.67810.69130.74590.71760.70030.7219
Special Purpose Machinery0.73550.67900.64860.68830.74810.76710.73650.7136
General Purpose Machinery0.68000.69140.62720.71100.74530.75580.72940.7045
Processing of Timbers and Wood, Bamboo, Rattan, Palm and Straw0.68160.66300.66080.53040.56320.48010.50540.5783
Metal Products0.65290.61800.58230.50800.55670.51400.54350.5658
Rubber and Plastic0.63030.64110.55830.49230.52410.56190.55970.5647
Textile0.61990.57280.58940.50500.52680.47120.50160.5387
Medicines0.58270.56480.52690.46570.49080.51690.52120.5228
Other Manufacturing0.64860.69190.69380.57410.34640.37450.42280.5165
Low green innovation group0.61850.46150.42140.37060.42590.40610.40140.4381
Processing of Food from Agricultural Products0.62280.60700.46540.44520.46960.40750.41740.4844
Chemical Fiber0.58050.45980.46320.42320.47880.48230.44960.4747
Foods0.68600.60560.51720.39770.41760.38770.38290.4730
Ferrous Metals1.00000.43980.41930.34970.41480.43970.37830.4614
Non-ferrous Metals0.55960.44550.43400.39100.43450.41140.42370.4402
Paper and Paper Products0.54400.42270.46600.38460.46430.39230.41630.4387
Non-metallic Mineral Products0.48280.51210.47140.38200.38510.40110.40670.4318
Chemical Raw Materials and Chemical Products0.47630.45050.33000.37250.42240.38970.39830.4031
Processing of Petroleum, Coking and Processing of Nuclear Fuel1.00000.34800.29430.29230.36920.35860.37800.3949
Liquor, Beverages and Refined Tea0.47190.39030.40750.29820.41630.40290.37010.3907
All manufacturing0.72890.65520.61390.56310.59740.57430.57850.6137
Table 4. Results of threshold-effect tests.
Table 4. Results of threshold-effect tests.
ModelF-Statisticp-ValueCritical Value
1%5%10%
Single threshold69.023 **0.02083.96140.3687.323
Double threshold22.655 ***0.00216.9885.4842.372
Triple threshold0.0030.95426.64515.4059.609
Note: The p-value and the critical value are obtained from 500 bootstrap replications. * p < 0.05; ** p < 0.01; *** p < 0.001.
Table 5. Results of threshold estimators and confidence intervals.
Table 5. Results of threshold estimators and confidence intervals.
ModelThreshold Estimators95% Confidence Intervals
Single threshold0.001[0.001, 0.001]
Double threshold0.007[0.007, 0.007]
0.001[0.001, 0.001]
Triple threshold0.010[0.008, 0.011]
Table 6. Results of dynamic threshold regression.
Table 6. Results of dynamic threshold regression.
Coef.Std. Err.tP > |t|95% Conf. Interval
L1.−0.548 ***0.083−6.6000.000−0.711−0.385
L2.−0.058 ***0.020−2.9400.003−0.096−0.019
REG−0.0600.142−0.4200.674−0.3370.218
FDI0.761 ***0.1057.2400.0000.5550.967
AGG0.161 ***0.01411.4600.0000.1330.188
SCA0.000 ***0.0004.1900.0000.0000.000
END−0.002 ***0.001−2.9100.004−0.003−0.001
EKS(LEV ≤ 0.001)−0.752 ***0.246−3.0600.002−1.235−0.270
EKS(0.001 < LEV ≤ 0.007)−0.072 **0.033−2.2000.028−0.137−0.008
EKS(LEV > 0.007)−0.069 **0.033−2.1200.034−0.133−0.005
_cons1.4661.5170.9700.334−1.5084.440
Note: GMM-type: L(2/.).GIG; * p < 0.05; ** p < 0.01; *** p < 0.001.
Table 7. AR(1) and AR(2) test.
Table 7. AR(1) and AR(2) test.
OrderzProb > z
AR(1)2.45120.0142
AR(2)−0.26480.7912

Share and Cite

MDPI and ACS Style

Hou, J.; Chen, H.; Xu, J. External Knowledge Sourcing and Green Innovation Growth with Environmental and Energy Regulations: Evidence from Manufacturing in China. Sustainability 2017, 9, 342. https://doi.org/10.3390/su9030342

AMA Style

Hou J, Chen H, Xu J. External Knowledge Sourcing and Green Innovation Growth with Environmental and Energy Regulations: Evidence from Manufacturing in China. Sustainability. 2017; 9(3):342. https://doi.org/10.3390/su9030342

Chicago/Turabian Style

Hou, Jian, Heng Chen, and Jianzhong Xu. 2017. "External Knowledge Sourcing and Green Innovation Growth with Environmental and Energy Regulations: Evidence from Manufacturing in China" Sustainability 9, no. 3: 342. https://doi.org/10.3390/su9030342

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop