Interaction mechanism between sustainable innovation capability and capital stock: Based on PVAR model

Abstract

The promotion of sustainable innovation capability of the manufacturing industry is a strategic move to the mid-to-high end of the global value chain. We construct the evaluation index system of sustainable innovation capability based on the panel data of China’s manufacturing industry from 2000 to 2015. We analyze the interaction mechanism between sustainable innovation capability and capital stock by using the PVAR model. The results show that there is a long-term co-integration relationship between them, and both of them have the characteristic of self-inertia. The sustainable innovation capability has a positive impact on capital stock, while the capital stock has a negative effect on sustainable innovation capability. The capital stock has a positive impact on sustainable innovation capacity in the low-tech and medium-tech industries, while it has a negative impact in the high-tech industry. Additionally, the capital stock has the least impact on sustainable innovation capacity in the low-tech industry, and sustainable innovation capacity has the greatest impact on capital stock in the medium-tech industry.

Keywords

China’s manufacturing industry sustainable innovation capability capital stock PVAR model interaction mechanism

1 Introduction

The manufacturing industry is the pillar industry of national economic development. After the subprime crisis, the U.S government proposed the strategy of Manufacturing Back-flow. Germany, the UK and Japan successively launched the “Industry 4.0”, “Industry 2050” and “Revitalization Strategy”, with manufacturing as the core of economic recovery. Against the background of international industrial transformation, the Chinese government launched the action program of “Made in China 2025” in 2015, aiming to realize the strategic goal of transforming from a big manufacturer to a strong manufacturer. China’s manufacturing industry has made remarkable achievements and great improvement in innovation capacity. By the end of 2018, the added value of the manufacturing industry had exceeded $4 trillion, accounting for 28% of the global share. Its contribution to China’s economic growth reached 58.5%, becoming an important engine of global industrial growth (Data from the National Bureau of Statistics of China). While the added value of the U.S. manufacturing industry is only 58% of that in China. The China-U.S. trade friction is the dispute between manufacturing industries, and it is the powerful measure taken by the U.S. to curb the technological progress of China’s manufacturing industry. Under the background of increasingly fierce trade friction, how to improve the innovation capability of the manufacturing industry to achieve sustainable development is the core and commanding point of international competition. It’s also the focus of social concern.

Innovation is the basis for the manufacturing industry to participate in market competition, which is the key to maintain a competitive position. Sustainable innovation is the organic combination of economy, society, and ecosystem with economic growth, technological development and environmental protection, which drives economic, environmental and social improvement at a national level (Adriana, et al.) [1]. Among them, the economic growth is the basic goal of industrial development and the material guarantee of technological innovation. Technology development is the core of sustainable development, and the sustainable technologies are very important for the development of the organizations or industries [2]. The social and environmental responsibilities make it have social values beyond economic benefits [3]. With the increasing strictness of environmental regulations [4], the manufacturing industry needs to consider environmental issues to achieve sustainable development.

The research on sustainable innovation capacity mainly includes the following two representative viewpoints. First, it means continuous incremental innovation of innovation subjects, which is a process of extensive and concentrated participation [5]. Second, it means the process of improving existing behaviors in a planned, organized and systematic manner with the goal of continuous improvement of business performance [6]. Sustainable innovation capability is an important channel to gain competitive advantages and economic benefits, which can be seen as a strategic measure for the transformation and upgrading of the manufacturing industry. It affects the division of labor in the global value chain and is affected by human, financial and material resources. The most critical influencing factor is capital, which can meet the resource demand for innovation activities, thus leading to higher innovation output [7 –9].

However, the “Solow paradox” exists in China’s manufacturing industry, which is manifested by a stable increase in capital investment, but the low quality of innovation output and technological progress. The core technologies of the manufacturing industry depend heavily on other countries. The interaction mechanism between these two variables needs to be further discussed which is of great significance to accelerating the innovation-driven development of China’s manufacturing industry and improving its global competitiveness.

The existing research mainly focuses on the impact of social capital, intellectual capital and FDI on the innovation capacity of the manufacturing industry [10 –13], lacking the interaction mechanism analysis between capital stock and innovation capacity from a macro perspective. Given the data availability, the existing evaluations of innovation capacity are mainly based on patent applications and economic benefits [14, 15], with less consideration of social and ecological benefits brought by innovation capability. The evaluation and dynamic analysis of the sustainable innovation capability of the manufacturing industry are far less. Additionally, the widely quoted conclusions mainly come from empirical analysis on the entire manufacturing industry, rather than the sub-sectors, which exits differences in economic development, capital scale, and technology level. It may lead to deviations in the estimates of sustainable innovation capacity and capital stock and most studies ignore the endogenous problem and cannot fully explain the interaction mechanism between sustainable innovation ability and capital stock and the industry heterogeneity.

Based on the panel data of China’s manufacturing industry from 2000 to 2015, we construct the evaluation index system of sustainable innovation ability by considering the economic, social and ecological benefits. We make a systematical analysis of the interaction between sustainable innovation capability and capital stock from dimensions of time series and industry heterogeneity by adopting the panel vector autoregressive model (PVAR).

2 Logical relationship between sustainable innovation ability and capital stock in China’s manufacturing industry

Innovation is the fundamental driving force for economic growth and social development [16 –18], and it is also the key determinant of sustainable development of the manufacturing industry. On the one hand, innovation is the recombination of production factors [16], which can promote the optimal allocation of labor, capital, raw materials, and other factors. It can reduce the production cost of unit factors, bringing the low-cost advantage for the manufacturing industry to maximize excess profits [19, 20]. On the other hand, innovation means that manufacturing enterprises with higher or new technology become the precursors in market competition, which is an important prerequisite for obtaining competitive advantages [21, 22]. This effectively improves the industrial profit margin and economic benefits, enabling it to have a sufficient financial base for further technological innovation and accelerate innovation speed. The improvement of sustainable innovation capability brings low-cost advantage and competitive advantage to the manufacturing industry. The resulting high profit is an important source of capital for innovation investment, which stimulates the driving force for technological innovation. Therefore, the manufacturing industry has stronger incentive to strengthen its competitive advantage by pursuing efficiency and forms the phenomenon of relying on its inertial development.

The sustainable innovation capability plays a positive role in capital stock, and the impact path lies in the following. First, capital has the profit-seeking nature [23], and always flows to industries or fields with a higher return on investment. The high added value and industry profitability brought by the improvement of the sustainable innovation capability of the manufacturing industry will inevitably have a strong attraction to capital, accelerating the capital accumulation to further increase the amount of capital stock. This process meets the judging criteria of the capital allocation efficiency proposed by Wurgler [24]. Second, technological innovation is a fundamental way to promote the manufacturing industry to get rid of resource dependence and transform into the capital-intensive and technology-intensive industry. It changes the input structure and allocation efficiency of industrial factors through the quantitative effect of capital accumulation and structural adjustment, which affects the capital deepening [25]. This can be reflected in the capital stock.

Due to the high degree of uncertainty, information asymmetry and positive externalities of innovation activities [26], sustainable innovation activities have adverse selection and moral hazard and are susceptible affected by financing constraints [27, 28]. As capital is the material basis for the smooth implementation of innovation activities [29], the rational input and distribution of capital elements is an important prerequisite for manufacturing technology development. When the manufacturing capital stock is at a relatively high level, its support for innovation activities is strong, which helps to improve the technology innovation efficiency. This process accelerates the speed of sustainable innovation and transfers capital value to innovative products and services. So the capital stock promotes the sustainable innovation capability of the manufacturing industry.

Meanwhile, industries with higher capital stocks can use the capital markets to prevent competitors from acquiring external funds, thereby gaining a higher competitive advantage and attracting more capital. When capital continuously gathers to the highly innovative manufacturing industries, given its profit-seeking nature, such industries will inevitably be required to reintegrate or allocate resources to improve the efficiency of resource allocation. As the innovation capability is one of the strategic resources [30], the effectiveness of innovation decisions of manufacturing enterprises will also improve. According to the innovation theory put forward by Schumpeter [16], the recombination of resources and the resulting services are the basis for new products, technologies, and processes. Therefore, while the capital stock is affected by sustainable innovation capability, it also has a positive effect on sustainable innovation capability. The logical relationship between them is shown in Fig. 1.

Fig. 1

Logical relationship between sustainable innovation capability and capital stock in the manufacturing industry.

3 Measurement and evaluation of sustainable innovation capability and capital stock of China’s manufacturing industry

3.1 Measurement of sustainable innovation capacity of China’s manufacturing industry

The core variable of our paper is the sustainable innovation capability of the manufacturing industry. Since the single indicator cannot reflect the true situation of the sustainable innovation capability accurately, our paper measures it from aspects of the innovation input capacity, innovation implementation capability, innovation output capability, and sustainable development capability by following the principles of sustainability, scientificity, comprehensiveness, and availability. Combining with the definition of sustainable innovation capability, the economic benefits, social benefits and ecological benefits are included in the evaluation scope. The evaluation index system of sustainable innovation capability of the manufacturing industry is shown in Table 1.

Table 1
Evaluation index system of sustainable innovation capability of manufacturing industry

First class index Second class index Third class index Definition

Innovation input capacity (Innov1) Capital investment Intensity of R&D expenditure R&D expenditure/sales revenue

Growth rate of R&D expenditure R&D expenditure in the current period/R&D expenditure in the last period-1

Personnel investment Intensity of R&D personnel R&D personnel/number of employees

Growth rate of R&D personnel R&D personnel in the current period/R&D personnel in the last period-1

Innovation implementation capability (Innov2) R&D capability R&D institutions Number of R&D institutions

R&D projects Number of R&D projects

Number of new products Number of new products

Production &marketing capability Fixed assets equipment rate Fixed assets/number of employees

Expenses on new products development Expenses on new products development in the current period

Intensity of marketing expense Marketing expense/sales revenue

Expenditure for technology acquisition Expenditure for technology acquisition in the current period

Innovative output capacity (Innov3) Patent output Patent applications Piece of patent applications

Invention in force Piece of the invention in force

Economic benefits Total profits Total profits in the current period

Operating profit ratio Operating profit/sales revenue

Sales revenue of new product Sales revenue of new product in the current period

Social benefits Number of employees Number of employees in the current period

Employee salary and benefits Employee compensation/number of employees

Labor productivity Industrial added value/number of employees

Ecological benefits Pollutant emissions Pollutant emissions per unit of production value

Energy consumption Energy consumption per unit of production value

Volume of solid wastes utilized Volume of solid wastes utilized per unit of production value

Sustainable development capability (Innov4) Development potential Growth rate of sales revenue Sales revenue in the current period/Sales revenue in the last period-1

Growth rate of total assets total assets in the current period/total assets in the last period-1

Growth rate of operating profit Profit in the current period/profit in the last period-1

First class index	Second class index	Third class index	Definition
Innovation input capacity (Innov1)	Capital investment	Intensity of R&D expenditure	R&D expenditure/sales revenue
		Growth rate of R&D expenditure	R&D expenditure in the current period/R&D expenditure in the last period-1
	Personnel investment	Intensity of R&D personnel	R&D personnel/number of employees
		Growth rate of R&D personnel	R&D personnel in the current period/R&D personnel in the last period-1
Innovation implementation capability (Innov2)	R&D capability	R&D institutions	Number of R&D institutions
		R&D projects	Number of R&D projects
		Number of new products	Number of new products
	Production &marketing capability	Fixed assets equipment rate	Fixed assets/number of employees
		Expenses on new products development	Expenses on new products development in the current period
		Intensity of marketing expense	Marketing expense/sales revenue
		Expenditure for technology acquisition	Expenditure for technology acquisition in the current period
Innovative output capacity (Innov3)	Patent output	Patent applications	Piece of patent applications
		Invention in force	Piece of the invention in force
	Economic benefits	Total profits	Total profits in the current period
		Operating profit ratio	Operating profit/sales revenue
		Sales revenue of new product	Sales revenue of new product in the current period
	Social benefits	Number of employees	Number of employees in the current period
		Employee salary and benefits	Employee compensation/number of employees
		Labor productivity	Industrial added value/number of employees
	Ecological benefits	Pollutant emissions	Pollutant emissions per unit of production value
		Energy consumption	Energy consumption per unit of production value
		Volume of solid wastes utilized	Volume of solid wastes utilized per unit of production value
Sustainable development capability (Innov4)	Development potential	Growth rate of sales revenue	Sales revenue in the current period/Sales revenue in the last period-1
		Growth rate of total assets	total assets in the current period/total assets in the last period-1
		Growth rate of operating profit	Profit in the current period/profit in the last period-1

Since the entropy method can determine the weight of each indicator objectivity, it’s used to calculate the weight of indicators of sustainable innovation capability and then obtain the comprehensive index.

First, to avoid the dimensionality and multicollinearity of different indicators, the original data is normalized. Assuming that there are m evaluation indicators, n samples, and the original matrix is A = (a_ij) _m×n, then R = (r_ij) _m×n is obtained after normalization. The formula is as follows. $r_{ij} = \frac{a_{ij} - m in (a_{ij})}{m ax (a_{ij}) - m in (a_{ij})}$ (1)

Second, the entropy value is obtained by the following formula. $h_{i} = - k \sum_{j = 1}^{n} f_{ij} ln f_{ij}$ (2)

Where $f_{ij} = r_{ij} / \sum_{j = 1}^{n} r_{ij}$ and k = 1/ln n. When f_ij is 0, f_ijlnf_ij is 0. Then the entropy weight is obtained by the following formula. $w_{i} = \frac{1 - h_{i}}{m - \sum_{i = 1}^{m} h_{i}} (0 ⩽ w_{i} ⩽ 1, \sum_{i = 1}^{m} w_{i} = 1)$ (3)

The entropy weights of innovation input capacity, innovation implementation capability, innovative output capacity, and sustainable development capability are 6.28%, 29.38%, 60.81, and 3.53%, respectively. The sustainable innovation capability is obtained by the following formula.

$\begin{matrix} Innov = 6.28 % Innov 1 + 29.38 % Innov 2 \\ + 60.81 % Innov 3 + 3.53 % Innov 4 \end{matrix}$ (4)

3.2 Measurement of capital stock of China’s manufacturing industry

We measure the capital stock of China’s manufacturing industry by using the internationally widely adopted perpetual inventory method (Goldsmith) [31]. The calculation formula is shown below. $K_{it} = (1 - σ_{it}) * K_{it - 1} + I_{it}$ (5)

Where Kit means the amount of capital stock of the i industry in t period, σ means the capital depreciation rate, I is the investment of fixed asset, which is converted based on the unchanged price in 2000. Referring to Kohli [32], the following formula is constructed to estimate the capital stock in the initial year. $K_{i 2000} = I_{i 2000} / (σ_{i 2000} + r_{i})$ (6)

Where r is the real growth rate of fixed asset investment at a fixed price in China’s manufacturing industry. As the different capital structure of each year, the capital depreciation rate is also varied. To improve the authenticity and accuracy of the estimated results, we measure the capital depreciation rates of different industries. The formula is shown below. $σ_{it} = \frac{depreciation in the current yea r_{it}}{original value of fixed {assets}_{it - 1}}$ (7)

3.3 Samples and data sources

The manufacturing industry is divided into 31 industry segments according to the industry classification standard of China’s Securities Regulatory Commission. To ensure the consistency of the statistical calibre, the automobile manufacturing industry and the railway, shipbuilding, aerospace, and other transportation equipment manufacturing industry are unified into the transportation equipment manufacturing industry; the rubber products industry and the plastic products industry are unified into the rubber and plastic products industry; the waste resource comprehensive utilization industry and the metal products, machinery and equipment repair industry are eliminated. Finally, 28 manufacturing industry segments are obtained.

Due to a large number of manufacturing industry segments, their technological level and industry focus are different, which may reduce the practical significance of the research conclusion. Concerning the OECD (2007) technical classification criteria for the manufacturing industry, China’s manufacturing industry is divided into three major categories, namely, the low-tech industry, the medium-tech industry, and the high-tech industry. The industry classifications and their codes are shown in Table 2.

Table 2
Manufacturing industry segments and their codes

Name Code

Agricultural and sideline food processing industry 01

Food manufacturing Industry 02

Beverage manufacturing Industry 03

Tobacco industry 04

Textile industry 05

Textile clothing, shoes, and hats manufacturing industry 06

Leather, fur, feather, and related products 07

Wood processing and wood, bamboo, rattan and palm grass products industry 08

Furniture manufacturing industry 09

Paper-making and paper products industry 10

Printing and reproduction of recording media industry 11

Handcrafts and other manufacturing industry 12

Cultural, educational and sporting goods manufacturing industry 13

Petroleum processing, coking, and nuclear fuel processing industry 14

Chemical raw materials and products manufacturing industry 15

Chemical fiber manufacturing industry 16

Rubber and plastic products industry 17

Non-metallic mineral products industry 18

Ferrous metal smelting and rolling processing industry 19

Non-ferrous metal smelting and rolling processing industry 20

Metal products industry 21

General equipment manufacturing industry 22

Special equipment manufacturing industry 23

Transportation equipment manufacturing industry 24

Electrical machinery and equipment manufacturing industry 25

Communication equipment, computer, and other electronic equipment manufacturing 26

Instrumentation and culture, office machinery manufacturing industry 27

Pharmaceutical manufacturing 28

Name	Code
Agricultural and sideline food processing industry	01
Food manufacturing Industry	02
Beverage manufacturing Industry	03
Tobacco industry	04
Textile industry	05
Textile clothing, shoes, and hats manufacturing industry	06
Leather, fur, feather, and related products	07
Wood processing and wood, bamboo, rattan and palm grass products industry	08
Furniture manufacturing industry	09
Paper-making and paper products industry	10
Printing and reproduction of recording media industry	11
Handcrafts and other manufacturing industry	12
Cultural, educational and sporting goods manufacturing industry	13
Petroleum processing, coking, and nuclear fuel processing industry	14
Chemical raw materials and products manufacturing industry	15
Chemical fiber manufacturing industry	16
Rubber and plastic products industry	17
Non-metallic mineral products industry	18
Ferrous metal smelting and rolling processing industry	19
Non-ferrous metal smelting and rolling processing industry	20
Metal products industry	21
General equipment manufacturing industry	22
Special equipment manufacturing industry	23
Transportation equipment manufacturing industry	24
Electrical machinery and equipment manufacturing industry	25
Communication equipment, computer, and other electronic equipment manufacturing	26
Instrumentation and culture, office machinery manufacturing industry	27
Pharmaceutical manufacturing	28

Note: Industries from Code 01 to 12 are low-tech industries; Industries from Code 13 to 24 are medium technology industries; Industries from Code 25 to 28 are high-tech industries.

The data of China’s statistical yearbooks are only available until 2015. Given the data availability, the manufacturing sample period selected in our paper is from 2000 to 2015, and the linear interpolation method is used to measure the missing data. All data are from the China National Bureau of Statistics, China Statistical Yearbook, China Energy Statistical Yearbook, China Environmental Statistical Yearbook, China Science and Technology Statistical Yearbook and China Industrial Economic Statistical Yearbook.

3.4 Evaluation of sustainable innovation capacity and capital stock of China’s manufacturing industry

According to the previous measurement, the time-series changes of the sustainable innovation capacity and capital stock from 2000 to 2015 are shown in Fig. 2.

Fig. 2

Time series of sustainable innovation capacity and capital stock of manufacturing industry from 2000 to 2015.

It can be seen that sustainable innovation capacity increases from 0.057 in 2000 to 0.122 in 2015, with a growth rate of 113.63% and the fluctuation upward trend throughout the sample period. The reason may be that the manufacturing industry is the pillar of China’s national economic development and its innovative development is highly valued by the state and government departments. How to improve its sustainable innovation capability is an important strategic measure to promote China’s sustained economic growth and move towards the mid-to-high end of the global value chain. However, China’s manufacturing industry is still vulnerable affected and restricted by external factors such as national policies, resource endowments, and economic development. There is a certain fluctuation in the time series of sustainable innovation capability.

From the results of the capital stock, it increases from 2597 in 2000 to 11,527 in 2015, increasing year by year with a growth rate of 343.88%, which is similar to the changing trend of sustainable innovation capability. The results of sustainable innovation capacity and capital stock in different manufacturing industry segments are shown in Fig. 3.

Fig. 3

Annual average sustainable innovation capacity and capital stock of different industries.

As can be seen from Fig. 3, there are obvious heterogeneous characteristics of sustainable innovation capacity and capital stock in different industries. Among them, the Communication equipment, computer, and other electronic equipment manufacturing (Code26) has the highest comprehensive index of sustainable innovation capacity, reaching 0.2139, followed by the Electrical machinery and equipment manufacturing industry (Code25) and the Special equipment manufacturing industry (Code23) with the sustainable innovation capability of 0.1831 and 0.1640, respectively. The Printing and reproduction of recording media industry (Code11) and the Leather, fur, feather, and related products (Code07) have the lowest innovation capability, which is about 0.0240. It reflects the uneven development of sustainable innovation capability in different industries, which is consistent with the current status of China’s manufacturing industry. The reason may be that it’s greatly different in resource endowments, national policies and other innovation foundations in different industries, and the advanced manufacturing industries are more innovative.

The results of the capital stock show that the Ferrous metal smelting and rolling processing industry (Code19) is at the highest level, reaching to 19231.03, followed by the Non-ferrous metal smelting and rolling processing industry (Code20), the Pharmaceutical manufacturing (Code28) and the Chemical raw materials and products manufacturing industry (Code 15), with the value of 15784.73, 15316.99, and 10995.15, respectively. Further analysis finds that industries with high capital stock generally have one or more of the following characteristics: high-profit margins, knowledge or technology intensiveness, national policy support, and resource-based or national monopoly industry, which leads to the continuous accumulation of capital to the above industries and is reflected in the amount of capital stock. Conversely, the capital stock in the Furniture manufacturing industry (Code09) and the Cultural, educational and sporting goods manufacturing industry (Code13) is far below the average level the whole industry, which is 1252.12 and 1025.48, respectively. It indicates that the distribution of capital stock in various manufacturing industry segments is not even.

4 PVAR model analysis of sustainable innovation capacity and capital stock of China’s manufacturing industry

4.1 Model construction

To analyse the relationship between the sustainable innovation capability and capital stock of China’s manufacturing industry, theoretically, the empirical analysis can be performed on the data of individual industries to control possible industry differences. However, this method has the following defects: it ignores the similarity of different manufacturing industry segments, and the results are not reliable because of the small size sample. Given this, our paper makes research by selecting the panel data of 28 manufacturing industries and performs group regression based on the three industry categories, that is, the low-tech, medium-tech, and high-tech industries. It’s able to expand the sample size to obtain more robust results, and effectively utilize similar information existing in different industries.

From the previous studies, they mainly analyse the impact of social capital, human capital, and intellectual capital on the innovation capability of the manufacturing industry, but rarely involves sustainable innovation capability. The in-depth analysis and empirical research on the relationship between these two variables are far less. So our paper adopts the PVAR model to systematically analyse the relationship between sustainable innovation capability and capital stock of the manufacturing industry from 2000 to 2015. It also considers both time effect and fixed effect, which improves the reliability and stability of the results. The model constructed is shown in formula 1. $Y_{it} = α_{0} + \sum_{j = 1}^{k} α_{i} Δ Y_{it - j} + γ_{i} + μ_{t} + ɛ_{it}$ (8)

Where Y = (Innov, Capital) meaning the two-dimensional system vector composed of sustainable innovation capacity and capital stock. To eliminate the possible heteroscedasticity problem, the capital stock is logged in the empirical tests. i means the industry, t means the year, k means the lag order, α₀ means the intercept term vector, α_i is the coefficient vector of the error correction term, γ_i is the individual effect, μ_t is the time effect, and ɛ_it is the random interference term.

4.2 Unit root and co-integration test

Since macroeconomic indicators generally have time trends and non-stationary characteristics, the unit root test is required before regression analysis to further determine whether the co-integration test is needed. To ensure the accuracy and robustness of the test results, the commonly used LL [33] and IPS test methods [34] are used to make unit root test of the manufacturing panel data, and the results are shown in Table 3.

Table 3
Panel unit root test of the manufacturing industry

Variable Method Whole industry Low-tech industry Medium-tech industry High-tech industry

Innov LL –5.952^(0.000) –9.121^(0.000) –3.099^***(0.001) 3.382(1.000)

IPS 1.728(0.958) –1.355^(0.088) 1.436(0.925) 4.431(1.000)

Capital LL –0.132(0.448) 1.319(0.906) –1.174(0.120) –0.511(0.305)

IPS 3.547(1.000) 3.324(1.000) 2.421(0.992) –0.568(0.285)

dInnov LL –10.123^(0.000) –7.844^(0.000) –5.834^(0.000) –2.877^(0.002)

IPS –8.440^(0.000) –6.730^(0.000) –4.746^(0.000) –2.453^(0.007)

dCapital LL –7.428^(0.000) –3.677^(0.000) –6.750^(0.000) –1.790^(0.037)

IPS –5.857^(0.000) –2.626^(0.004) –5.742^(0.000) –1.003(0.158)

2dInnov LL –27.881^(0.000) –20.558^(0.000) –17.533^(0.000) –7.431^(0.000)

IPS –22.561^(0.000) –19.971^(0.000) –23.792^(0.000) –8.929^(0.000)

2dCapital LL –32.273^(0.000) –16.233^(0.000) –14.495^(0.000) –6.466^(0.000)

IPS –22.004^(0.000) –14.409^(0.000) –15.611^(0.000) –6.221^***(0.000)

Variable	Method	Whole industry	Low-tech industry	Medium-tech industry	High-tech industry
Innov	LL	–5.952^***(0.000)	–9.121^***(0.000)	–3.099^***(0.001)	3.382(1.000)
	IPS	1.728(0.958)	–1.355^*(0.088)	1.436(0.925)	4.431(1.000)
Capital	LL	–0.132(0.448)	1.319(0.906)	–1.174(0.120)	–0.511(0.305)
	IPS	3.547(1.000)	3.324(1.000)	2.421(0.992)	–0.568(0.285)
dInnov	LL	–10.123^***(0.000)	–7.844^***(0.000)	–5.834^***(0.000)	–2.877^***(0.002)
	IPS	–8.440^***(0.000)	–6.730^***(0.000)	–4.746^***(0.000)	–2.453^***(0.007)
dCapital	LL	–7.428^***(0.000)	–3.677^***(0.000)	–6.750^***(0.000)	–1.790^**(0.037)
	IPS	–5.857^***(0.000)	–2.626^***(0.004)	–5.742^***(0.000)	–1.003(0.158)
2dInnov	LL	–27.881^***(0.000)	–20.558^***(0.000)	–17.533^***(0.000)	–7.431^***(0.000)
	IPS	–22.561^***(0.000)	–19.971^***(0.000)	–23.792^***(0.000)	–8.929^***(0.000)
2dCapital	LL	–32.273^***(0.000)	–16.233^***(0.000)	–14.495^***(0.000)	–6.466^***(0.000)
	IPS	–22.004^***(0.000)	–14.409^***(0.000)	–15.611^***(0.000)	–6.221^***(0.000)

Note: The sample interval of the test is from 2000 to 2015; d and 2d represent the first-order and second-order differences of the variables, respectively; the P values are in brackets; ^***, ^**, and ^* mean significant at 1%, 5%, and 10% levels, respectively.

As can be seen from Table 3, part of the original data failure to reject the null hypothesis that there is a unit root, which means the data is the non-stationary time series with a time trend. Except for the P-value of the second-order differential statistics of the high-tech industry, the P-value of the first-order differential statistics of other variables is all significant at 1%, rejecting the null hypothesis. It means that the first-order differential variables of sustainable innovation capacity and capital stock in the whole industry, the low-tech industry, and the high-tech industry are smooth, and the second-order differential variables in the high-tech industry are smooth. It may lead to false regression if using the ordinary regression method directly. Therefore, the co-integration method is needed for analysis.

There may be strong correlations between 28 manufacturing industries; our paper uses the Pedroni method to perform the co-integration test on the panel data. This method is more flexible and allows different co-integration vectors and residual autoregressive coefficients to exist in the panel units, that is, spatial correlation. The results are shown in Table 4. As can be found that the latter two statistics reported by the Pedroni method are significant at 1% level, reject the null hypothesis that there is no co-integration relationship between sustainable innovation capacity and capital stock. It suggests that there is a long-term and stable equilibrium relationship between these two variables, but it can’t explain the specific interaction relationship, which needs further tests.

Table 4

Panel co-integration test of the manufacturing industry

Statistics	Whole industry	Low-tech industry	Medium-tech industry	High-tech industry
Modified Phillips-Perron t	0.549(0.292)	0.023(0.491)	0.798(0.212)	0.236(0.406)
Phillips-Perron t	–9.378(0.000)	–7.023(0.000)	–6.235(0.000)	–2.734(0.003)
Augmented Dickey-Fuller t	–10.506(0.000)	–8.225(0.000)	–6.370(0.000)	–2.515(0.006)

Note: The P values are in brackets.

4.3 PVAR model

To ensure the validity of the model estimation results, the optimal lag term of the PVAR model is first determined.

According to the criteria of LR, FPE, AIC, SC, and HQ, the maximum number of ^* statistics is considered as the optimal lag order. Meanwhile, the smallest one is selected to reduce the loss of sample freedom. The results are shown in Table 5. It can be seen that the VAR model with lag order of 4 in the whole manufacturing industry has the goodness of fit and the residual sequence is relatively stable. The optimal lag order of the low-tech, medium-tech, and high-tech industries are 5.

Table 5
Testing results of AIC, BIC, and HQIC

Lag Whole industry Low-tech industry Medium-tech industry High-tech industry

AIC BIC HQIC AIC BIC HQIC AIC BIC HQIC AIC BIC HQIC

1 –1.953 –1.311 –1.698 –5.183 –4.636 –4.961 –2.421 –1.900 –2.209 0.683 1.151 0.860

2 –1.883 –1.156 –1593 –5.716 –5.056 –5.448 –2.627 –2.002 –2.237 0.804 1.453 1.044

3 –2.511 –1.687 –2.182 –9.919 –9.133^* –9.600 –2.864 –2.122 –2.562 1.147 1.910 1.452

4 –2.934^* –1.999^* –2.559^* –9.959 –9.030 –9.581 –3.169 –2.296 –2.814 1.358 2.414 1.727

5 –2.823 –1.758 –2.394 –10.05^* –8.957 –9.607^* –6.111^* –5.089^* –5.696^* –1.457^* –0.175^* –1.032^*

Lag	Whole industry	Low-tech industry	Medium-tech industry	High-tech industry
1	–1.953	–1.311	–1.698	–5.183	–4.636	–4.961	–2.421	–1.900	–2.209	0.683	1.151	0.860
2	–1.883	–1.156	–1593	–5.716	–5.056	–5.448	–2.627	–2.002	–2.237	0.804	1.453	1.044
3	–2.511	–1.687	–2.182	–9.919	–9.133^*	–9.600	–2.864	–2.122	–2.562	1.147	1.910	1.452
4	–2.934^*	–1.999^*	–2.559^*	–9.959	–9.030	–9.581	–3.169	–2.296	–2.814	1.358	2.414	1.727
5	–2.823	–1.758	–2.394	–10.05^*	–8.957	–9.607^*	–6.111^*	–5.089^*	–5.696^*	–1.457^*	–0.175^*	–1.032^*

Note: ^*represents the minimum value of AIC, BIC, and HQIC, that is, the optimal lag period.

The System-GMM method is usually selected in the PVAR model. As the estimation bias may be caused by time and fixed effects, the mean difference method and forward mean difference method (Helmert) are used to eliminate their effects and the System-GMM method is used for regression. The results are shown in Table 6.

Table 6

GMM estimation of the PVAR model

Variable	Whole industry		Low-tech industry		Medium-tech industry		High-tech industry
	h_Innov	h_Capital	h_Innov	h_Capital	h_Innov	h_Capital	h_Innov	h_Capital
L.h_Innov	0.830^***(0.101)	4.322(4.255)	–0.372^***(0.097)	–0.592(0.648)	–0.040(0.171)	1.256^***(0.492)	0.554^**(0.234)	–7.339(6.611)
L.h_Capital	–0.007(0.006)	0.630^**(0.360)	0.015(0.019)	0.378^**(0.188)	0.070^*(0.043)	0.251^**(0.105)	–0.003(0.004)	0.607^*(0.351)
L2.h_Innov	0.087(0.086)	1.253(1.344)	–0.101(0.079)	0.017(0.552)	0.044(0.160)	1.861^***(0.449)	0.361(0.237)	6.201(5.100)
L2.h_Capital	0001(0.002)	–0.067(0.086)	0.002(0.020)	0.389^***(0.103)	–0.002(0.002)	0.025^***(0.005)	–0.001(0.002)	–0.032(0.042)
L3.h_Innov	–0.034(0.045)	–1.859^**(0.825)	0.058^**(0.027)	–0.215(0.227)	0.169(0.109)	0.100(0.305)	0.037(0.147)	0.252(1.956)
L3.h_Capital	0.001(0.001)	0.017(0.023)	0.001(0.002)	–0.008(0.020)	–0.001(0.002)	0.017^***(0.004)	0.000(0.001)	–0.007(0.021)
L4.h_Innov	–0.008(0.035)	–0.507(1.074)	0.001(0.021)	0.296^*(0.156)	0.152(0.105)	0.329(0.246)	–0.128(0.096)	–1.824(1.825)
L4.h_Capital	0.001^**(0.000)	–0.014(0.012)	0.001(0.002)	0.039^**(0.016)	0.002^*(0.001)	0.041^***(0.003)	–0.002^***(0.001)	0.021(0.023)
L5.h_Innov	—	—	0.030(0.025)	–0.106(0.179)	0.232^**(0.106)	0.079(0.290)	–0.040(0.069)	1.589(1.695)
L5.h_Capital	—	—	0.003(0.003)	0.044^***(0.016)	–0.001(0.002)	–0.004(0.008)	0.004^***(0.001)	–0.003(0.019)

Note: h_Innov and h_Capital represent the sequences of Innov and Capital after Helmert conversion and elimination of fixed effects; L., L2., L3., L4. and L5. represent the lag period from 1 to 5; standard deviations in the brackets; ^***, ^**, and ^* mean significant at 1%, 5%, and 10% levels, respectively.

From the GMM estimation results of the entire manufacturing industry, it can be found that the impact of sustainable innovation capacity with 1 period lag on itself is 0.830, which is significant at 1% level. It means that the sustainable innovation capability of China’s manufacturing industry has a significant positive role in promoting its development, showing the characteristics of relying on its inertial development. The impact of the capital stock lagging four phases on sustainable innovation capacity is 0.001, which is significant at 5% level. It suggests that capital stock has a positive impact on sustainable innovation capacity, but has a lagging effect. When the capital stock is used as the explanatory variable, the lag capital stock has a positive effect on itself, significant at 5% level. The other lag periods have no significant impact on themselves, indicating that the capital stock of China’s manufacturing industry develops depending on its self-inertia. The inertia of capital stock is lower than that of sustainable innovation capability.

From the GMM estimation results of the various manufacturing segments, it can be found that the lag sustainable innovation capacity in the low-tech industry has a negative effect on itself, significant at 1% level. The three-period lag sustainable innovation capacity has a positive effect on itself, significant at 5% level, remaining positive in the subsequent periods. The reason may be that the self-reinforcing effect of sustainable innovation capacity is closely related to the foundation of innovation. The sustainable innovation ability in the low-tech industry is at a low level, and its promotion effect of itself in the first phase is less than the hindrance effect of resources consumption, which brings “crowding effect” for sustainable innovation capability (such as R&D expenditure, R&D personnel, etc.). It is difficult to achieve self-driving. Over time, the benefits brought by resource consumption gradually emerge, which provides convenient conditions for innovation implementation and is conducive to improving the innovation output capacity and sustainable development capacity of the manufacturing industry. The five-period lag sustainable innovation capacity in the medium-tech industry and the one period lag sustainable innovation capacity in the high-tech industry has a positive effect on themselves, with the coefficient of 0.232 and 0.554, which are both significant at 5% level. It indicates that such industries have a good foundation for innovation, and there is significant self-inertia in sustainable innovation capability. Meanwhile, the lag and the four-period lag capital stock in the medium-tech industry play a significant positive role in sustainable innovation capacity.

When the capital stock is used as the explanatory variable, the lag, 2 periods lag, 4 periods lag and 5 periods lag capital stock in the low-tech industry can promote itself. The first four periods lag capital stock in the medium-tech industry and the one-period lag capital stock in the high-tech industry has a significant positive impact on itself, indicating that capital stock of these industries has its inertia. Meanwhile, sustainable innovation capacity has a promotion effect on capital stock. In the low-tech industry, the impact of the four-phase lag sustainable innovation capacity on the capital stock is 0.296, which is significant at 10% level. In the medium-tech industry, the impact of one-phase lag and two-phase lag sustainable innovation capacity on the capital stock is 1.256 and 1.861, respectively, both of which are significant at 1%. Therefore, the sustainable innovation capacity is conducive to improving the capital stock of the low-tech and medium-tech industries, and it has a hysteretic nature in the former industry. However, in the high-tech industry, there is no significant correlation between sustainable innovation capabilities and capital stock.

4.4 Impulse response and variance decomposition

4.4.1 Impulse response

The GMM estimation results reflect the dynamic simulation process between sustainable innovation capability and capital stock, but not enough to reveal the transmission mechanism of these two variables. To further investigate the inter-temporal influence relationship and interaction mechanism between manufacturing sustainable innovation capability and capital stock, our paper makes the impulse response analysis. The duration of the response is set as phase 6, and the impulse response diagram between sustainable innovation capacity and the capital stock is obtained. The impulse response of sustainable innovation capability to itself is shown in Fig. 4, where the solid line represents the impulse response curve, the dotted lines mean the estimated values of the 95% and 5% sub-points, respectively. The horizontal axis is the number of predict periods, and the vertical axis is the degree of impact response.

Fig. 4

Impulse response of sustainable innovation capability to itself.

It can be seen from the Fig. 4(a) that the sustainable innovation capacity of the whole manufacturing industry is greater than 0 after being shocked by a standard deviation, indicating that it has a stable positive impact on itself, but shows a gradually weakening trend.

As can be seen from Fig. 4(b)–(d), the sustainable innovation capability of the low-tech industry is sensitive to its shocks. After being shocked by the first standard deviation, it has a self-reinforcing effect in the current period, and rapidly decline in the following period, performing a negative response. As the number of forecast periods increases, it shows a trend of fluctuating development and gradually converges to 0. The reason may be that the innovation development of China’s manufacturing industry has a significant role in promoting itself to high innovation ability in the current period, and it pays more attention to technological innovation, energy conservation, and emission reduction. But once the new technology is invented, it is quickly adopted by other industries because of its “quasi-public goods” nature (Macho and Perez) [35]. The resulting free-riding behaviour reduces the driving force for sustainable technological innovation in the manufacturing industry. The responses of the sustainable innovation ability of the medium-tech and high-tech industries to themselves are always positive, and the peak value is higher in the high-tech industry. It means that the sustainable innovation ability of such industries has a promotion effect on its development, that is, self-inertia. Such effect declines in the fluctuation.

The impulse response of capital stock to itself is shown in Fig. 5.

As can be seen from the Fig. 5(a), the capital stock in the whole industry responses quickly after being shocked by a standard deviation, and is greater than 0 in the current period and subsequent periods. That means the development of manufacturing capital stock has a stable self-inertia. The response in the current period reaches the maximum and gradually declines in the following forecast periods and eventually tends to 0. That’s because capital has the law of diminishing marginal utility (Keynes) [36], that is, the expected profit rate generated by capital investment decreases with the expansion of capital scale. Then the self-reinforcement effect bought by capital gradually weakens. It also shows that more capital stock does not mean better performance. The manufacturing industry should pay more attention to the optimized capital allocation instead of the blind expansion of capital scale.

The results of Fig. 5(b)–(d) show that the impact of capital stock in the low-tech and medium-tech industries on itself is complex. The overall response value is positive, but it declines rapidly in the first phase and flattens after the fourth phase showing a trend of decreasing volatility. The effect of the capital stock of the high-tech industry on itself is similar to the result of the whole industry and tends to be stable in the optimal lag period, indicating that the capital stock in this industry has self-inertia, but gradually weakens.

Fig. 5

Impulse response of capital stock to itself.

The impulse response of sustainable innovation capability to the capital stock in the manufacturing industry is shown in Fig. 6.

Figure 6(a) shows that the impulse response of capital stock is positive in the current period when the sustainable innovation capacity is shocked by a standard deviation. It reaches the maximum in the second period, and then decreases, indicating that the sustainable innovation capacity has a significant and continuous positive impact on capital stock. With the increase of sustainable innovation capability, it brings higher added value and industry profitability to the manufacturing industry. The profitable nature of capital makes high innovation capability industries attract more capital, which further verifies the measurement standard of capital allocation efficiency proposed by Wurgler [24].

Fig. 6

Impulse response of sustainable innovation capability to the capital stock.

The results of Fig. 6(b)–(d) present that the current impulses in the low-tech and high-tech industries are positive when the capital stock is shocked by the sustainable innovation capability. They quickly drop to the negative level, then rise to a positive value in the second period, showing the fluctuant development trend. However, the cumulative effects are positive, indicating that sustainable innovation capacity has a positive impact on capital stock, but it should be based on the long-term accumulation of sustainable innovation ability. In the medium-tech industry, the response value is always greater than 0, reaching a peak in the second phase, and gradually converging to 0.020% after the third phase. It means that sustainable innovation ability has a significant and sustained promoting effect on capital stock in the medium-tech industry.

The impulse response of capability capital stock to the sustainable innovation capability of China’s manufacturing industry is shown in Fig. 7.

Fig. 7

Impulse response of capability capital stock to sustainable innovation capability.

As can be seen from Fig. 7(a), the current response of sustainable innovation capacity is 0 when the capital stock is shocked by the first standard deviation. It gradually turns negative in the next two periods, indicating that the capital stock does not play its expected positive role. The increase of capital stock hinders the improvement of sustainable innovation capacity, which is contrary to the research conclusions of existing scholars [37 –40]. The reason may be that China’s manufacturing industry still belongs to the extensive development pattern, which mainly relies on cheap labour, land and other natural resources and high environmental pollution tolerance. It attaches great importance to economic benefits and ignores the sustainable innovation capability, resulting in insufficient R&D investment, especially environmental protection. Under the premise that sustainable innovation capability cannot be followed up, it is difficult to play its due role only by the increase of capital stock.

Additionally, many Chinese high-tech industries still belong to the labour-intensive industries and are at the low end of the global technology industry chain. Even if a large amount of capital is invested, it is difficult to translate smoothly into R&D results. The capital allocation efficiency and the R&D quality need to be improved. Once a new technology emerges, it will bring negative shocks to the existing technologies or innovation projects, resulting in the capital loss of the manufacturing industry and then affect the sustainable innovation ability.

The results of Fig. 7(b)–(d) show the current responses of sustainable innovation capacity in the low-tech and medium-tech industries are 0 after shocked by the capital stock. They reach to the peak in the first phase and keep positive in the subsequent forecast periods, indicating that the capital stock is funding basis for such industries to improve their sustainable innovation capability, which is conducive to promoting technological progress and innovation behaviour. The impact result of capital stock on sustainable innovation capacity in the high-tech industry is similar to that of the entire manufacturing industry. In addition to the above reasons, it may also be the following reasons. First, the sustainable innovation capacity of the high-tech industry encounters a bottleneck or not reaches a certain threshold, making it difficult for capital stock to play its due role. Second, the growth opportunities of enterprises exacerbate underinvestment [41].

4.4.2 Variance decomposition

The variance decomposition can reflect the interpretation degree of the shock of a single variable to the changes of other endogenous variables, which can systematically compare the relative importance of each shock. Therefore, our paper adopts this method to further analyse the interaction degree between sustainable innovation ability and capital stock of China’s manufacturing industry. The results are shown in Table 7.

Table 7
Variance decomposition of sustainable innovation capacity and capital stock

Variable S Whole industry Low-tech industry Medium-tech industry High-tech industry

Innov Capital Innov Capital Innov Capital Innov Capital

Innov 10 0.849 0.151 0.982 0.018 0.872 0.128 0.912 0.088

Capital 10 0.124 0.876 0.024 0.976 0.408 0.592 0.025 0.975

Innov 20 0.849 0.151 0.979 0.021 0.860 0.140 0.905 0.095

Capital 20 0.126 0.874 0.023 0.977 0.509 0.492 0.026 0.974

Innov 30 0.849 0.151 0.979 0.021 0.854 0.146 0.905 0.095

Capital 30 0.126 0.874 0.023 0.977 0.551 0.449 0.026 0.974

Variable	S	Whole industry	Low-tech industry	Medium-tech industry	High-tech industry
Innov	10	0.849	0.151	0.982	0.018	0.872	0.128	0.912	0.088
Capital	10	0.124	0.876	0.024	0.976	0.408	0.592	0.025	0.975
Innov	20	0.849	0.151	0.979	0.021	0.860	0.140	0.905	0.095
Capital	20	0.126	0.874	0.023	0.977	0.509	0.492	0.026	0.974
Innov	30	0.849	0.151	0.979	0.021	0.854	0.146	0.905	0.095
Capital	30	0.126	0.874	0.023	0.977	0.551	0.449	0.026	0.974

Note: S means the forecast periods.

As can be seen from the table that the variance decomposition results of the 20th and 30th phases are almost the same, indicating that the fluctuation of sustainable innovation capacity and the capital stock has stabilized since the 20th phase. Seen from the error term decomposition results of the sustainable innovation ability of the whole industry, it can be found that it contributes 84.9% of the explanatory power, and capital stock contributes 15.1%, indicating that the development of sustainable innovation capabilities mainly depends on its self-inertia and is supported by capital stock. The error term decomposition results of the capital stock show that it contributes 87.4% and the sustainable innovation ability contributes 12.6%, indicating that the development of capital stock is mainly dependent on its self-inertia and is weakly affected by the sustainable innovation ability. The sustainable innovation ability of the manufacturing industry may enter the bottleneck period or has not reached a certain threshold, resulting in the limited attracts for capital.

Besides, the explanatory power of capital stock to sustainable innovation ability (15.1%) is greater than the explanatory power of sustainable innovation ability to capital stock (12.6%), showing that the impact of capital stock on sustainable innovation capacity is greater than that of sustainable innovation capacity on capital stock in the manufacturing industry.

The variance decomposition results of different industries show that the explanatory power of sustainable innovation ability in the low-tech, medium-tech, and high-tech industries to explain themselves is 97.9%, 86.0%, and 90.5%, respectively. It means the sustainable innovation capability of these three industries has the characteristic of self-inertial. The capital stock has the least impact on sustainability innovation ability in the low-tech industry (only 2.1%) and the greatest impact in the medium-tech industry (14.0%). Moreover, the decomposition results of the capital stock have the industry heterogeneity. In the low-tech and the high-tech industry, the explanatory power of capital stock to itself is 97.7% and 97.4% respectively, and the sustainable innovation ability contributes 2.3% and 2.6%. In the medium-tech industry, capital stock contributes 59.2% to itself and sustainable innovation ability contributes 40.8%. It shows that the development of capital stock in the above industries has self-inertia, but the sustainable innovation capacity plays a significant role in the capital stock of the medium-tech industry

5 Conclusions and implications

5.1 Conclusions

Based on the panel data of China’s manufacturing industry from 2000 to 2015, our paper measures its sustainable innovation capability and capital stock, explores their dynamic relationship by using the PVAR model. The following research conclusions are obtained.

First, sustainable innovation ability and the capital stock have industry heterogeneity. The sustainable innovation capabilities in the Communication equipment, computer, and other electronic equipment manufacturing industry, the Electrical machinery and equipment manufacturing industry and the Transportation equipment manufacturing industry are at the highest level. The capital stocks in the Ferrous metal smelting and rolling processing industry, the Non-ferrous metal smelting and rolling processing industry and the Pharmaceutical manufacturing are at the highest level.

Second, the sustainable innovation capacity and capital stock of China’s manufacturing industry presents the characteristics self-inertial, and the former is greater than the latter. As can be seen from the results of different industries, the development of sustainable innovation capacity and capital stock of the low-tech, medium-tech, and high-tech industries also rely on their own inertial. The self-inertial in the former industries are more significant. In the low-tech and high-tech industries, the effect of capital stock on sustainable innovation ability is more significant, and in the low-tech and the medium-tech industries the effect of sustainable innovation ability on the capital stock is more significant.

Third, the sustainable innovation capabilities of the entire manufacturing industry, the medium-tech and high-tech industries have a stable positive impact on themselves, while the sustainable innovation capability of the low-tech industry changes from positive to negative, with a positive cumulative effect. In all of the manufacturing industries, the capital stock has the nature of self-inertial and gradually weakened. The sustainable innovation capacity of the entire industry and the medium-tech industry have a continuously positive impact on capital stock, while in the low-tech and high-tech industries, the impact changes from positive to negative. Additionally, the capital stock inhibits the development of sustainable innovation capability in the entire industry and the high-tech industry, however, in the low-tech and medium-tech industries, the capital stock has a stable positive role in sustainable innovation.

Forth, the development of sustainable innovation capacity and capital stock of the whole manufacturing industry in China mainly depends on its self-inertia. The impact of sustainable innovation capacity on the capital stock is weaker than that of capital stock on sustainable innovation capacity. The capital stock has the least impact on sustainable innovation capacity in the low-tech industry, and the sustainable innovation capacity has the greatest impact on capital stock in the medium-tech industry.

5.2 Implications

Our paper has the following enlightenment. China’s manufacturing industry still belongs to the extensive development pattern, which is at the low and middle end of the global value chain. To form a positive interaction between sustainable innovation capability and capital stock, the following measures should be taken.

From the micro perspective, the capital stock has self-inertia and plays a positive role in sustainable innovation capability in the low-tech and medium-tech industries. Such industries should focus on advantageous resources to strengthen their R&D investment, breakthrough core technologies, and improve the innovation ability to enhance the agglomeration effect on capital. In the high-tech industry, the impact of the capital stock on sustainable innovation capacity is negative. It should focus on improving original and major innovation and break through the bottleneck of innovation capacity to improve the economic, social and ecological benefits to achieve sustainable development. Meanwhile, such industries should pay more attention to the R&D investment efficiency to prevent invalid or duplicate capital investment. Only realizing the effective integration of technological innovation and capital deepening, can it solve the practical dilemma of the coexistence of increasing capital stock and decreasing technological progress in China’s manufacturing industry, especially the high-tech industry. Therefore, the manufacturing industry can improve its sustainable innovation capability through both internal-driven and capital-driven approaches, and form the interaction and mutual promotion of sustainable innovation capability and capital stock. It will eventually push China’s manufacturing industry from “Made in China” to “Created in China”.

From the macro perspective, the Chinese government should strengthen its policy support for the sustainable innovation capability of the manufacturing industry especially the low-tech and medium-tech industries. It helps to stimulate their inherent innovation momentum and promote capital flows to such industries, which is conducive to fully utilizing the role of capital stock in promoting sustainable innovation capacity. Meanwhile, the Chinese government should promote market-oriented reforms, adjust investment-oriented science and technology management policies, and change the extensive development pattern driven by capital investment in the high-tech industry. It enables the high-tech industry to breakthrough technological innovation bottlenecks, and reduces invalid capital investment and improves the efficiency of innovation resource utilization. Only by realizing the virtuous circle of sustainable innovation capacity and capital stock can we continuously improve the quality and economic benefits of the manufacturing industry, and then promote China’s economic growth and global competitiveness.

Footnotes

Acknowledgments

The work is supported by the China Social Science Foundation (15BJY065) and the International Exchange Project of Independent Innovation of Wuhan University of Technology (2018-JL-008).

References

Grigorescu

, Maer-Matei

M.M.

, Mocanu

and Zamfir

A.M.

, Key Drivers and Skills Needed for Innovative Companies Focused on Sustainability, Sustainability12 (2020), 102–115.

Zhang

, Song

, Yang

X.M.

and Li

, What are Important Technologies for Sustainable Development in the Trucking Industries of Emerging Markets? Differences between Organizational and Individual Buyers, Sustainability12 (2020), 224–246.

Collins

and Porras

J.I.

, Built to Last: Successful Habits of Visionary Companies, NewYork: Harper Collins Publishers (1997).

Fang

, Bai

and Bilan

, Evaluation Research of Green Innovation Efficiency in China’s Heavy Polluting Industries, Sustainability12 (2020), 146–166.

Bessant

, Caffyn

, Gilbert

, Harding

and Webb

, Rediscovering Continuous Improvement, Technovation14 (1994), 17–29.

Boer

, Berger

, Chapman

and Gertsen

, CI Changes: From Suggestion Box to Organisational Learning-Continuous Improvement in Europe and Australia, Aldershot: Ashgate Publishing (2000).

Griliches

, Issues in Assessing the Contribution of Research and Development to Productivity Growth, Bell Journal of Economics10 (1979), 92–116.

Jonker

, Henny

and Adam

, Technological Effort, Technological Capabilities and Economic Performance: A Case Study of the Paper Manufacturing Sector in West Java, Technovation26 (2006), 121–134.

Wang

, Liu

and Chen

, Evaluating Firm Technological Innovation Capability Under Uncertainty, Technovation28 (2008), 349–363.

10.

Gloet

and Terziovski

, Exploring the Relationship between Knowledge Management Practices and Innovation Performance, Journal of Manufacturing Technology Management15 (2004), 402–409.

11.

Leonard

J.A.

and Waldman

, An Empirical Model of the Sources of Innovation in the U.S. Manufacturing Sector, Business Economics42 (2007), 33–45.

12.

Priit

, Does FDI Spur Productivity, Knowledge Sourcing and Innovation by Incumbent Firms? Evidence from Manufacturing Industry in Estonia, World Economy34 (2011), 1308–1326.

13.

, Delin

, Shumin

, Xiang

and Jizhen

, Open Innovation and Firm Performance: Evidence from the Chinese Mechanical Manufacturing Industry, Journal of Engineering and Technology Management48 (2018), 76–86.

14.

Merih

and Tian.

, Acquiring Innovation, AFA 2012 Chicago Meetings Paper (2012).

15.

Andrea

, Marion

and Grazia

L.G.

, The Impact of Internationalization on Innovation at Countries’ Level: the Role of Absorptive Capacity, Cambridge Journal of Economics41 (2017), 413–439.

16.

Schumpeter

J.A.

, The Theory of Economic Development, Cambridge: Harford University Press (1912).

17.

Lucas

R.E.

, On the Mechanics of Economic Development, Journal of Monetary Economics22 (1988), 3–42.

18.

Hekkert

M.P.

, Suurs

R.A.

, Negro

S.O.

, Kuhlmann

and Smits

R.E.

, Functions of Innovation Systems: A New Approach for Analysing Technological Change, Technological Forecasting and Social Change74 (2007), 413–432.

19.

Mourad

and Ahmed

Y.S.E.

, Perception of Green Brand in an Emerging Innovative Market, European Journal of Innovation Management15 (2012), 514–537.

20.

D’Angelo

, Majocchi

, Zucchella

and Buck

, Geographical Pathways for SME Internationalization: Insights from an Italian Sample, International Marketing Review30 (2013), 80–105.

21.

Alpert

F.H.

and Kamins

M.A.

, Pioneer Brand Advantage: A Conceptual Framework and Propositional Inventory, Journal of the Academy of Marketing Science22 (1994), 244–253.

22.

Clemens

B.W.

and Douglas

T.J.

, Understanding Strategic Responses to Institutional Pressures, Journal of Business Research58 (2005), 1205–1213.

23.

Liu

and Zhang

, Joint Decisions on Pricing and Advertising of Dominant Manufacturer with Different Advantage Strategies, International Journal of Modelling and Simulation36 (2016), 97–105.

24.

Marx

, Capital: A Critique of Political Economy. Penguin Classics (1992).

25.

Wurgler

, Financial Markets and the Allocation of Capital, Journal of Financial Economics58 (2001), 187–214.

26.

Madsen

Jakob B.

, Growth and Capital Deepening Since: Is It All Technological Progress?, Journal of Macroeconomics32 (2009), 641–656.

27.

Berger

A.N.

and Udell

G.F.

, Collateral, Loan Quality and Bank Risk, Journal of Monetary Economics25 (1990), 21–42.

28.

Hall

B.H.

and Lerner

, The Financing of R&D and Innovation - Handbook of the Economics of Innovation, North-Holland (2010), 609–639.

29.

Howell

, Firm R&D, innovation and easing financial constraints in China: Does corporate tax reform matter?, Research Policy45 (2016), 1996–2007.

30.

Furman

J.L.

and Hayes

, Catching Up or Standing Still?: National Innovative Productivity among Follower Countries, -, Research Policy33 (2004), 1329–1354.

31.

Vincent

, Laure

, N’Doli

G.A.

and Mauricio

, Evaluating Innovative Processes in French Firms: Methodological Proposition for Firm Innovation Capacity Evaluation, Research Policy43 (2014), 608–622.

32.

Goldsmith

R.W.

, A Perpetual Inventory of NationalWealth. NBER Studies in Income and Wealth, New York: National Bureau of Economic Research (1951).

33.

Kohli

, A Gross National Product Function and the Derived Demand for Imports and Supply of Exports, Canadian Journal of Economics11 (1978), 167–182.

34.

Levin

, Lin

C.F.

and Chu

C.J.

, Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties, Journal of Econometrics108 (2002), 1–24.

35.

K.S.

, Pesaran

H.M.

and Shin

, Testing for Unit Root in Heterogeneous Panels, Journal of Econometrics115 (2003), 53–74.

36.

Macho

S.I.

and Perez

C.D.

, Optional Enforcement Policy and Firms’ Emissions and Compliance with Environmental Taxes, Journal of Environmental Economics and Management51 (2004), 110–131.

37.

Keynes

J.M.

, The General Theory of Employment Interest and Money, Kessinger Publishing (1936).

38.

Borensztein

, Gregorio

and Lee

J.W.

, How does Foreign Direct Investment Affect Economic Growth?, Journal of International Economics45 (1995), 115–135.

39.

Laursen

, Masciarelli

and Prencipe

, Regions Matter: How Localized Social Capital Affects Innovation and External Knowledge Acquisition, Organization Science23 (2012), 177–193.

40.

Blandine

, Knowledge-Capital and Innovation, Springer, New York (2013).

41.

Sulistyo

and Siyamtinah , Innovation Capability of SMEs Through Entrepreneurship, Marketing Capability, Relational Capital and Empowerment, Asia Pacific Management Review21 (2016), 196–203.