Green Technology Innovation, Capital-Factor Allocation, and Manufacturing-Export Resilience

Abstract

Green technology innovation, with its two-fold benefits of protecting the environment and promoting economic growth, is an increasingly necessary strategy for China’s manufacturing exports. This study examines the impact and mechanisms of green technology innovation on China’s manufacturing-export resilience in the aftermath of the 2008 global financial crisis. The study findings demonstrate that green technology innovation considerably boosts manufacturing-export resilience. Specifically, green technology innovation enhances manufacturing-export resilience by improving capital allocation efficiency. This is especially significant in the high labour-mismatch region and the eastern–central region. It is worth noting that the impact of green technology innovation on manufacturing-export resilience is negatively regulated by intellectual-property protection, as well as being subject to the single-threshold effect of government intervention. That is, reinforcing intellectual-property protection inhibits the upgrading effect of green technology innovation on manufacturing-export resilience, and the effect of green technology innovation on manufacturing-export resilience becomes insignificant when the government intervention goes beyond a certain level.

1. Introduction

The gradual disappearance of the demographic dividend, the recurrence of financial risks and the sudden outbreak of COVID-19 impacts China’s supply chain and industrial chain development. Together, these and other challenges, coupled with the further weakening of the economic growth momentum brought about by the traditional mode of development in the country, exacerbate the uncertainty and vulnerability of the development of manufacturing exports. In light of the complex and volatile development situation at home and abroad, how to maintain long-term sustainable and stable development in the face of the impact on manufacturing exports is a major practical problem that needs to be faced and solved. Although China’s manufacturing exports are due to the complete industrial chain and support an export “substitution effect”, relative to other countries, they show strong resilience in the short term. It is important to note, however, that in addition to shocks from short-cycle fluctuations such as financial crises [1], manufacturing-export development is also exposed to shocks from long-term slow disturbances such as climate change and resource depletion [2], which puts it in a long-term evolutionary process that gives rise to differences in export resilience. In the face of the impact of short-cycle fluctuations, the government, through its intervention policy, can enable manufacturing exports to return to their original growth state relatively quickly after suffering a shock, which is reflected in the ability of manufacturing exports to resist and recover from a single contingency. In the face of the impact of long-term disturbances, the failure to consider the long-term path derivation and industrial upgrading process and an over-reliance on the original development path, factor utilisation, and institutional framework will make manufacturing exports, even if they return to the original growth equilibrium, inevitably decline due to the low-end locked-in development path when faced with a new round of external shocks in the future. Therefore, to repair the growth process of China’s manufacturing exports that is being impacted, and through the mutual adjustment and co-evolution of internal factors to ensure that manufacturing exports achieve the ability to undergo adaptive development, such that the growth pattern of manufacturing exports moves from recovery to innovation, the traditional mindset that relies on the advantages of cheap resource factors to promote growth must be transformed. Against the background of the decline in the marginal increase in the global supply of resources, the alleviation of the ecological and environmental crisis can provide a rare historical opportunity for manufacturing exports to cope with the violent external impacts of the financial crisis, COVID-19, and so on. If we can make reasonable use of the opportunities brought about by external forces, find the direction of the development of manufacturing exports, and upgrade towards digitisation, intelligence, and greening [3,4,5,6], we can not only make the manufacturing exports have a better resistance and recovery ability in the face of shocks but to even gradually adapt to the external environment and to carry out self-renewal and orientation. To realise the transformation and upgrading of the manufacturing-export growth mode towards greening, innovation and technological progress are the key factors. Green technological innovation will bring about industrial innovation and extinction, thus promoting the adjustment of the manufacturing production system and the reconstruction of the export structure.

So, in the face of the slow perturbation of global climate change and resource depletion, how can the pressure from the global green transition be transformed into an opportunity for China’s manufacturing industry to enhance its international competitiveness? Can green technology innovation become a new engine to enhance manufacturing-export resilience? This requires us to examine the risk-absorbing capacity of the manufacturing-export system from a more in-depth and comprehensive perspective and to explore the path of reshaping export resilience.

Compared with existing studies, the marginal contributions of this paper are as follows: Firstly, by identifying and analysing manufacturing-export resilience, it enriches the meaning and measurement of export resilience. Secondly, it analyses the intrinsic logic of how green technological innovation affects manufacturing-export resilience and empirically examines the mediating effect played by capital allocation efficiency; at the same time, it explores the moderating effect of intellectual-property protection and the threshold effect of government intervention, which is a further exploration of the path of how green technological innovation affects manufacturing-export resilience. Thirdly, based on how different types of regions provide differentiated evidence of green technological innovation affecting manufacturing-export resilience, clear green technological innovations to enhance manufacturing-export resilience may exist in the boundary conditions, which is a useful supplement to the analysis of geographical differences in export resilience.

The remainder of the paper is structured as follows. Section 2 presents the literature review. Section 3 explains the selection of variables, measurement methods, and sources of data collection; designs and fits the empirical model; and conducts a mechanistic analysis. In addition, we test the heterogeneity and robustness of the empirical model and discuss the findings. Section 4 presents the empirical results. Section 5 further investigates the moderating effect of intellectual-property protection and the threshold effect of government intervention. Section 6 discusses the empirical results in detail. Section 7 summarises the findings of the paper.

2. Literature Review

Resilience is an interdisciplinary concept and its perception has evolved from “engineering resilience” [7] and “ecological resilience” [7] to “adaptive resilience [8]”. While engineering resilience is used to describe the ability of a system to return to an equilibrium or stable state after a disturbance [9], ecological resilience breaks away from the idea of engineering resilience as a “single equilibrium” [10] by suggesting that ecosystems that are disturbed will change the original equilibrium, recover, and evolve to a new equilibrium state. Although resilience is an inherent property of a system, it is independent of shocks and evolves according to the external environment, so that there is no stable equilibrium [2]. Therefore, scholars began to try to introduce complexity theory and evolutionary ideas to promote the idea of resilience from a static equilibrium and a rebound to non-equilibrium, evolution, and diversity and thus put forward the concept of adaptive resilience in line with the logical core of evolutionary economic geography [11]. Adaptive resilience no longer focuses on whether the system can reach an equilibrium state after a shock but emphasises the system’s coping and dynamic self-adjustment ability in the face of shocks [12]. Martin [1], based on the concept of adaptive resilience, defines economic resilience as an adaptive, dynamic adjustment capacity, which includes the ability of economic agents to resist crises, recover from crises, self-renew, and reposition themselves, which has been affirmed and adopted by several scholars [13].

With the deepening of scholars’ understanding of the concept of resilience, the measurement and evaluation methods of economic resilience are also being enriched. Currently, there are two main categories of methods for measuring and evaluating resilience: one is the single-indicator sensitivity analysis method, which judges the resilience of an economic system to a crisis by calculating the gap between the trend and actual values of the core variables before and after the occurrence of an external disturbing shock. In empirical studies, the core variables in a single indicator mainly contain the GDP growth rate, unemployment rate, and employment [14,15,16]. The other category is the indicator system approach, which measures economic resilience by screening economic indicators that are highly correlated with resilience and constructing a basket of indicator systems [17]. Briguglio et al. [18] constructed an economic-resilience index system from four fundamentals: macroeconomic stability, social development, market efficiency, and economic governance. Yu et al. [19] measured regional economic resilience from six dimensions, including the economic level, innovation level, and upon opening up. Ubago et al. [20] constructed a composite index of regional economic resilience in Spain from the dimensions of industrial structure, capital value, labour force, economic level, and so on. Since most existing studies define export resilience as the ability of exports to resist and recover in the face of external shocks, the sensitivity analysis of a single index is mainly used to measure it, which mostly selects the export growth rate and export value as core variables to evaluate the export resilience in different periods [21].

Through the empirical analysis of factors affecting export resilience, in the case of external shocks, the path to improving export resilience may come from export policies, structural upgrading, technological innovation, local market size, and other aspects. With the rise of the global green revolution, the field of international trade has gradually manifested itself as a combination of environmental governance and trade measures, so that the traditional resource- and environment-consuming mode of export development at home has been constrained, and the green growth model has become an inevitable choice for achieving benign growth in manufacturing exports. Braun and Wield [22] believe that green technology is a technology that reduces environmental pollution and the consumption of energy and raw materials, so green technology innovation is considered to be an important means of ameliorating ecological damage and achieving sustainable development [23], and it is committed to the pursuit of a “win–win” situation between the environment and the economy [24]. Most scholars believe that technological innovation can be used to improve the environment [25]. Environmental regulation and technological innovation are linked when examining the impact on export trade. Technological innovation triggered by environmental regulation can improve the quality of export products [26,27], enhance export incentives, and protect the international competitiveness of products [3]. Haddoud et al. [28], using a sample of Polish family firms, find that strategic commitment to environmental issues has a positive impact on technological innovation, which increases export intensity. Bertarelli and Lodi [29], using microdata, empirically find that environmental taxes positively affect the propensity for technology innovation and, consequently, indirectly contribute to the propensity to export. In addition to the linear effect, some scholars have found that technological innovation under environmental regulation has a “U-shaped” promotional effect on the export competitiveness of China’s manufacturing industry [30], and that technological innovation hurts export trade before reaching the inflexion point. Qiang et al. [31] also found that, in the long run, and when environmental regulation reaches a certain intensity, technological innovation has a promotional effect on exports. Thus, green technological innovation is bound to have an important impact on manufacturing-export resilience.

In summary, the existing research has laid a good foundation for this paper, but research on resilience in the economic field started late, and that research is more concentrated on the regional economy and urban economy. Therefore, there are still some problems that need to be discussed more deeply: Firstly, there are few empirical studies directly on export resilience, and there is a lack of identification and analysis of export resilience in the manufacturing industry. The existing research on export resilience is mostly based on the short-term perspective, focusing on the static timepoint analysis of the resistance and recovery ability of a single emergency without considering the long-term path derivation and industrial upgrading process. Secondly, the research on the influencing factors of export resilience is not sufficient, and the impact of exports on long-term slow disturbances (such as climate change, resource depletion) is not discussed in depth. Thirdly, the existing research has recognised the importance of improving the level of green technology innovation in the manufacturing industry, but there is little analysis on the mechanism of how green technology innovation acts on export trade.

So, what is the path of action for green technology innovation in affecting manufacturing-export resilience? From a supply-side perspective, technology innovation will certainly increase firm productivity [32], reduce production costs, and enhance export competitiveness through price advantages. It has been shown in the literature that technological innovation is an important measure to optimise capital allocation efficiency in many industries [33], and capital allocation efficiency largely affects firm productivity [34], so green technology innovation is bound to have an impact on manufacturing-export resilience by changing the capital allocation efficiency. Based on this, this paper puts green technology innovation, capital allocation efficiency, and manufacturing-export resilience under the same research framework to explore whether green technological innovation can adjust the manufacturing production capacity to ensure long-term export resilience by optimising the capital allocation efficiency.

3. Materials and Methods

3.1. Measures

3.1.1. Dependent Variable

Manufacturing-export resilience (Res). In this paper, the financial crisis of 2008 is selected as the research backdrop, the study of Wei et al. [35] is referred to, and the deviation of the manufacturing-export value of each province in the corresponding year from the manufacturing-export value of 2008 is used to express the manufacturing-export resilience. The specific calculation method is as follows:

$${Res}_{it}=\frac{{Export}_{it}-{Export}_{i,2008}}{{Export}_{i,2008}}$$

(1)

In Equation (1), $i$ and $t$ represent the region and year, respectively, ${Res}_{it}$ represents the manufacturing-export resilience in year $t$ of region $i$, and ${Export}_{it}$ and ${Export}_{i,2008}$ represent the export value of the manufacturing industry in each region in the corresponding year and in 2008, respectively. The larger the value of ${Res}_{it}$ is, the stronger the manufacturing-export resilience is, and vice versa; the weaker the manufacturing-export resilience, the smaller the value of ${Res}_{it}$.

3.1.2. Independent Variable

Green technology innovation ($LnGti$). Patent-related data are often used by scholars as a measure of technological innovation [36]. Green patents can intuitively reflect the output of provincial green technological innovation activities, and the number of authorisations can directly reflect the quality of output of green technological innovation compared with the number of applications [37]. Therefore, this paper refers to Feng et al. [38] and selects the logarithm of the number of green invention patents granted as a proxy indicator for measuring provincial green technological innovation.

3.1.3. Mediating Variable

Capital allocation efficiency ($MisK$). This paper uses the degree of capital mismatch to measure the height of the capital allocation efficiency in the manufacturing industry, and this variable is negatively correlated with the capital allocation efficiency. The greater the degree of capital mismatch, the lower the capital allocation efficiency of the manufacturing industry, and vice versa. This paper refers to the treatment of Bai and Liu [39] and uses the capital-mismatch index ${\tau }_{Ki}$ to indicate the degree of regional capital mismatch as follows:

$${\gamma }_{Ki}=\frac{1}{1-{\tau }_{Ki}}$$

(2)

In Equation (2), ${\gamma }_{Ki}$ is the coefficient of the absolute factor price distortion, which represents the additive case when capital is relatively undistorted. The coefficient of relative price distortion is used instead in the actual measurement:

$${\widehat{\gamma }}_{Ki}=\frac{\frac{K_i}{K}}{\frac{s_i{\beta }_{Ki}}{{\beta }_K}}$$

(3)

In Equation (3), $\frac{K_i}{K}$ denotes the actual proportion of regional capital employed $K_i$ to the total capital $K$ of the economy as a whole, $s_i=\frac{p_i{y}_i}{Y}$ denotes the share of regional output $y_i$ to the output $Y$ of the economy as a whole, ${\beta }_K={\sum }_i^N{s}_i{\beta }_{Ki}$ denotes the value of the weighted contribution of capital to regional output, and $\frac{s_i{\beta }_{Ki}}{{\beta }_K}$ denotes the theoretical proportion of capital employed in the region when there is an efficient allocation of capital. The ratio of the two indicates the extent to which the amount of capital deviates from its effective allocation when it is used; if the ratio is greater than 1, it means that the region has been over-allocated capital; conversely, a ratio of less than 1 means that the region is under-allocated capital. Therefore, to calculate the capital-mismatch index ${\tau }_{Ki}$, one must first estimate the factor-output elasticity of capital in each region ${\beta }_K$. Assuming that the production function is a C–D production function with constant returns to scale, the specific form is as follows:

$$Y_{it}=A{K}_{it}^{{\beta }_{Ki}}{L}_{it}^{1-{\beta }_{Ki}}$$

(4)

Taking the natural logarithm of both sides at the same time, and adding the individual effect ${\mu }_i$ and the time effect ${\lambda }_t$ to the model, the collation can be obtained:

$$\ln \left(\frac{Y_{it}}{L_{it}}\right)=lnA+{\beta }_{Ki}\ln \left(\frac{K_{it}}{L_{it}}\right)+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(5)

In Equation (5), $Y_{it}$ is an output variable expressed as the GDP for each province, with 2009 as the base period, and the GDP for other years is transformed into the real GDP at 2009 constant prices using the GDP deflator. $L_{it}$ is the quantity of labour input, expressed as the annual average of employment in each province, the arithmetic average of the number of persons employed at the beginning of the year, and the number of persons employed at the end of the current year. $K_{it}$ is the amount of capital inputs, expressed in terms of the fixed capital stock of each province, which is calculated using the perpetual inventory method with the following formula:

$$K_t=\frac{I_t}{P_t}+\left(1-{\delta }_t\right)K_{t-1}$$

(6)

In Equation (6), $K_t$ denotes the fixed capital stock in the current period, $I_t$ denotes the nominal gross fixed-capital formation in the current period, $P_t$ is the fixed-asset investment price index, ${\delta }_t$ denotes the depreciation rate, which is taken to be 9.6%, and $K_{t-1}$ denotes the fixed-capital stock in the previous period. After estimating the factor-output elasticity of each province, the capital-mismatch index ${\tau }_{Ki}$ for each province is calculated based on the above equation.

Due to the existence of both under-capitalisation $\tau >0$ and over-capitalisation $\tau <0$, the absolute value of ${\tau }_{Ki}$ is treated to make the regression direction consistent. Larger values indicate a more serious capital mismatch.

3.1.4. Moderating Variable

Intellectual-Property Protection (IPP). China’s intellectual-property protection is characterised by a “dual-track system” of judicial and administrative protection. Unlike judicial protection, administrative protection takes fuller account of regional realities. This paper uses the number of intellectual-property protection systems established by local governments to indicate the level of intellectual-property protection, specifically including local laws and regulations, local normative documents, local government regulations, and local working documents.

3.1.5. Threshold Variable

Government Intervention (GI). This paper uses the ratio of government fiscal expenditure to regional GDP for each province.

3.1.6. Control Variables

This paper refers to the research of existing scholars and selects the manufacturing industry indicators of each province as control variables. These include:

Asset Size (AS). This indicator is expressed using the total assets of the manufacturing industry in each province. The larger the asset size of a manufacturing industry, the larger its scale of production and the more difficult it is to recover pre-shock export levels in the face of a shock.

Industry Profitability (IP). This indicator is expressed as the ratio of operating profit to the main business revenue of the manufacturing industry in each province. Higher industry profitability means weaker competition in its markets, less product substitutability, and greater post-crisis recovery.

Asset–Liability Ratio (ALR). This indicator is expressed as the ratio of total liabilities to total assets of the manufacturing industry in each province. A larger asset–liability ratio indicates that the manufacturing industry faces fewer financing constraints and can face export risks effectively.

Labour-Force Size (LFS): This indicator is expressed as the average number of workers employed in the manufacturing sector in each province. A larger labour-force size indicates greater production capacity and greater resilience to shocks.

3.2. Data Source and Variables Description

This paper takes the financial crisis of 2008 as the research backdrop. Using the export data of 2008 as the basis for comparison and considering the need for continuity of the data, we take 30 Chinese provinces as samples from 2009–2020, examining the effect of green technological innovation on manufacturing-export resilience by constructing a double fixed-effect model and a mediated-effect model. (We excluded Tibet, Hong Kong, Macao and Taiwan from our study because the data we needed for our study were not available in full). Provincial manufacturing-import and export-value data are obtained from the National Research Network statistical database, matching the HS two-digit code (96th edition) with the National Standard Industry Classification Code (2017). (The HS Code is the customs code that stands for Harmonised System, the full name of which is the International Convention for Harmonised Commodity Description and Coding System). Considering the availability and completeness of the data, the three manufacturing sub-industries of Other Manufacturing, Comprehensive Utilisation of Waste Resources and Metal Products, and Machinery and Equipment Repair were excluded. The number of green invention patents granted is from the China Research Data Service Platform (CNRDS). The number of local laws and regulations, local normative documents, local government regulations, and local working documents required for intellectual-property protection is from the Peking University Fabulous Database. The data on the GDP, average annual employment and fixed capital stock of each province, the fiscal expenditure required for government intervention, and the control variables required for the capital-mismatch index are obtained from the EPS database. The definitions of the variables selected in this paper are given in

Table 1. Variable definitions.

Variable Classification	Variable	Variable Definition
Dependent variable	Res	Deviation of manufacturing exports of each province in the corresponding year from manufacturing exports in 2008
Independent variable	LnGti	The logarithm of the number of patents granted for green inventions
Mediating variable	MisK	Capital-mismatch index
Moderating variable	IPP	Number of local governments supplying intellectual-property protection systems
Threshold variable	GI	The ratio of local government fiscal expenditure to regional GDP
Control variables	AS	Total asset size of the manufacturing industry
	IP	The ratio of operating profit to main business income in the manufacturing sector
	ALR	The ratio of total manufacturing liabilities to total industry assets
	LFS	Average number of workers in the manufacturing sector

, and the results of the statistical characteristics of the relevant variables are shown in

Table 2. Descriptive statistics.

Variable	N	Mean	Std. Dev	Min	Max
Res	360	0.8074	1.5391	−0.7169	10.5308
LnGti	360	5.8937	1.5178	0	9.2113
MisK	360	0.2689	0.2943	0.0015	1.9343
IPP	360	40.7278	31.3458	1	144
GI	360	0.2447	0.1021	0.0964	0.6430
AS	360	3.1218	2.6810	0.1291	15.0718
IP	360	0.0653	0.0277	−0.2074	0.1574
ALR	360	0.5827	0.0561	0.4208	0.7608
LFS	360	3.0271	3.2422	0.1042	15.6800

. As can be seen from Table 2, the minimum value of manufacturing-export resilience is −0.7169, while the maximum value is 10.5308, which indicates that there is a great difference in manufacturing-export resilience in different regions, and the mean is 0.8074 (>0), which indicates that manufacturing exports in various regions have shown some resilience. The maximum value of green technology innovation is 9.2113, while the minimum value is 0, indicating that there are large differences in the development of green technology innovation in various regions, which also reflects the imbalance in the development of various regions.

3.3. Research Design

This paper uses the overall data, and there is no random sampling problem. In order to eliminate the interaction between time and the individual, which leads to a deviation in the estimation results, we have chosen to construct a double-fixed effect model.

3.3.1. Basic Model

To explore the direct impact of green technology innovation on manufacturing-export resilience, this paper constructs the following econometric model:

$${Res}_{it}={\alpha }_0+{\alpha }_1{LnGti}_{it}+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(7)

In Equation (7), $i$ and $t$ represent regions and years, respectively, and the dependent variable ${Res}_{it}$ represents the manufacturing-export resilience of region $i$ in year $t$. The independent variable ${lnGti}_{it}$ represents the level of green technological innovation of region $i$ in year $t$, Cons represents the control variables, ${\mu }_i$ and ${\lambda }_t$ represent the region fixed-effect and year fixed-effect, respectively, and ${\epsilon }_{it}$ represents the random disturbance term. If the coefficient estimate of green technology innovation, ${\alpha }_1$, is greater than zero, it indicates that green technology innovation can enhance the manufacturing-export resilience in each region.

3.3.2. A Mediating-Effect Model

This paper follows the methodology proposed by Muller et al. [40] to construct a mediating-effect model for empirical testing, and the constructed econometric model is as follows:

$${MisK}_{it}=b_0+b_{1}Ln{Gti}_{it}+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(8)

$${Res}_{it}=c_0+c_1{LnGti}_{it}+c_2{MisK}_{it}+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(9)

In Equation (8), the mediating variable is ${MisK}_{it}$, and the rest of the settings are consistent with the benchmark model above. In this paper, we first test the coefficient estimate $a_1$ of Equation (7)’s $Ln{Gti}_{it}$; when $a_1$ is significant, then we test $b_1$ and $c_2$ of Equations (8) and (9) in turn; when $b_1$ and $c_2$ are significant and $c_1$ < ${\alpha }_1$, it indicates that there is the existence of a mediating effect.

4. Results

4.1. Green Technology Innovation and Manufacturing-Export Resilience

Based on Equation (7), this paper first conducts a regression estimation of the effect of green technological innovation on manufacturing-export resilience, and the results are reported in

Table 3. Baseline regression.

Variable	(1) Res	(2) Res	(3) Res	(4) Res
LnGti	0.467 ***	0.833 ***	0.264 ***	0.495 ***
	(0.094)	(0.129)	(0.080)	(0.180)
AS	−0.175 ***	−0.009	−0.399 ***	−0.051
	(0.053)	(0.057)	(0.075)	(0.053)
IP	1.203	5.936 **	8.389 ***	7.410 ***
	(1.733)	(2.301)	(2.153)	(2.988)
ALR	−1.013	−3.625 *	−1.348	−1.501
	(1.608)	(2.089)	(1.456)	(2.037)
LFS	0.008	0.279 ***	0.223 ***	0.464 ***
	(0.027)	(0.085)	(0.045)	(0.093)
Cons	−0.912	−3.196 *	0.058	−2.964 *
	(1.284)	(1.790)	(1.196)	(1.775)
Area FE	N	Y	N	Y
Year FE	N	N	Y	Y
N	360	360	360	360
R-squared	0.110	0.753	0.250	0.776

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

. Column (1) adds only control variables. Column (2) further controls for area-fixed effects. Column (3) controls for year-fixed effects. Column (4) controls for both area-fixed effects and year-fixed effects. The coefficient estimate of $Ln{Gti}_{it}$ is positive and significant. This indicates that the higher the level of green technology innovation, the stronger the manufacturing-export resilience; there is a significant positive relationship between green technology innovation and manufacturing-export resilience in each province. In summary, the regression results in Table 3 indicate that green technology innovation can enhance manufacturing-export resilience.

Regarding the estimation results of the control variables, the coefficient estimates of AS and ALR are negative in all regression results, which is consistent with the symbolic direction in the theoretical expectation, but they are no longer significant after controlling for regional fixed effects and two-way fixed effects. The coefficient estimate of IP is significantly positive after adding the regional fixed effect, time fixed effect, and two-way fixed effect, indicating that the enhancement of the profitability of the manufacturing industry has a significant contributing effect on the enhancement of export resilience. The coefficient estimate of the LFS is also significantly positive after adding the three fixed effects, indicating that the labour-force size of the manufacturing also positively contributes to the enhancement of the manufacturing-export resilience.

4.2. Mediating Effect of Capital Allocation Efficiency

Using Equations (7)–(9), this paper examines the mediating role played by capital allocation efficiency in the impact of green technology innovation on manufacturing-export resilience. The regression results are shown in

Table 4. Mechanism test.

Variable	(1) Res	(2) MisK	(3) Res
LnGti	0.495 ***	−0.049 *	0.406 **
	(0.180)	(0.025)	(0.169)
MisK			−1.816 ***
			(0.576)
AS	−0.051	−0.036 ***	−0.116 *
	(0.053)	(0.006)	(0.066)
IP	7.410 ***	0.057	7.513 ***
	(2.988)	(0.309)	(2.723)
ALR	−1.501	−0.682 **	−2.740
	(2.037)	(0.283)	(2.037)
LFS	0.464 ***	−0.033 **	0.404 ***
	(0.093)	(0.015)	(0.084)
Cons	−2.964 *	1.165 ***	−0.848 ***
	(1.775)	(0.299)	(1.741)
Area FE	Y	Y	Y
Year FE	Y	Y	Y
N	360	360	360
R-squared	0.776	0.915	0.786

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

. After adding the control variables and controlling for area-fixed effects and year-fixed effects, the coefficient estimate of $Ln{Gti}_{it}$ in column (2) is significantly negative, indicating that green technological innovation curbs capital mismatch and increases the capital allocation efficiency. The coefficient estimate of capital mismatch in column (3) is significantly negative, indicating that an increase in capital mismatch inhibits the enhancement of export resilience. The improvement in capital allocation efficiency helps to improve manufacturing-export resilience. The coefficient estimate of $Ln{Gti}_{it}$ in column (3) is less than 0.495, which proves that there is a partial mediating effect of capital allocation efficiency. In summary, capital allocation efficiency plays a positive mediating role in the impact of green technology innovation on manufacturing-export resilience.

4.3. Heterogeneity Analysis

4.3.1. Differences in Labour-Mismatch Levels

Based on the methodology of calculating the capital-mismatch index above, this paper can similarly calculate the labour-mismatch index. (The coefficient of absolute distortion of labour factor prices ${\gamma }_{Li}=\frac{1}{1-{\tau }_{Li}}$, where ${\tau }_{Li}$ is the labour-mismatch index. The coefficient of relative distortion of labour prices is used in the actual measurement ${\widehat{\gamma }}_{Li}=\left(\frac{L_i}{L}\right)/\left(\frac{s_i{\beta }_{Li}}{{\beta }_L}\right)$, where $\frac{L_i}{L}$ denotes the actual proportion of regional use of labour $L_i$ to the total labour force $L$ of the whole economy, ${\beta }_L={\sum }_i^N{s}_i{\beta }_{Li}$ denotes the regional output-weighted value of the contribution of labour, and $\frac{s_i{\beta }_{Li}}{{\beta }_L}$ denotes the theoretical proportion of regional use of labour when there exists an efficient allocation of labour, and the rest of the variables are interpreted and computed in the same way as above). First, the average value of the labour-mismatch index for each province from 2009 to 2020 is calculated; then, it is sorted to find the median. In this paper, the areas higher than the median are defined as high labour-mismatch areas, and the converse as low labour-mismatch areas. According to the labour-mismatch level of the sub-sample regression, the regression results are shown in

Table 5. Heterogeneity test based on labour-mismatch levels.

Variable	High Labour-Mismatch Region			Low Labour-Mismatch Region
	(1) Res	(2) MisK	(3) Res	(4) Res	(5) MisK	(6) Res
LnGti	1.068 ***	−0.162 ***	0.907 ***	0.143	0.033 **	0.233
	(0.235)	(0.041)	(0.257)	(0.205)	(0.015)	(0.227)
MisK			−0.995 **			−2.738
			(0.471)			(2.137)
AS	−0.085	−0.044 ***	−0.129 **	−0.032	−0.017 **	−0.079
	(0.054)	(0.009)	(0.061)	(0.125)	(0.007)	(0.147)
IP	3.219	2.486 ***	5.692	8.483 *	0.527 *	7.042 *
	(4.104)	(0.756)	(4.688)	(4.833)	(0.298)	(3.832)
ALR	−1.586	−0.248	−1.832	−5.172 *	0.013	−5.137
	(3.289)	(0.470)	(3.291)	(3.256)	(0.201)	(3.282)
LFS	0.409 ***	−0.047	0.362 ***	0.301 **	−0.001	0.298 **
	(0.114)	(0.034)	(0.102)	(0.131)	(0.011)	(0.128)
Cons	−6.345 **	1.641 ***	−4.712 *	1.902	0.096	2.164
	(2.476)	(0.442)	(2.684)	(2.564)	(0.168)	(2.510)
Area FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
N	180	252	252	108	108	108
R-squared	0.773	0.936	0.780	0.791	0.865	0.797

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

. From the regression results in Table 5, it can be seen that in the high labour-mismatch region, green technology innovation can significantly enhance the manufacturing-export resilience, and the capital allocation efficiency plays a positive intermediary effect, but in the low-labour-mismatch region, the indicators of green technology innovation and capital allocation efficiency are no longer significant.

4.3.2. Geographical Differences

Because China’s regional development is unbalanced, and there are differences in the factor resource endowment, institutional and legal environment, and marketisation level, this paper divides the research sample into the eastern–central region (including Beijing, Tianjin, Hebei and Shanxi Provinces, Inner Mongolia Autonomous Region, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangxi, and Hainan Provinces) and the western region for regression, to further examine the impact of geographic location differences on the relationship between green technological innovation, capital allocation, and manufacturing-export resilience. The results of the regression are shown in

Table 6. Heterogeneity test based on geographical differences.

Variable	Eastern–Central Region			Western Region
	(1) Res	(2) MisK	(3) Res	(4) Res	(5) MisK	(6) Res
LnGti	0.971 ***	−0.122 ***	0.908 ***	0.023	0.019	0.118
	(0.180)	(0.036)	(0.187)	(0.219)	(0.019)	(0.260)
MisK			−0.511 *			−5.019
			(0.274)			(3.137)
AS	−0.098 **	−0.044 ***	−0.120 ***	1.471 ***	−0.69 **	1.126 **
	(0.041)	(0.008)	(0.043)	(0.326)	(0.028)	(0.464)
IP	5.355 **	2.053 ***	6.403 **	1.571	0.077	1.957
	(2.494)	(0.771)	(2.731)	(4.416)	(0.225)	(4.264)
ALR	−0.791	−0.986 ***	−1.295	−10.993 *	0.582	−8.069
	(1.634)	(0.339)	(1.660)	(6.164)	(0.412)	(5.621)
LFS	0.373 ***	−0.030 *	0.358 ***	1.400	−0.065	1.072
	(0.083)	(0.016)	(0.082)	(0.997)	(0.044)	(0.818)
Cons	−6.418 ***	1.746 ***	−5.526 ***	3.735	−0.011	3.680
	(1.469)	(0.344)	(1.566)	(4.531)	(0.345)	(4.538)
Area FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
N	252	252	252	108	108	108
R-squared	0.786	0.933	0.788	0.836	0.881	0.848

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

. Green technological innovation in the eastern–central region can significantly enhance manufacturing-export resilience, and capital allocation efficiency plays a positive mediating effect, but in the western region, it is no longer significant.

4.4. Robustness Test

4.4.1. Replacement of the Manufacturing-Export Resilience Measurement Methodology

This paper conducts robustness tests by changing the measure of manufacturing-export resilience. The specific methods are as follows:

Referring to the studies of Bergeijk et al. [41] and Martin et al. [16], the sensitivity-analysis method with a single indicator is used to calculate export resilience by comparing the change in the real export trade of the manufacturing industry with the expected export-trade change. The expected change in manufacturing exports of the region is calculated as:

$${\left({∆C}_i^{t+k}\right)}^{expect}=C_i^t·g_N^{t+k}$$

(10)

In Equation (10), ${\left({∆C}_i^{t+k}\right)}^{expect}$ denotes the expected change in manufacturing exports of region i in the period [t, t + k]; $C_i^t$ denotes the number of manufacturing exports of region i in period t; and $g_N^{t+k}$ denotes the rate of change in the number of manufacturing exports of the whole country in the period [t, t + k].

The regional manufacturing-export resilience is calculated by the formula:

$${Res}_{it}=\frac{{∆C}_i^t-{\left({∆C}_i^t\right)}^{expect}}{\left|{\left({∆C}_i^t\right)}^{expect}\right|}$$

(11)

In Equation (11), ${∆C}_i^t$ represents the change in actual manufacturing export of region $i$ in period $t$, and ${\left({∆C}_i^t\right)}^{expect}$ represents the expected change in manufacturing export of region $i$ in period $t$. The value of ${Res}_{it}$ represents the manufacturing-export resilience of region $i$ in year $t$. The larger the value is, the stronger the export resilience is, and vice versa; the weaker the manufacturing-export resilience, the smaller the value of ${Res}_{it}$.

Table 7. Robustness test.

Variable	(1) Res	(2) MisK	(3) Res
LnGti	0.568 ***	−0.049 *	0.493 **
	(0.218)	(0.025)	(0.207)
MisK			−1.549 **
			(0.762)
AS	−0.121	−0.036 ***	−0.176 **
	(0.075)	(0.006)	(0.089)
IP	8.815 ***	0.057	8.903 ***
	(2.988)	(0.309)	(3.060)
ALR	5.826 **	−0.682 **	4.770 *
	(2.862)	(0.283)	(2.800)
LFS	0.939 ***	−0.033 **	0.887 ***
	(0.153)	(0.015)	(0.147)
Cons	−9.417 ***	1.165 ***	−7.613 ***
	(2.437)	(0.299)	(2.269)
Area FE	Y	Y	Y
Year FE	Y	Y	Y
N	360	360	360
R-squared	0.832	0.915	0.835

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

shows that the coefficient estimates of green technology innovation pass the significance test, and the sign of the coefficients does not change. These further demonstrate that green technology innovation can promote manufacturing-export resilience, and that capital allocation efficiency plays a mediating effect in green technology innovation to enhance the manufacturing-export resilience.

4.4.2. Replacement of the Independent Variable

Although the estimation bias caused by omitted variables is minimised by constructing a two-way fixed effect model in this paper, the correlation between green technology innovation and the random disturbance term and the reverse causality between green technology innovation and export resilience may still lead to endogeneity, causing estimation bias. In this paper, we will adopt the regression using the lagged first-order variable ($L.LnGti$) of green technology innovation as a proxy variable.

Table 8. Endogeneity test.

Variable	(1) Res	(2) MisK	(3) Res
L.LnGti	0.733 ***	−0.050 **	0.664 ***
	(0.202)	(0.022)	(0.194)
MisK			−1.366 **
			(0.614)
AS	−0.050	−0.034 ***	−0.097
	(0.059)	(0.007)	(0.072)
IP	6.930 ***	−0.109	6.781 ***
	(2.514)	(0.248)	(2.496)
ALR	−2.179	−0.657 **	−3.076
	(2.141)	(0.289)	(2.173)
LFS	0.376 ***	−0.024	0.343 ***
	(0.093)	(0.015)	(0.086)
Cons	−3.506 *	1.142 ***	−1.945
	(2.025)	(0.286)	(2.017)
Area FE	Y	Y	Y
Year FE	Y	Y	Y
N	360	360	360
R-squared	0.813	0.933	0.817

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

illustrates that the coefficient estimates of green technology innovation all pass the significance test. In addition, the sign of the coefficients has not changed, further arguing that green technology innovation can promote manufacturing-export resilience, and that capital allocation efficiency plays a mediating effect in green technology innovation to enhance the manufacturing-export resilience.

5. Further analysis

5.1. Moderating Effects of Intellectual-Property Protection

This paper follows the methodology proposed by [40] to construct a moderating-effect model for empirical testing, and the constructed econometric model is as follows:

$${Res}_{it}=e_0+e_{1}Ln{Gti}_{it}+e_2{IPP}_{it}+e_{3}ln{Gti}_{it}\ast {IPP}_{it}+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(12)

In Equation (12), the moderating variable is ${IPP}_{it}$, $Ln{Gti}_{it}\ast {IPP}_{it}$ is the cross-multiplier of green technological innovation and intellectual-property protection, and the rest of the settings are consistent with the benchmark model above. If the estimated coefficient of the cross-multiplier of green technological innovation and intellectual-property protection, $e_3$, is less than zero, it indicates that intellectual-property protection plays a negative moderating role in the impact of green technological innovation in enhancing manufacturing-export resilience. The regression results are shown in

Table 9. Moderating-effects test.

Variable	(1) Res	(2) Res	(3) Res
LnGti	0.495 ***	0.483 ***	0.434 **
	(0.180)	(0.179)	(0.172)
IPP		0.003	0.004 *
		(0.002)	(0.002)
LnGti × IPP			−0.010 **
			(0.005)
AS	−0.051	−0.061	−0.071
	(0.053)	(0.056)	(0.055)
IP	7.410 ***	7.335 ***	7.504 ***
	(2.988)	(2.524)	(2.504)
ALR	−1.501	−1.464	−0.742
	(2.037)	(2.038)	(2.157)
LFS	0.464 ***	0.466 ***	0.465 ***
	(0.093)	(0.093)	(0.089)
Cons	−2.964 *	−2.997 *	−0.391
	(1.775)	(1.778)	(1.517)
Area FE	Y	Y	Y
Year FE	Y	Y	Y
N	360	360	360
R-squared	0.776	0.777	0.781

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively. Interaction terms were introduced into the model to avoid the problem of covariance, and the variables were first standardised.

, and the coefficient estimate of $ln{GTI}_{it}\ast {IPP}_{it}$ in column (3) is significantly negative, which proves the negative moderating role played by intellectual-property protection.

5.2. Threshold Effects of Government Intervention

To further explore whether the impact of green technology innovation on manufacturing-export resilience has a non-linear characteristic with government intervention as the threshold, we draw on the panel threshold-regression model proposed by Hansen [42] to construct a threshold-regression model to test the threshold effect of government intervention, e.g., with single and double thresholds. The specific threshold regression model is as follows:

$${Res}_{it}={\phi }_0+{\phi }_{1}Ln{Gti}_{it}\ast I\left({Gi}_{it}\le {\varphi }_1\right)+{\phi }_{2}Ln{Gti}_{it}\ast I\left({Gi}_{it}>{\varphi }_1\right)+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(13)

$${Res}_{it}={\phi }_0+{\phi }_{1}Ln{Gti}_{it}\ast I\left({Gi}_{it}\le {\varphi }_1\right)+{\phi }_{2}Ln{Gti}_{it}\ast I\left({\varphi }_1<{Gi}_{it}\le {\varphi }_2\right)+{\phi }_{3}ln{Gti}_{it}\ast I\left({Gi}_{it}>{\varphi }_2\right)+\phi Cons+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(14)

In Equations (13) and (14), ${Gi}_{it}$ denotes the threshold variable, $I$ is the threshold value to be tested, and $I\left(\cdot \right)$ is the indicator-line function of the threshold model, with $I$ being 1 if the parentheses are true and 0 otherwise, and the rest of the settings are consistent with the benchmark model above.

Table 10. Results of the threshold-model test.

Threshold Variable	Model	Threshold Value	F-Value	p-Value	Critical Value
					1%	5%	10%
GI	Single threshold	0.2789 **	37.35	0.049	55.174	37.116	30.363
	Double threshold	0.2083	18.77	0.263	63.954	38.797	29.115

Notes: The threshold value, F-value, p-value, and critical value are obtained by repeating the Bootstrap analysis 3000 times; the fixed-effect model is adopted; ** indicates significance at the 5% level.

demonstrates the results of the test with government intervention as the threshold variable. Controlling for area and year fixed effects, the threshold test results show that the single-threshold model passes the test at the 5% significance level. The regression results in

Table 11. Results of the threshold regression.

Variable	(1) Res	(2) Res
Threshold value	GI $\le$ 0.2789	GI $>$ 0.2789
LnGti	0.521 **	0.238
	(0.205)	(0.227)
AS	−0.129
	(0.100)
IP	5.647
	(3.364)
ALR	0.221
	(2.720)
LFS	0.523 ***
	(0.166)
Cons	−4.330 *
	(2.432)
Area FE	Y
Year FE	Y
N	360
R-squared	0.501

Notes: Numbers in parentheses in the table are robust standard errors of the coefficient estimates; *, **, and *** represent significance at the 10%, 5%, and 1% significance levels, respectively.

indicate that when government control is not greater than the threshold value of 0.2789, the coefficient estimate of the impact of green technological innovation on manufacturing-export resilience is 0.521, which is significant at the 5% significance level, indicating that when government intervention is lower than the threshold value, the impact of green technological innovation on manufacturing-export resilience has a significant positive effect. When government intervention is greater than the threshold value, the coefficient estimate is no longer significant, indicating that green technological innovation does not affect the manufacturing-export resilience.

6. Discussion

Figure 1 shows the results of this paper, which visually illustrate the interaction between green technology innovation, manufacturing-export resilience, capital allocation efficiency, intellectual-property protection, and government intervention. Green technology innovation promotes the improvement of capital allocation efficiency, which, in turn, promotes the improvement of manufacturing-export resilience, while intellectual-property protection inhibits the improvement of green technology innovation on manufacturing-export resilience, and government intervention exerts a threshold effect on the impact of green technology innovation on manufacturing-export resilience.

6.1. Mediating Effect of Capital Allocation Efficiency

Table 12. The result of green technology innovation on manufacturing-export resilience.

	Manufacturing-Export Resilience	Capital Allocation Efficiency	Manufacturing-Export Resilience
Green Technology Innovation	promotion; significant	promotion; significant
Capital Allocation Efficiency			promotion; significant

shows that green technology innovation can significantly enhance China’s manufacturing-export resilience. Green technology innovation not only helps to improve the industry’s production efficiency and reduce the production costs of exporters [43,44] but also reduces the burden of environmental penalties and non-compliance costs on exporters [45]. At the same time, green technology innovation can also facilitate the production of differentiated and premium products [46], giving manufacturing exporters access to key resources to reshape their international competitive advantage. In addition, as the public’s environmental awareness increases, consumers are increasingly attracted to green products and show green purchasing behaviour [47], pushing governments to change their trade management policies to be more biased towards importing green products. Green technological innovation enables enterprises to signal to the public that they are concerned about environmental protection [48], which not only better meets the green consumption demand in the international market but also better responds to the constraints of international environmental barriers. Thus, in the context of the complex and changing global trade environment and the increasingly prominent contradiction between the domestic production mode and the resource environment, green technology innovation not only provides an effective path for China’s manufacturing industry to cope with the production contradiction but also provides more certainty for manufacturing exports to face sudden external shocks and long-term slow perturbations and to realise sustainable development.

The technological change brought about by green technological innovation can break the constraints imposed by the traditional factor market on production, promote enterprises to recombine and transform the factored resources invested in production, and build a new production-factor input system and allocation method. Table 12 also shows that green technology innovation can improve the capital allocation efficiency, which then enhances the manufacturing-export resilience. To be specific, green technology innovation will not only change the original production of the capital-factor input ratio but will also allow the capital factor to develop more ways of configuration and alleviate the contradiction between the supply and demand of capital factors, improving the degree of capital mismatch and the capital allocation efficiency. The improvement in capital allocation efficiency can make the internal resources of the enterprise more reasonably configured, which ensures that the production link in the manufacturing enterprise receives the due input and support, enhances the level of collaboration between the elements and the return on capital, and further promotes the optimisation and upgrading of the industrial structure to improve the quality of the manufacturing industry exports and the export growth rate. At the same time, regions and industries with higher capital allocation efficiency are more likely to obtain financial support, thus expanding their output and export scale and increasing export growth.

In the heterogeneity analysis, we find that there are different impacts of green technology innovation on manufacturing-export resilience in different regions, and the mediating effect of the capital allocation efficiency also changes. Green technology innovation and capital-allocation-efficiency indicators are no longer significant in low labour-mismatch areas compared with high labour-mismatch areas. Compared with the capital-factor market, the labour-marketisation process is lagging. When green technological innovation reallocates the input ratios of factors of production, the capital factor can improve the degree of factor mismatch promptly according to changes in market prices, but because the quality of the labour force cannot match green technological innovation in a short period and because of the existence of barriers to mobility in the existing labour market, the labour factor cannot be adjusted promptly, which may exacerbate the degree of mismatch in the labour factor. An increase in the degree of labour-factor mismatch can cause a decline in labour productivity and factor allocation efficiency in the short term, thus affecting output and export size. However, in the medium and long term, if the labour-factor mismatch can stimulate enterprises to increase R&D investment and better exploit the scale effect, it will instead help to improve product quality and the export scale. In other words, labour mismatch can, to a certain extent, provide a new operating space for the rational input and flow of capital. Green technology innovation and capital-allocation-efficiency indicators are also no longer significant in the western region compared with the eastern–central region. The possible reason is that the economic development of the western region is relatively backward, the local resource factors are not rich enough, and the institutional and legal environment and infrastructure are not perfect, which not only restricts the level of green technological innovation but also leads to the marginal output value of investment is not high, the capital factor cannot give full play to its due role, resulting in capital allocation efficiency cannot play the role of the channel.

6.2. Moderating Effects of Intellectual-Property Protection and Threshold Effects of Government Intervention

In addition, we further explore the role played by intellectual-property protection and government intervention in the impact of green technology innovation on manufacturing-export resilience. Most studies agree that strengthening intellectual-property protection can safeguard the legitimate rights and interests of innovative enterprises and provide a guarantee for the emergence of new technologies and products, which can promote enterprises to carry out innovative activities [49,50,51]. However, intellectual-property protection can essentially lead to the monopolisation of intellectual-property rights, which is not conducive to the diffusion of green technologies. The prolonged lack of dissemination of green technologies not only discourages others from inventing but also can no longer provide incentives for inventors of green technologies to encourage them to continue to innovate. Therefore, some scholars have pointed out that intellectual-property protection may hinder knowledge flow and technological learning, which hurts innovation [52,53]. There are also studies based on the national level, which found that strengthening intellectual-property protection promotes innovation in developed countries, but is detrimental to innovation in developing countries [49,54,55]. Considering that China is still a developing country, compared with developed countries, the correlation between various industries is limited, and a complete industrial chain that can be extended cannot be formed. Therefore, it is not enough for China’s manufacturing industry to stimulate its green innovation capacity through its innovation system, and exogenous technology transfer is needed as a catalyst for the formation of learning effects and the enhancement of green innovation capacity. However, increasing the strength of domestic intellectual-property protection will increase the cost of imitation, hinder manufacturing enterprises from learning and absorbing international advanced technology, and hurt their green technological innovation activities, which is not conducive to the enhancement of manufacturing-export resilience. Manufacturing firms are key players in the market and are the main actors in innovation activities. The government, on the other hand, is an important agent of the market, and its intervention influences the decision of manufacturing firms to implement green technological innovation. It has been shown in the literature that government intervention has two sides to the innovative activities of manufacturing firms. Government intervention can be effective in providing resource support to manufacturing firms to alleviate cost constraints and can ameliorate the externalities faced by firms by providing effective incentives and compensation to reduce the loss of technological innovation levels caused by externalities [56]. However, government intervention also undermines the free market and distorts the market mechanism, thus reducing the incentive for manufacturing enterprises to innovate [57]. Considering the two sides of government intervention affecting innovation activities, we find reasonable government intervention is conducive to green technological innovation to enhance manufacturing-export resilience, and excessive government intervention will, on the contrary, adversely affect the relationship between the two.

7. Conclusions

Taking the 2008 global financial crisis as an external shock, this paper empirically analyses the impact of green technology innovation on manufacturing-export resilience in the face of external shocks as well as the mechanism of its role. It not only has important reference significance and practical value for the implementation of green development strategy and intellectual-property protection policy in China but also provides an empirical basis for most developing countries to enhance manufacturing-export resilience. Our main conclusions are as follows:

Green technology innovation can significantly enhance manufacturing-export resilience, so we should actively improve the level of green technology innovation and cultivate new momentum for the development of manufacturing exports. The government should make special allocations to increase the investment in environmental protection research and development and establish green technology research and development centres mainly by enterprises and in cooperation with enterprises, universities and scientific research institutes. Developing countries should introduce advanced green technologies and equipment from developed countries, improve and update energy-consuming and highly polluting production equipment and processes, strengthen international energy technology cooperation and actively participate in international cooperation on clean development mechanism projects.

Green technology innovation can improve capital allocation efficiency, thereby enhancing manufacturing-export resilience. From the perspective of manufacturing enterprises, it is necessary to enhance the attention to the capital allocation situation, avoid capital mismatch, adjust capital allocation in time, and put the results of green technology innovation into practice. Manufacturing enterprises also need to actively build modern enterprise systems, improve corporate governance and supervision systems, improve the quality of capital investment, promote deeper integration with modern service industries, and reduce average operating costs. From the government’s perspective, it is necessary to rationalise intergovernmental financial relations, promote the modernisation of economic governance capacity, implement a coordinated development strategy, and focus on the coordinated development of the manufacturing industry and a reasonable division of labour.

In China’s high labour-mismatch regions, the enhancement of green technology innovation on manufacturing-export resilience is more significant compared with the low labour-mismatch regions, and the enhancement effect in the eastern–central region is also significantly higher than that in the western region. Labour factors still need to adapt to changes in technological progress and capital-factor allocation to achieve the appropriate ratio of factor allocation and to maximise the stimulation of the efficiency of capital allocation, so we need to further promote the reform of the labour-factor market to improve the degree of labour-factor mismatch. At the same time, a sound capital-factor market reduces the threshold of entry into and exit from the industry and promotes a more reasonable flow of capital, which weakens the inhibition of labour-factor mismatch on manufacturing exports, maximises the release of enterprise production potential, and optimises the way resources are allocated. The government should also create a good institutional and legal environment and perfect market mechanism according to the economic level of each region, with the aim of creating favourable conditions to ensure the effective implementation of green technology innovation activities and to effectively play a role in promoting the manufacturing-industry export trade.

Strengthening IPP by the government inhibits the enhancement of green technology innovation on manufacturing-export resilience. We should identify the differences with developed countries, formulate intellectual-property protection policies as appropriate, and establish a realistic intellectual-property protection system. Local governments should strengthen legislation in the field of IPP, establish a comprehensive legal system for IPP, accelerate the convergence with the international IPP system, and, at the same time, strike a proper balance between individual and social interests and dynamically adjust IPPR-related enforcement procedures or measures. When government intervention exceeds a certain limit, the impact of green technology innovation on manufacturing-export resilience will no longer be significant. Therefore, we should deal with the relationship between the market and the government, regulate the government’s administrative power, clarify the boundaries of government functions, adhere to the decisive position of the market in resource allocation, maximise the release of the potential of green technology innovation, and promote the development of manufacturing exports.

The research in this paper also has some limitations. This paper focuses on the changes in manufacturing exports in different regions of China and discusses the impact of green technology innovation on manufacturing-export resilience in each region. Considering the rich variety of manufacturing industries, further research can focus on analysing the impact of green technology innovation on the export resilience of each manufacturing sub-industry. In addition, this paper selects the output perspective to measure green technological innovation, but its scope is more complex, and further research should also give full consideration to the input perspective (such as R&D expenditure) to enrich the measurement method of green technology innovation.