Environmental Regulations and Urban Technological Innovation: China’s Two Control Zones Policy as Evidence

Abstract

In China, the Two Control Zones (TCZ) policy is an important practice in promoting sustainable development. This study aimed to investigate whether the TCZ policy promotes urban technological innovation. In this research, a DID model was built with the TCZ policy as an exogenous shock using panel data covering over 223 Chinese cities from 1995 to 2003. The empirical outcomes indicate that the TCZ policy is beneficial for augmenting urban technological innovation. Through heterogeneity analysis, it is further observed that the promotional effects of technological innovation resulting from the TCZ policy are highest in the Western region, followed by the Eastern region, and finally the Central region, and higher in cities with a higher proportion of state-owned enterprises than in cities with a higher proportion of non-state-owned enterprises. The findings of this paper align with Porter’s hypothesis and hold significant importance for other countries in devising appropriate environmental regulation policies to facilitate sustainable development.

1. Introduction

The issue of striking a balance between economic development and environmental preservation has gained international attention as environmental conservation becomes more and more important for all nations to address. Especially for developing countries that are accelerating their industrialization, this challenge is even more serious. Studies have shown that rapid economic growth and environmental protection are always incompatible, and the experience of developed countries has also confirmed this [1].

As a representative of a developing country that has undergone rapid industrialization, China has inevitably encountered environmental pollution problems associated with industrialization. As the world’s manufacturing plant, China has many highly polluting industries such as steel, power generation, and coal. The industrial exhaust emissions from these highly polluting enterprises are the main cause of air pollution in China. The impact of air pollution in China is alarming. The World Bank reports that air pollution contributes to an astounding 350,000 to 400,000 deaths of minors each year in China [2]. Since sulfur dioxide (SO₂) is one of the primary air pollutants, its large emissions have resulted in acidic precipitation and severe air pollution. According to the 1995 China Environmental Status Report, China’s SO₂ emissions in that year were 13.96 million tons and were on an upward trend, increasing by 4.1% over the previous year. In Central China, which suffered the most severe acid rain pollution caused by SO₂ emissions, the annual average pH value of acid rain in its central area was below 4.0, and the frequency of acid rain was as high as 80% or more. Given the external impacts of production and the finite nature of conventional energy sources, the adoption of environmental regulation policies by the government is an essential tool to control environmental pollution [3]. The Chinese government implemented the TCZ policy in January 1998 to curb the emissions of SO₂. It has been demonstrated that a variety of environmental protection measures, such as the creation of eco-industrial parks, environmental legislation, or environmental protection fees, can aid in the reduction in environmental pollution [4,5,6,7].

Beyond its primary function of mitigating environmental pollution, environmental regulation policies invariably influence other aspects of the economy, such as technological innovation. Schumpeter’s innovation theory explored the importance of technological innovation in a pioneering way, helping to provide a dynamic perspective on the role of technological innovation for economic development rather than economic growth. The impact of environmental control regulations on technological innovation is the main subject of this study. Based on compliance cost theory, some scholars propose that environmental regulation increases firms’ production costs, which imposes more constraints on firms’ production decisions [8,9]. Furthermore, environmental regulation increases the uncertainty faced by firms. Such uncertainties may potentially hinder investment decisions and adversely affect the development of new products and production processes [10,11]. However, the perspective of compliance cost theory is relatively static. Porter and other academics note that a dynamic viewpoint is necessary to analyze the connection between enterprise technical innovation, company competitiveness, and environmental regulation. The innovation compensation theory suggests that when faced with rising costs caused by environmental regulations, enterprises will compensate for these costs through innovation [12]. Porter’s hypothesis argues that appropriate environmental regulation stimulates enterprises to engage in technological improvement and R&D innovation activities [13]. In fact, for enterprises lacking the capacity to engage in technological innovation, environmental regulation eliminates these highly polluting enterprises [14,15], which is actually beneficial to the overall technological upgrading of the industry. A case study conducted in China has confirmed the beneficial effects of environmental regulating measures on enterprise technological innovation [16].

China’s TCZ policy is a model program for preventing and controlling air pollution. The effects of the TCZ policy have been the subject of several studies. The research results indicate that the TCZ policy can eliminate low-technology businesses in polluting industries [17], encourage the modernization of the urban industrial structure [18,19], enhance the caliber of export goods [20,21], raise the productivity of businesses that use green total factors [22], decrease foreign direct investment (FDI) that contributes to the pollution haven effect [23], and lower infant mortality in relevant cities [24]. Nonetheless, there are still few pertinent studies examining how the TCZ policy affects technological innovation in cities. Does the TCZ policy’s implementation encourage technological innovation in urban areas? This essay seeks to provide an answer to that query. Therefore, we use China’s TCZ policy as a kind of quasi-natural experiment to determine what kind of effects this policy has on urban technological innovation.

Furthermore, considering that different characteristics of the location where the policy is applied can lead to different effects of policy implementation [25,26], it has also been argued that this difference in the characteristics of the location where the policy is implemented is responsible for the contrary findings in the literature related to compliance cost theory [27]. Using China as an example, economic development, the amount of energy consumed from fossil fuels, the percentage of secondary sectors, and the population distribution vary significantly between regions. These differences lead to different intensities of air pollutant emissions across regions [28,29], different spatial distribution patterns of air pollution across regions [30], and differences in the cost of emission reduction [31]. In addition, the implementation of environmental regulation policies varies among regions, which also leads to differences in the final effectiveness of policies [32]. Thus, this study investigates further if diverse impacts on the implementation effects of the TCZ policy can be attributed to urban disparities in area characteristics and company ownership.

Compared to prior works, this paper makes the following advances and contributions. (1) Taking the TCZ policy in China as an example, this paper empirically analyzes the impact of this environmental regulation policy on urban technological innovation. This study’s findings contribute to the current literature in the topic. (2) The findings of the heterogeneity analysis are of great practical importance. They can help governments identify priorities when formulating environmental regulation policies.

The rest of this paper is organized as follows. The study’s policy backdrop is introduced in the second section, which also includes the data and methodology used for the empirical research. The examination of the empirical findings is presented in the third section. The policy implications of our findings are covered in the fourth section. The fifth part is the conclusion.

2. Materials and Methods

2.1. Policy Background

As industrialization and urbanization continue to advance, SO₂ has emerged as a primary contributor to air pollution in China. China is currently the world’s third-largest emitter of SO₂ [33]. In Southern China, the increase in SO₂ concentration is likely to cause acid rain pollution. In 1997, the average annual precipitation pH value in over 40% of the nation’s cities was less than 5.6. Reducing SO₂ emissions and dealing with acid rain has been a national concern since 1990. In order to effectively control SO₂ pollution and acid rain, the central government imposed SO₂ emission restrictions, which were first established in The Five-Year Plan in 1996. In 1995, the Standing Committee of the National People’s Congress revised the Law of the People’s Republic of China on the Prevention and Control of Atmospheric Pollution and suggested establishing the TCZ to limit overall SO₂ emissions.

In January 1998, the State Council formally approved the TCZ policy proposed by the State Bureau of Environmental Protection. The TCZ refer to the SO₂ pollution control zones and acid rain control zones. Severe SO₂ pollution or acid rain pollution may occur or has occurred in these areas. The TCZ encompass 175 cities throughout 27 provinces in China, where the SO₂ pollution control zones mainly cover cities in Northern China and the acid rain control zones mainly cover cities in Southern China. This coverage is wide, for these cities accounted for 60% of China’s SO₂ emissions, 62.4% of its GDP, and 40.6% of the country’s population [34].

Specifically, the following criteria must be met in order to define the boundaries of SO₂ pollution control zones: (1) recent years have seen an increase in the annual average concentration of SO₂ in ambient air over the national secondary standard; (2) the daily average concentration of SO₂ exceeds the national tertiary standard; (3) large SO₂ emissions; and (4) cities are the basic control units. The following are the fundamental requirements for drawing the boundaries of acid rain control zones: (1) precipitation pH value ≤ 4.5 is currently being monitored; (2) sulfur deposition exceeds the critical load; and (3) areas with large SO₂ emissions. Southern cities where SO₂ pollution as well as acid rain are both severe are delineated as acid rain control zones.

2.2. Methodology

To examine the impact of environmental regulation policies on technological innovation, the empirical evaluation in this work is conducted using the difference-in-difference (DID) approach. The DID approach, which originated in the field of economics in the late 1970s, is an econometric technique based on experimental or quasi-experimental design [35], which efficiently addresses the endogeneity issue in regressions and bases causal effects on the concept of randomized experiments [36]. The DID method helps the researcher divide the sample into experimental and control groups based on exogenous shocks. According to this classification, the experimental group will be subjected to an exogenous shock, while the control group will remain unexposed. It allows the researcher to empirically test the impact of this exogenous shock using the DID method. The TCZ policy can be used as an exogenous shock in this research to investigate how it affects urban technological innovation using the DID approach.

In this paper, a DID model is developed of a quasi-natural experiment with the exogenous shock of the 1998 TCZ policy. Applying the DID approach is predicated on satisfying the randomness assumption. Otherwise, the problem of sample selection bias will occur. For this paper, whether each city is designated as the TCZ is determined randomly, because the TCZ policy is primarily intended to prevent acid rain and SO₂ pollution from getting worse, not to advance technological advancement. According to related research, the TCZ policy’s exogeneity is comparatively strong [20,23]. Therefore, the sample selection in this paper satisfies the randomness assumption. This study uses the DID approach to compare and analyze the differences in urban technological innovation before and after the implementation of the policy, between cities designated as TCZ and cities not designated as such. The model is set as follows:

$$\ln paten{t}_{it}={\beta }_0+{\beta }_{1}trea{t}_i·pos{t}_t+X_{it}·\gamma +{\theta }_i+{\theta }_t+{\epsilon }_{it}$$

(1)

where i is the city and t is the year. θi and θt represent city fixed effects and time fixed effects. ${\mathsf{\epsilon }}_{\mathrm{it}}$ is a random error term. The explanatory variable is the logarithm of patent numbers. The variable $trea{t}_i$ is a dummy variable for the selected city, which indicates whether the city is delineated as a TCZ. $pos{t}_t$ is a time dummy variable to indicate whether the current year is before or after the implementation of the TCZ policy. Therefore, the regression coefficient of ${\mathrm{treat}}_{\mathrm{i}}·{\mathrm{post}}_{\mathrm{t}}$ can represent the impact of TCZ policy on technological innovation.

In order to accurately estimate the regression coefficient of $trea{t}_i·pos{t}_t$, it is necessary to ensure the exogeneity of $trea{t}_i$ and the randomness of $pos{t}_t$, because whether a city is delineated as a TCZ may be related to urban economic development and other factors. To ensure convincing regression results in this study, this paper further controls city fixed effects and time fixed effects, and sets a series of control variables. Control variables $X_{it}$ include other factors that affect technological innovation, such as urbanization, industrial structure, education, R&D investment, and gross regional product.

2.3. Variables and Data

Due to the fact that the establishment of a TCZ is based on the city as a unit, this paper mainly uses relevant city-level panel data for empirical analysis. The original data are from the China City Statistical Yearbook. The yearbook is authored by the Department of Urban Social and Economic Surveys of the National Bureau of Statistics and is published annually by the China Statistics Press. City-level panel data from 1995 to 2003 are selected for empirical analysis in this paper. Since the TCZ policy was implemented in 1998, the data for this period (1995–2003) can meet the needs for empirical analysis using the DID method. China enacted the Cleaner Production Promotion Law in 2003, which was accompanied by the Interim Measures on Cleaner Production Audit in 2004. Therefore, we use data up to 2003 to avoid shocks to our study from the enactment of these new regulations. The administrative code of China was changed several times during the data period of this paper, so we use the latest version of the administrative code in 2002 to unify the data. By organizing and cleaning the data through the requirements of the later analysis, we finally obtained panel data for 223 cities. According to government documents, 158 cities, 13 regions, and 4 municipalities directly under the central government are included in the scope of China’s TCZ policy implementation. By matching the panel data above with cities in the TCZ, we finally identified the panel data of 135 cities as the policy implementation group.

Dependent variable: In this work, the number of patents awarded in cities is logarithmically valued as $lnpatent$ to measure technological innovation. Compared with indicators such as research inputs and the number of personnel engaged in scientific and technological activities in industrial enterprises above a designated size, the number of patents granted, as an important representation of innovation output, can better measure the technological innovation capacity of enterprises and avoid neglecting external technological influences or neglecting innovation inputs that are not included in statistics.

Key explanatory variables: This paper’s interaction term $trea{t}_i·pos{t}_t$ is the key explanatory variable. $trea{t}_i$ is included to distinguish if the city is designated as one of the TCZs, where $trea{t}_i$= 1 means the city in the TCZ belongs to the experimental group, and $trea{t}_i$= 0 means the city not in the TCZ belongs to the control group. $pos{t}_t$ is included to distinguish between the years prior to and following the TCZ policy, where $pos{t}_t$ = 0 means the year before the policy implementation year 1998 and $pos{t}_t$ = 1 means the year 1998 and after.

Control variables: Considering that other factors may also have an impact on technological innovation, we select the following control variables in this paper. (i) Urbanization (urb). An increase in the level of urbanization helps to enhance urban technological innovation [37]. This paper uses the percentage of the non-agricultural population multiplied by one hundred to calculate the level of urbanization. (ii) Industrial structure (str). An advanced industrial structure contributes to more technological innovation [38]. The measurement of the industrial structure is undertaken by calculating the proportion of tertiary industry in total output multiplied by one hundred. (iii) Education level (edu). The level of education affects the stock and increment of human capital, which is an important factor affecting technological innovation [39]. The level of education is measured as the percentage of students pursuing general higher education in the total population multiplied by one hundred. (iv) R&D investment (RnD). R&D investment is one of the determinants of technological innovation [40]. This study uses the proportion of scientific researchers multiplied by one hundred to represent the R&D investment. (v) Economic development level (lngrp). The degree of technological innovation in a city is also influenced by its economic development. In general, the more developed the economy, the higher the technological innovation [41]. The level of economic development of a city is characterized by the logarithm of gross regional product.

Statistics on these variables were obtained from the 1995–2003 China City Statistical Yearbook, and descriptive statistics for each variable are shown in

Table 1. Descriptive statistics.

Variable	Obs	Mean	Std. Dev.	Min	Max
lnpatent	2007	4.521	1.294	0	9.545
treat	2007	0.605	0.489	0	1
urb	2007	31.580	16.513	8.080	93.700
str	1917	36.490	10.134	8.370	74.650
edu	1742	0.503	0.674	0.010	6.170
RnD	1886	39.474	15.409	12.600	83.510
lngrp	1917	14.506	1.023	11.265	17.804

.

3. Results

3.1. Main Results

Table 2. Benchmark regression results.

Variables	(1)	(2)
	FE: lnpatent	RE: lnpatent
treat_post	0.278 ***	0.246 ***
	(0.0230)	(0.0240)
urb	0.013 ***	0.017 ***
	(0.0018)	(0.0015)
str	0.002 *	0.001
	(0.0011)	(0.0011)
edu	0.119 ***	0.156 ***
	(0.0243)	(0.0248)
RnD	0.012 ***	0.009 ***
	(0.0008)	(0.0008)
lngrp	0.351 ***	0.526 ***
	(0.0238)	(0.0207)
Constant	−1.571 ***	−4.163 ***
	(0.3600)	(0.3090)
Observations	1729	1729
R-squared	0.679
Number of code	216	216

Note: * p < 0.1, *** p < 0.01. The standard error is located beneath the coefficient ().

shows the regression results of the baseline model, where columns (1) and (2) show the fixed and random effects regression results, respectively. According to the Hausman test, the chi2 value is 249.1 and the p-value is 0.000, indicating that fixed effects regression should be selected and random effects regression rejected. Therefore, the analysis in this paper makes use of the fixed effect regression model.

The results of fixed effects regression in Table 2 show that, after controlling for time fixed and city fixed effects, the coefficient of the core explanatory variable $trea{t}_i·pos{t}_t$ is 0.278 and is significantly positive at the level of 1%, indicating that the technological innovation level of cities in the TCZ is 27.8%, which is higher than cities not in the TCZ. The findings support Porter’s notion that environmental regulations might foster urban technological innovation. One possible explanation for this could be that businesses will have to adhere to more stringent pollution reduction standards following the TCZ policy’s adoption, and they may develop green technologies to achieve energy saving and emission reduction to meet the policy requirements. Otherwise, they will face higher environmental governance costs and may increase production efficiency through technological innovation to offset these costs. Meanwhile, in order to guarantee the effective execution of the policy, the government has increased financial support for green production, which reduces the cost of green production for enterprises and allows them to invest more in technological innovation. For the control variables, the empirical results show that urbanization (urb), industrial structure (str), education level (edu), R&D investment (RnD), and economic development level (lngrp) all significantly improve urban technological innovation, which is in line with the conclusions of the current research.

3.2. Robustness Checks

3.2.1. Parallel Trend Hypothesis and Dynamic Test

Figure 1 shows the average number of patents granted in the experimental and control groups. It is clear from this figure that, prior to the policy’s implementation, the trend in the average number of patents granted in the two groups of cities was relatively stable. However, following the policy’s implementation, the average number of patents granted in the experimental group increased significantly more quickly than in the control group.

The premise of the policy impact assessment using the DID model is that the time trends of the experimental and control groups prior to policy initiation should be the same, and therefore a parallel trend test of this assumption is required. The following econometric model is developed here with reference to Fajgelbaum et al. (2020) [42]:

$$lnpaten{t}_{it}={\beta }_0+{\sum }_{j=-3}^5{\beta }_j·trea{t}_i·I\left(ET=j\right)+X_{it}·\gamma +{\theta }_i+{\theta }_t+{\epsilon }_{it}$$

(2)

where i represents the city, t represents the year, ${\theta }_i$ represents city fixed effects, and ${\theta }_t$ represents time fixed effects. The dependent variable remains the logarithm of patent numbers. I is the relative time indicator variable and j represents the relative time, where j = 0 refers to the current policy implementation period, j = 1 refers to a period after the policy implementation, and j = −1 refers to a period before the policy implementation. From 1995 to 2003, j is defined as −3 to 5 in that order. Variables $trea{t}_i$ and $X_{it}$ match the baseline model (1). The parallel trend test excludes the base period (1995) and results are shown in Figure 2.

Figure 2 illustrates the levels of technological innovation in the experimental group and the control group. The results show that although technological innovation levels in the two groups varied prior to policy implementation, their overall trends remained the same. After the implementation of the TCZ policy, the level of technological innovation in experimental and control group cities began to show significant differences. Therefore, these results passed the parallel trend test, and it is feasible to use the DID model for policy impact assessment.

3.2.2. Placebo Test

In accordance with the relevant literature, this study conducts a placebo test on the results of previous calculations to validate the robustness of the findings [43]. Assuming that we randomly chose cities from the 223 cities as the experimental group, the remaining cities served as the control group, and the DID process was repeated 500 times. Correspondingly, we received 500 distinct interaction term coefficients. The significance and distribution of the interaction term’s coefficients might then be used to assess how reliable the earlier regression results were.

Figure 3 demonstrates the placebo test results, which indicate the distribution of the interaction term coefficients acquired from the aforementioned random repeated experiments. The interaction term coefficients have an essentially normal distribution and are concentrated around the zero point. The p-values of most of the interaction term coefficients are greater than 0.1, indicating that before and after the TCZ policy was put into place, the technological innovation effects determined by the experimental groups chosen at random showed no discernible differences. The results show that, in reality, the promotion of technological innovation in the cities where the TCZ policy is implemented is caused by this policy and not by other factors. Therefore, the baseline model’s regression results are extremely reliable and robust. One could argue that the TCZ policy’s implementation helps to foster urban technological innovation.

3.2.3. Propensity Score Matching (PSM)

In order to reduce the subjective bias that may exist when delimiting cities in the TCZ and avoid biased estimates, this study adds some key control variables that may influence city selection in the TCZ using the benchmark model, such as the size of the population in the year, the overall investment amount in social fixed assets, and the amount of foreign direct investment in the year [44], and on this basis, used the PSM-DID approach to estimate, so as to eliminate the bias. As shown in

Table 3. PSM-DID regression results.

Variables	(1)	(2)
	OLS: lnpatent	FE: lnpatent
treat_post	0.156 ***	0.244 ***
	(0.0328)	(0.0231)
urb	0.269 ***	0.255 ***
	(0.0380)	(0.0567)
str	0.350 ***	0.00498
	(0.0657)	(0.0478)
edu	0.183 ***	0.132 ***
	(0.0168)	(0.0199)
RnD	−0.0500	0.395 ***
	(0.0417)	(0.0286)
lngrp	0.933 ***	0.397 ***
	(0.0173)	(0.0242)
Constant	−10.72 ***	−3.356 ***
	(0.396)	(0.435)
Observations	1664	1664
R-squared	0.741	0.680

Note: *** p < 0.01. The standard error is located beneath the coefficient ().

, the PSM-DID regression results still confirm the research hypothesis.

3.3. Heterogeneity Analysis

Although the previous analysis shows that the TCZ policy does promote urban technological innovation, the effect of this promotion may vary according to the location where the policy is implemented. This section discusses the heterogeneous effects of policy implementation in terms of regional development characteristics and urban enterprise ownership.

3.3.1. Heterogeneity of Regional Development Characteristics

Due to China’s vast territory, factors affecting economic development such as geographical conditions vary greatly from region to region, resulting in significant differences in the level of economic development between the Eastern, Central, and Western regions of China. Specifically, in terms of the level of economic development, the Central and Western regions are less developed than the Eastern region. In terms of regional industrial development realities, pollution emissions are higher in the Eastern and Central regions and lower in the Western region. Therefore, this paper divides the sample into these three regions and the regression results for each region are shown in

Table 4. Impact of heterogeneity in regional development characteristics.

Variables	(1)	(2)	(3)
	Western Region	Central Region	Eastern Region
	lnpatent	lnpatent	lnpatent
treat_post	0.439 ***	0.150 ***	0.243 ***
	(0.0579)	(0.0377)	(0.0247)
urb	0.000	0.004	0.014 ***
	(0.0047)	(0.0027)	(0.0019)
str	0.000	0.000	0.002 **
	(0.0027)	(0.0015)	(0.0011)
edu	0.099 *	0.151 ***	0.121 ***
	(0.0514)	(0.0381)	(0.0270)
RnD	0.013 ***	0.014 ***	0.0123 ***
	(0.0018)	(0.0012)	(0.0008)
lngrp	0.021	0.311 ***	0.407 ***
	(0.0612)	(0.0365)	(0.0256)
Constant	3.320 ***	−0.991 *	−2.412 ***
	(0.9000)	(0.5540)	(0.3890)
Observations	257	694	778
Number of code	34	88	94
R-squared	0.649	0.647	0.696

Note: * p < 0.1, ** p < 0.05, *** p < 0.01. The standard error is located beneath the coefficient ().

. The coefficients of the key explanatory variable show that once the TCZ policy was put into effect, the level of technological innovation in the Western region increased the most, while the Central region was least affected and the Eastern region was in the middle.

With regard to economic and industrial development, the Eastern region leads the way, so it is able to achieve green development through technological innovation when facing environmental regulations. Economic growth in the Western region is largely based on the abundance of natural resources. As a result, industrial development in the Western region is highly lagging behind. Environmental regulatory policies are putting less pressure on enterprises in Western China. Since the Western region has a relatively low overall technology level, there is great potential for technological innovation and improvement. The Central region relies heavily on traditional high-emission manufacturing. At the same time, the Central region does not have the same economic capacity as the Eastern region, which leads to greater difficulties in the green transformation of enterprises as the intensity of environmental protection regulation increases. Therefore, in comparison to the Eastern and Western regions, the Central region does not promote technological innovation to the same extent.

3.3.2. Heterogeneity of Urban Enterprise Ownership

The ownership status of enterprises also profoundly affects the economic performance. In this paper, enterprise ownership is broadly classified into two types: SOEs (state-owned enterprises) and non-SOEs. Among them, state-owned enterprises refer to enterprises invested in and controlled by central or local governments, including wholly state-owned enterprises and state-controlled enterprises; non-state-owned enterprises refer to enterprises with ownership types other than those of state-owned enterprises. Due to the distinct differences in shareholder characteristics between SOEs and non-SOEs, the impact of environmental regulations on enterprises with different ownership may differ. To investigate the heterogeneous impacts on urban technological innovation that may result from differences in the ownership status of enterprises, this study separates the sample into two groups by proportion of SOEs. Specifically, the proportion of state-owned enterprises in fixed asset investment is used to measure this urban enterprise ownership difference. In this paper, cities with a higher-than-average proportion of state-owned fixed asset investment are classified as cities with a high proportion of SOEs, and cities with a lower-than-average proportion of state-owned fixed asset investment are classified as having a low proportion.

The results are displayed in

Table 5. Impact of heterogeneity in urban enterprise ownership.

Variables	(1)	(2)
	High Proportion of SOEs	Low Proportion of SOEs
	lnpatent	lnpatent
treat_post	0.330 ***	0.223 ***
	(0.0403)	(0.0272)
urb	0.012 ***	0.010 ***
	(0.0037)	(0.0020)
str	0.002	0.001
	(0.0019)	(0.0013)
edu	0.359 ***	0.108 ***
	(0.0683)	(0.0250)
RnD	0.013 ***	0.012 ***
	(0.0013)	(0.0009)
lngrp	0.441 ***	0.326 ***
	(0.0467)	(0.0268)
Constant	−2.756 ***	−1.144 ***
	(0.7030)	(0.4060)
Observations	605	1124
Number of code	78	138
R-squared	0.752	0.646

Note: *** p < 0.01. The standard error is located beneath the coefficient ().

, which indicates that cities in the high-proportion SOE group benefit more from the TCZ policy in terms of technological innovation promotion than those in the low-proportion SOE group. Possible reasons for this are as follows. First, the government has greater control over SOEs and is more likely to regulate their use of subsidies and funds, which, as relevant studies have also shown, gives SOEs an advantage in terms of access to preferential policies and financial support [45], and thus SOEs tend to have more funds for technological innovation. Second, SOEs, which are usually larger in size, are more likely to receive financial support for technological innovation, and thus have greater capacity to afford the cost of upgrading traditional production equipment.

4. Discussion

Establishing the TCZ policy is one of China’s important initiatives to focus on ecological and environmental governance after its rapid development. The policy has curbed SO₂ emissions in various regions of China, and is of great significance for the transformation of the economic development pattern and the realization of sustainable development in China. Balancing economic development and environmental protection is a common problem that countries around the world, especially developing countries, must face. For this reason, environmental regulation policies are essential.

Taking the TCZ policy as an example, this paper empirically examines the impact of this policy on urban technological innovation in the hope that the results of this study can provide a reference for other countries in transforming their economic development patterns and promoting sustainable development. Findings from this study demonstrate the effectiveness of the TCZ policy, by showing that it does not inhibit urban technological innovation, but rather promotes it. The policy has imposed higher emission reduction requirements and pollution control costs, forcing enterprises to engage in green technology innovation to meet the policy’s emission reduction requirements and production technology innovation to improve production efficiency to offset these costs. However, this strict environmental policy did not affect the incentive for technological innovation, as the TCZ policy provided financial subsidies to reduce the cost of green technological innovation, making enterprises more capable of technological innovation.

The heterogeneity analysis’s findings can assist us in determining the variables that affect how the TCZ policy fosters technological innovation in urban areas.

First, the TCZ policy has a more favorable impact on technological innovation in the Western region, followed by the Eastern region, and the Central region is the least affected. As previously analyzed, the Eastern region has a comparatively high level of economic development, while the Western and Central regions have very low levels. Meanwhile, the Western region experiences comparatively little policy compliance pressure. while the policy compliance pressure in the Central and Eastern regions is relatively high. When faced with environmental regulatory policies, regions with higher levels of economic development are more capable of paying additional costs for technological innovation, while regions with lower policy compliance pressures require minimal compliance costs, which therefore do not affect their innovation.

Secondly, cities with a higher percentage of SOEs are more likely to benefit from the TCZ policy in terms of technological innovation than those with a smaller percentage of SOEs. This implies that SOEs do not neglect pollution control because of their greater affordability of additional costs, but rather take advantage of this to better innovate. As the majority shareholder of SOEs is the central or local government, SOEs’ behavior tends to be more in line with government policy objectives. This is supported by studies showing that SOEs are more legally sensitive and more likely to comply with these increasingly stringent environmental requirements [46]. The findings of this paper suggest that, despite their greater ability to withstand external risks and additional costs, SOEs have not responded negatively to environmental protection requirements because of these characteristics, but have instead been able to use these advantages of their own to accelerate technological innovation. In order to ensure that policy implementation is rigorous and effective, the role of state-owned enterprises should be fully utilized.

The following policy recommendations are made in light of the primary findings of this article.

First, governments must implement appropriate environmental regulatory policies as early as possible. Specific measures include formulating comprehensive prevention and control plans, restricting the extraction and utilization of conventional energy sources that are highly polluting (like high-sulfur coal), limiting pollutant emissions from highly polluting industries, developing emission reduction technologies and equipment, and implementing proper sewage fee collection and management. In the long term, the impact of environmental regulation policies on technological innovation has been positive, reflecting the fact that such policies promote technological upgrading by firms to meet policy environmental requirements. Therefore, the early implementation of environmental regulatory policies is a necessary way for a country to achieve sustainable development.

Secondly, in formulating environmental regulations, governments ought to focus more on regions with comparatively low levels of economic development or where there is greater pressure for policy compliance. Regions with these types of characteristics will have greater difficulty in upgrading technological innovation. Governments should adopt appropriate supportive measures to ensure the implementation of environmental regulation policies. Among the various supportive measures, government subsidies and incentives are the most critical. For example, China’s TCZ policy clearly states that “In order to prevent and control SO₂ pollution, projects and funds for technological transformation, comprehensive utilization and clean production should be tilted to the TCZ.” Policy facilitation and coverage should be enhanced to ensure that enterprises have sufficient funds to carry out various technological innovations, including green production technology innovations.

Finally, governments should strengthen their control over state-owned enterprises, especially those that are stronger but emit more. In order to better ensure SOEs’ performance on environmental protection issues, it is important to strengthen their obligation to implementing environmental regulation policies and contribute to SOEs’ leading role in promoting environmental protection. Governments may require them through administrative orders to complete the renewal of production equipment, the application of energy-saving technologies, and the optimization of production processes within a specified time frame.

5. Conclusions

In order to determine whether the TCZ policy in China promotes urban technological innovation, a DID model is built in this paper by treating the TCZ policy as an exogenous factor, and this model is used to examine how the policy affects urban technological innovation based on panel data of 223 cities from 1995 to 2003. According to the findings, there has been a notable increase in urban technological innovation as a result of the TCZ policy, which is in line with Porter’s hypothesis. Heterogeneity analysis shows that it is further observed that the promotional effects of technological innovation resulting from the TCZ policy are highest in the Western region, followed by the Eastern region, and finally the Central region, and higher in cities with a higher proportion of state-owned enterprises than in cities with a higher proportion of non-state-owned enterprises. The findings of the study are informative for other countries to formulate appropriate environmental regulatory policies. The results of the heterogeneity analysis in this study can help other countries identify priorities when formulating environmental regulation policies.