Balancing Industrialization with Pollution: Evidence from the Marine Ecological Civilization Demonstration Zone Program in China

Abstract

This study provides empirical evidence on how the Marine Ecological Civilization Demonstration Zone program, a policy integrating industrial upgrading with ecological protection, achieves dual objectives in a developing economy context. Theoretically, we uncover three key mechanisms, including green technology innovation, productivity gains, and government intervention, that explain the program’s impacts, offering a framework for reconciling industrialization with environmental sustainability. Our novel findings demonstrate that the program significantly boosts industrial structure upgrading and coastal water quality, with pronounced effects in the Pearl River Delta. We highlight the need for regionally tailored measures, expanded green R&D support, and strategic zone expansion insights critical for China and other developing nations navigating similar trade-offs. By bridging the gap between environmental regulation and industrial transformation, this work contributes to the debates on sustainable development in coastal regions.

1. Introduction

The blue economy is a sustainable marine economy that emerges when economic activity balances with the long-term capacity of marine ecosystems to support such activity while maintaining resilience and health [1]. It is growingly acknowledged as a critical driver for national economic development, especially in boosting the high-quality development of marine industries and building an ecological civilization [2,3]. China has a territorial sea width of 12 nautical miles and an exclusive economic zone of 320,000 square kilometers, and it has enacted multiple blue economy strategies targeting eco-innovation-driven marine sustainability and ecological conservation. In 2012, the State Oceanic Administration (SOA) released the “Opinions on Carrying out the Development of Marine Ecological Civilization Demonstration Zones” (MECDZ, hereafter), aiming to achieve two goals: promoting industrial structure upgrading (ISU, hereafter) and reducing pollution. Specifically, marine ecological civilization is a sustainable development paradigm that harmonizes ocean utilization with ecological conservation through institutional, technological, and cultural innovation, and as an integral part of national strategy, it promotes economic growth and fosters harmonious human-sea relations [4].

Our work thus studies the influences of implementation of a large program, MECDZ, on the ISU of coastal cities in China. In 2013, the National Oceanic Administration approved 12 cities as the first batch of demonstration cities, followed by another 11 cities in 2015 (for details, see Section 2). The existing literature presents a mixed consensus on whether environmental regulations promote industrial upgrades. On one side, some studies indicate that enforcing environmental regulations can sometimes contradict the optimization of industrial structures. This contradiction arises because complying with new standards and requirements necessitates substantial investments and pollution control costs [5,6]. On the other side, environmental policies facilitate the upgrading of industrial structures. Ref. [7] holds that environmental protection policies can boost regional economic growth by refining industrial structures, fostering specialized industrial clusters, and promoting industrial transfer. As industries adapt to these regulations, they not only comply with environmental standards but also unlock new opportunities for growth and development, ultimately contributing to a more sustainable future [8].

Given these confounding effects of environmental regulation, we characterize the effectiveness of MECDZ by adopting staggered difference-in-differences (SDID). Our SDID estimates show that, relative to unregulated cities, regulated cities promote their ISU after the implementation of this program. Regulated cities achieved these upgrades by increasing green innovation, green total factor productivity (GTFP, hereafter) and local government support. Our main results still remain unchanged after carrying out a set of robustness checks. Heterogeneous analyses indicate that the MECDZ drives significant ISU in the Pearl River Delta demonstration cities, and its effects prove statistically insignificant in both the Bohai Rim and Yangtze River Delta regions. It is also significant in the cities identified as focus cities for environmental protection in the “China’s 11th Five-Year Environmental Plan” released by the State Council in 2007. Finally, we provide evidence of industrial upgrading spillovers by showing the effects on ocean pollutions, i.e., water quality, active phosphate concentration, and inorganic nitrogen concentration. These results corroborate the notion that the MECDZ policy has a positive impact on improving coastal water quality via green technological innovation.

This paper yields valuable contributions to our understanding of whether the MECDZ program aimed at improving industrial structure and protecting coastal cities’ ecological environment is effective in the largest developing country. First, unlike previous work that has primarily assessed the influences of environmental supervision on industrial structure [9,10,11], our research shows nuanced knowledge of how the MECDZ program influences these aspects. By enriching the discourse on environmental policy that balances ISU with pollution reduction [4,12,13], our findings provide actionable policy implications for emerging economies confronting analogous marine sustainability challenges.

Second, this paper investigates the underlying economic channels that may influence the effectiveness of the MECDZ policy by analyzing green innovation, GTFP, and local government support. Notably, we utilize the super efficiency slack-based measurement to assess urban GTFP, thereby addressing the radial and angular issues inherent in methods [14]. This deepens the evaluation of the environmental governance effects of the MECDZ policy compared to the existing literature on environmental policies. Our study provides critical evidence for the expansion of pilot areas and policy optimization in China, facilitating a more sustainable industrial transformation while concurrently addressing pollution concerns.

2. Conceptual Framework and Literature Review

2.1. Marine Ecological Civilization Demonstration Zone Policy

The SOA issued the “Opinions on Promoting the Construction of Marine Ecological Civilization Demonstration Zone” and its revised draft, which serve as critical guidelines for the establishment of the first and second batches of MECDZ in 2013 and 2015, respectively [15,16]. The list of coastal cities is shown in Figure 1. These documents emphasize the importance of pollutant emission reduction as a key component in evaluating ecological protection and construction efforts. These objectives are regarded as the most significant goals within the MECDZ framework [17].

The MECDZ program mandates the establishment of a rich and professional target system and assessment methods. The first objective focuses on ISU, which entails a scientific selection of marine development methods and activities. This represents a paradigm shift from prioritizing economic expansion at the expense of environmental conservation to a balanced development model that concurrently emphasizes environmental conservation and economic advancement. This policy calls for accelerated industrial restructuring, a gradual elimination of outdated production capacities, and the phasing out of enterprises that are heavily polluting.

The second objective aims at reducing pollutant emissions in coastal areas. It advocates for a coordinated approach between land and sea. This involves strengthening marine protection and restoration efforts while intensifying comprehensive pollution control measures. The integration of these objectives is essential for achieving sustainable development within marine environments, ensuring that ecological integrity is maintained while fostering economic growth.

Central to the ocean ecological civilization is the recognition of the ocean’s intrinsic value, not only as a resource for economic exploitation but also as a vital component of the Earth’s biosphere. The initiative encourages the adoption of sustainable practices in fishing, tourism, and maritime industries, aiming to minimize ecological footprints while maximizing socio-economic benefits. This involves implementing policies that regulate resource extraction, enhance marine conservation efforts, and promote the marine habitat rehabilitation.

The construction of ocean ecological civilization represents a comprehensive approach to managing marine resources and ecosystems in a sustainable manner. By integrating economic development with ecological protection, this concept seeks to create a future where humans and the ocean can thrive together. As the challenges facing our oceans continue to grow, the commitment to building an ocean ecological civilization becomes increasingly urgent, requiring concerted efforts at local, national, and international levels.

2.2. The MECDZ Policy and Upgrading of Industrial Structure

The effect of environmental policies on ISU is a subject of ongoing scholarly debate, revealing complex mechanisms influenced by regulatory design and context. While some scholars argue that environmental policies hinder ISU due to the short-term compliance cost effect, which refers to the static burden imposed on firms for adhering to regulations [18], the prevailing empirical evidence supports what is known as the promotion theory. In the long run, stringent environmental regulations exert what ref. [19] termed a survival-of-the-fittest effect. Small- and medium-sized enterprises unable to internalize rising pollution control costs are forced to exit the market. Surviving firms adapt by restructuring operations and reallocating resources towards cleaner production methods. Larger polluting firms generally find it easier to absorb these costs through means such as investment in equipment compared to small- and medium-sized enterprises, for whom compliance costs can threaten their viability. This selective pressure inherently promotes ISU.

Furthermore, ref. [20] described green barriers, increasing marginal and sunk costs for potential new entrants into polluting sectors. These barriers curb the expansion of polluting industries and reduce new firm entry [21], thereby positively influencing overall industrial structure. Governments also leverage environmental regulations, often combined with economic instruments like taxes and subsidies, to boost the competitive edge of cleaner sectors and drive the adoption of eco-friendly production methods across the economy, fostering higher-level industrial structures. Ref. [22], employing a tri-dimensional indicator system encompassing industrial development, ecology, and green potential, found that marine environmental regulations significantly promote marine industrial green transformation, with market-incentive instruments proving more effective than command-and-control types. Studies on coastal manufacturing, such as that by ref. [9], have demonstrated that increasing marine environmental regulation stringency first promotes the transfer of polluting industries away from coastal areas, subsequently facilitating local ISU. Ref. [22] corroborated this sequential effect of transfer followed by upgrading in Hainan.

The MECDZ policy, as a stringent legislative and administrative policy, establishes distinct green barriers through specific instruments. These barriers prevent new high-pollution firms from entering marine sectors, phase out obsolete enterprises, and compel existing resource-intensive firms to upgrade their structures. Drawing upon the evidence, the hypothesis is proposed:

Hypothesis 1.

The MECDZ policy significantly promotes ISU.

2.3. Environmental Regulation and Green Technology Innovation

Recent studies have highlighted a significant role that environmental regulations play in influencing industry structure upgrades through green technology innovation (GTI). The theoretical foundation for this relationship is strongly supported by the Porter Hypothesis (PH) [23], which argues that properly formulated environmental policies can boost innovation. Such innovation may counteract compliance costs, thereby enabling firms to boost competitiveness. Prior studies have found that environmental policies influence technological innovation [24,25,26,27]. Similarly, Ref. [28] evaluated urban environmental regulations using a comprehensive scoring system, revealing that the technological barriers established by these regulations promote innovation, leading to structural optimization in industries. Ref. [19] identified that GTI acts as a positive mediator between the stringency of environmental policies and ISU. However, the overall outcome depends on how the short-term cost burden interacts with long-term efficiency improvements that eventually outweigh initial expenses [29,30]. Excessively stringent environmental regulations can exceed the capacity of enterprises, stifling their technological innovation.

The MECDZ policy emphasizes controlling pollutant emissions, compelling enterprises within demonstration zones to focus more on developing environmentally friendly technologies to meet regulatory requirements [31]. Stringent environmental rules require enterprises to meet stricter environmental criteria, which calls for spending on green tech R&D, such as upgrading production facilities or adopting clean energy sources [32,33]. Well-designed regulations can trigger the “Strong Porter Hypothesis,” driving innovations beyond mere compliance that reduce costs and create competitive advantage [34]. These innovations not only aim to meet environmental standards but also facilitate internal advancements in industrial structures [35]. Firms that fail to comply with these standards face the pressure of obsolescence and relocation. Grounded in the theoretical expectations of the PH, this work underscores the potential for environmental policies to serve as catalysts for industrial transformation and sustainability in coastal blue economies [36]. Based on the above analysis, hypothesis 2 is proposed:

Hypothesis 2.

GTI is the underlying economic channel through which the MECDZ policy affects the industrial upgrading of coastal cities.

2.4. Environmental Regulation and Green Total Factor Productivity

GTFP is a comprehensive production efficiency indicator that incorporates environmental factors, allowing for a more holistic assessment of the sustainability and environmental friendliness of economic development. Theoretically, GTFP extends traditional productivity measurement by internalizing environmental externalities through the inclusion of undesired outputs such as pollutants, thereby aligning economic performance with ecological constraints [37]. By taking into account input consumption and output pollution, GTFP delivers a more precise representation of resource usage and development efficiency [38]. A substantial body of empirical evidence indicates that environmental regulations can foster the growth of GTFP [39,40]. Ref. [41] identified a positive linear correlation between environmental regulations and GTFP in economically advanced urban agglomerations, while other urban clusters exhibited a U-shaped nonlinear relationship. Ref. [40] focused on China’s manufacturing sector, demonstrating that environmental regulations in moderately and lightly polluting industries follow a U-shaped curve concerning GTFP.

Environmental regulations influence the shift in China’s industrial growth pattern through their influence on GTFP, although there exists a threshold effect regarding the intensity of these regulations [42]. Ref. [43] also found significant impacts of industrial structure upgrades and environmental regulations on GTFP, exhibiting regional characteristics. Ref. [44] studied low-carbon pilot policies, revealing that such policies significantly affect the GTFP of pilot cities and can also promote GTFP through industrial structure adjustments.

Under strengthened environmental regulations, enterprises are compelled to innovate green technologies, optimize production processes, and improve resource allocation, leading to reduced waste, lower emissions, and improved GTFP [39,45]. This shift not only drives ISU towards high technology and low pollution [46] but also forces companies to boost spending on research and development investments in energy-saving and emission-control technologies in response to stringent policies like emission standards and carbon taxes [23]. Simultaneously, such regulation encourages the exit of high-pollution, low-efficiency enterprises from the market, facilitating the transfer of resources to high-tech and high-value-added industries [47], further reinforcing GTFP growth. MECDZ establishes stringent green barriers for the development of marine industries through rigorous standards, prohibitions, and evaluation systems, such as marine dumping permits and the marine ecological redline policy. These measures prevent highly polluting enterprises from entering the market and gradually phase out outdated firms, thereby compelling resource-intensive enterprises to reduce waste, cut emissions, and ultimately enhance GTFP. Based on the above analysis, hypothesis 3 is proposed:

Hypothesis 3.

GTFP is the underlying economic channel through which the MECDZ policy affects the industrial upgrading of coastal cities.

2.5. Environmental Regulation and Local Government Support

The core of environmental regulation promoting industrial structural upgrading through local government support lies in the government’s ability to internalize environmental externality costs via policy tools. This process drives enterprises to revise their production patterns and improve resource distribution, in turn steering the market towards green, high-efficiency industries and eventually realizing a qualitative change in industrial structure [48]. Governments can strategically intervene to coordinate the pace of industrial transformation, preventing premature deindustrialization and enhancing the upgrading of industrial structures [49].

Crucially, environmental regulation intrinsically empowers and incentivizes local government intervention (GI, hereafter). Stringent environmental standards and performance-based accountability systems, such as central environmental inspections and the integration of green GDP metrics into cadre evaluations, fundamentally realign local governments’ incentives. By elevating environmental goals to parity with economic growth in political tournaments, environmental regulation transforms local governments from passive regulators into active enforcers seeking legitimacy and competitive advantage through green governance [50,51].

Existing research has revealed the mechanisms through which environmental regulation, facilitated by government intervention, drives industrial structural upgrading from multiple dimensions. By utilizing fiscal expenditure tools, local government can expand the scale of capital accumulation within industries and enhance labor productivity, forming a sustained driving force for industrial structural upgrading [52]. However, the efficacy of such policies is restricted by dynamically adjusted local fiscal capacities. When local governments face lower fiscal pressure and have ample disposable financial resources, the industrial upgrading effects of environmental regulation become more pronounced [53].

Furthermore, the effects of intervention exhibit significant heterogeneity. In regions with a strong foundation in marine economies, government intervention can enhance industrial upgrading effects through resource integration and policy orientation. Conversely, in areas with a weak marine economic base, excessive intervention may stifle market vitality and hinder structural transformation [54]. Hence, the following hypothesis is drawn as follows:

Hypothesis 4.

Local government support is the underlying economic channel through which the MECDZ policy affects the industrial upgrading.

Overall, this paper’s research framework is depicted in Figure 2.

3. Data and Model Specification

3.1. Data

This works used 53 coastal Chinese cities spanning 2007 to 2019 as its research sample. The starting time was dictated by the fact that statistics on indicators for coastal prefecture-level cities have been included from 2007, such as those on industrial wastewater directly released into the ocean. It designated 22 cities from the two batches of MECDZ policy in 2013 and 2015 as the treatment group, while other 31 coastal cities served as the control group. The list of cities is presented in

Table 1. List of demonstration cities.

Policy Issued Time	Marine Ecological Civilization Demonstration Cities
2013	Shandong Province: Weihai City, Rizhao City, Yantai City; Zhejiang Province: Ningbo City, Taizhou City, Wenzhou City; Fujian Province: Xiamen City, Quanzhou City, Zhangzhou City; Guangdong Province: Zhuhai City, Shantou City, Zhanjiang City
2015	Liaoning Province: Panjin City, Dalian City; Shandong Province: Qingdao City; Jiangsu Province: Nantong City, Yancheng City; Zhejiang Province: Zhoushan City; Guangdong Province: Huizhou City, Shenzhen City; Guangxi Zhuang Autonomous Region: Beihai City; Hainan Province: Sanya City

.

The core independent variable is policy dummy variables (MEC), equal to one when a city has become a demonstration city in a time according to the MECDZ policy, otherwise zero. Specifically, 12 cities, including Weihai, Rizhao, Yantai, Ningbo, Taizhou, Wenzhou, Xiamen, Quanzhou, Zhangzhou, Zhuhai, Shantou, and Zhanjiang in 2013, and 10 cities, including Panjin, Dalian, Qingdao, Nantong, Yancheng, Zhoushan, Huizhou, Shenzhen, Beihai, and Sanya in 2015, were treated groups. The remaining 31 coastal prefecture-level cities were control groups.

The dependent variable was a measure of ISU. Following ref. [55,56,57,58], control variables included Lnperincome, Lnedu, Lnenterprises, Lnpop, and Lnfdi (all the variables are defined in Appendix A). The above data were sourced from the China City Statistical Yearbook and processed as follows: (1) we standardized measurement units across all variables; (2) we addressed isolated missing values through time-trend interpolation to maintain panel balance.

As to potential channels, the data for green technology innovation capability (GTI) were sourced from the China National Intellectual Property Administration and measured by the natural logarithm of the number of green patent applications. Data on green total factor productivity (GTFP) incorporating environmental factors were obtained from the CEINET Statistics Database and China City Statistical Yearbook, with detailed calculation procedures provided in Appendix C. The degree of government intervention (GI) originated from the China City Statistical Yearbook and was quantified as the proportion of fiscal expenditure of prefecture-level cities relative to provincial fiscal expenditure, with missing values in certain cities or years being supplemented through interpolation methods.

The data processing methods enhance the reliability and comparability of the results by minimizing measurement inconsistencies and preserving panel balance. Standardization ensures uniform scaling across variables, reducing bias from unit disparities, while interpolation mitigates the risk of sample attrition bias caused by missing data. Logarithmic transformations help normalize skewed distributions (e.g., income, patent counts) and improve model interpretability by approximating linear relationships. Such processing is common in empirical studies and strengthens cross-city comparability, further validating the findings.

The definition of variables is defined in Appendix A. The descriptive statistics results are presented in

Table 2. Statistics.

Variable	Obs	Min	p50	Max	Mean	Std. Dev
MEC	689	0	0	1	0.194	0.396
ISU	689	0.156	0.459	0.792	0.465	0.105
Lnperincome	689	6.778	8.631	10.271	8.530	0.558
Lnedu	689	3.540	5.448	6.294	5.440	0.314
Lnenterprises	689	2.996	7.440	9.841	7.291	1.237
Lnpop	689	4.904	6.311	7.923	6.323	0.564
Lnfdi	689	0.069	3.398	5.270	3.244	1.052
STRAD	689	5.956	6.638	7.836	6.652	0.330
GTI	689	0	3.135	7.857	3.296	1.888
GTFP	689	0.835	0.998	1.183	0.999	0.020
GI	689	0.009	0.045	0.502	0.081	0.095
WQ	689	1	2	5	2.814	1.567
LnAP	517	0	0.801	2.061	0.837	0.347
LnDIN	517	0	3.060	5.441	3.165	0.683

.

3.2. Model Specification

The MECDZ policy, formulated by legislative and administrative authorities, establishes green barriers for marine industries through rigorous standards, prohibitions, and evaluation systems [21]. These mechanisms prevent high-pollution enterprises from market entry while progressively phasing out obsolete operations, simultaneously compelling resource-intensive firms to drive ISU [19]. To test these, we study the causal between MECDZ and the ISU of coastal cities to estimate staggered difference-in-differences specification of the form

$$Y_{i,t}=\alpha +{\beta MEC}_{i,t}+{\gamma }^{\prime }{Controls}_{i,t}+u_i+v_t+{\epsilon }_{i,t}$$

(1)

where Y is the proportion of the tertiary industry in GDP (ISU); MEC is a dummy variable as a measure of environmental regulation effect proposed in the MECDZ program. Our interest is the estimate $\beta .$ $Controls$ account for economic level (Lnperincome), education expenses (Lnedu), industrialization (Lnenterprises), city size (Lnpop), and open economy (Lnfdi). We include city-level ($u_i$) and time-level ($v_t$) fixed effects. i is the city, and t is the annual period. ${\epsilon }_{i,t}$ is an error term.

4. Empirical Results

4.1. Baseline Regression Results

We examined the influences of MECDZ policy on the industrial upgrading by adopting SDID. The results are shown in

Table 3. Baseline regression.

	(1)	(2)
	ISU	ISU
MEC	0.0187 ***	0.0162 ***
	(0.00635)	(0.00607)
Lnperincome		−0.0134
		(0.0184)
Lnedu		0.0050
		(0.0129)
Lnenterprises		−0.0750 ***
		(0.0109)
Lnpop		−0.0043
		(0.0314)
Lnfdi		0.0070 **
		(0.0029)
Constant	0.420 ***	1.0440 ***
	(0.00582)	(0.2120)
Observations	689	689
R-squared	0.3997	0.4700
Time-fixed effect	Yes	Yes
City-fixed effect	Yes	Yes

Notes: this table shows estimates of Equation (1). All the variables are defined in Appendix A. The corresponding standard errors are reported in parentheses. *** and ** represent statistical significance at the 1% and 5% level, respectively.

. Column 1 shows the estimate without controls, while Column 2 shows the coefficients with controls using city and year fixed effects. The results show that the MECDZ program significantly promoted ISU with 1.62%.

As to control variables, more large-size firms hindered the upgrading (−7.50%) since the increase in the number of large-scale industrial enterprises expanded the added value of the secondary industry [56], while openness had a significant positive impact on the ISU (0.70%). Foreign direct investment brought financial support and advanced management systems to the host country, thereby enhancing the optimization and upgrading of the ISU in the host country [59].

We also provide parallel trend tests as shown in Figure 3. The estimate coefficients in pre-periods were insignificant, whereas those coefficients turned positive and significant starting from the second post-period. This identifies that our estimations are following the parallel trend.

We conduct placebo tests by randomly selecting from the sample to form a pseudo-treatment group. The baseline regression was then repeatedly estimated, and the bootstrap was employed to conduct 500 times. Figure 4 illustrates the results. The horizontal axis denotes the estimated coefficients of the MECDZ policy for randomly selected groups, with the left vertical axis (Y1) indicating p-values and the right vertical axis (Y2) presenting kernel probability density. The curve stands for the kernel density distribution. The original coefficient is marked by a vertical red line, positioned in the tail of the placebo test distribution. Most p-values exceeded 0.1 and were far from the baseline regression coefficient. This verifies that our findings are not attributable to randomness.

4.2. Robustness Check

4.2.1. Propensity Score Matching-Difference-in-Differences

The cities designated as demonstration areas for the MECDZ policy were not selected randomly; they were approved through a comprehensive evaluation by the government. This may lead to endogeneity arising from selection bias. We thus adopted the propensity score matching (PSM) method to address this concern. Then, we proceeded the staggered DID model.

Table 4. Balance test of PSM.

Variable	Unmatched/Matched	Mean		Bias (%)	Reduct Bias (%)	t-Test
		Treated	Control			t
Lnperincome	U	8.535	8.527	1.500		0.190
	M	8.535	8.493	7.600	−417.600	0.910
Lnedu	U	5.395	5.471	−24.80		−3.170
	M	5.395	5.421	−8.400	66.200	−1.090
Lnenterprises	U	7.407	7.210	15.80		2.070
	M	7.407	7.336	5.700	64.000	0.680
Lnpop	U	6.410	6.262	26.90		3.430
	M	6.410	6.451	−7.400	72.500	−1.000
Lnfdi	U	3.417	3.121	28.70		3.670
	M	3.417	3.397	2.000	93.200	0.260

Notes: this table shows balance tests using PSM to address endogeneity arising from selection bias. All the variables are defined in Appendix A.

shows the results of the balance test of PSM. After matching, the absolute values of standardized bias for each control variable across the treatment and control groups were below 20%.

Table 5. Robustness checks.

Variable	(1)	(2)	(3)	(4)
	ISU	ISU	STRAD	STRAD
MEC	0.0168 ***	0.0138 **	0.0240 **	0.0243 **
	(0.0062)	(0.0059)	(0.0117)	(0.0117)
Lnperincome		−0.0029		0.0370
		(0.0188)		(0.0355)
Lnedu		−0.0002		−0.0285
		(0.0148)		(0.0250)
Lnenterprises		−0.0696 ***		−0.0616 ***
		(0.0110)		(0.0211)
Lnpop		−0.0071		−0.1510 **
		(0.0305)		(0.0606)
Lnfdi		0.0071 **		−0.0031
		(0.0030)		(0.0056)
Constant	0.4210 ***	0.9720 ***	6.4940 ***	7.7600 ***
	(0.0056)	(0.2130)	(0.0107)	(0.4100)
R-squared	0.4290	0.4850	0.6790	0.6910
Observations	659	659	689	689
Time-fixed effect	Yes	Yes	Yes	Yes
City-fixed effect	Yes	Yes	Yes	Yes

Notes: this table shows estimates of robustness tests. All the variables are defined in Appendix A. The corresponding standard errors are reported in parentheses. *** and ** represent statistical significance at the 1% and 5% level, respectively.

shows the results using PSM-DID. Columns 1 and 2 also show positive and significant coefficients of MEC, consistent with the main results.

4.2.2. Alternative Measure of Industrial Upgrading

To provide stable results, we replaced ISU using advanced industrial structure (STRAD) as a new measure of industrial upgrading in Equation (1). STRAD refers to the process by which industries evolve from lower-level to higher-level stages, in which the industrial structure rises progressively in the order of primary, secondary, and tertiary industries. Drawing on refs. [28,60,61], we split the GDP into three components: primary, secondary, and tertiary sectors, and we show a collection of three-dimensional vectors:

$$x_0=\left(x_{\mathrm{1,0}},x_{\mathrm{2,0}}{,x}_{\mathrm{3,0}}\right)$$

(2)

Next, we calculated the angles ${\theta }_1$, ${\theta }_2$, and ${\theta }_3$, of $x_0$, and the industrial vectors $x_1$, $x_2$, and $x_3$, which are shown as follows:

$$x_1=\left(x_{\mathrm{1,1}},x_{\mathrm{2,1}}{,x}_{\mathrm{3,1}}\right)=\left(\mathrm{1,0},0\right)$$

(3)

where $x_1$ is the primary sector.

$$x_2=\left(x_{\mathrm{1,2}},x_{\mathrm{2,2}}{,x}_{\mathrm{3,2}}\right)=\left(\mathrm{0,1},0\right)$$

(4)

where $x_2$ is the secondary sector.

$$x_3=\left(x_{\mathrm{1,3}},x_{\mathrm{2,3}}{,x}_{\mathrm{3,3}}\right)=\left(\mathrm{0,0},1\right)$$

(5)

where $x_3$ is the tertiary sector.

We calculated ${\theta }_j$ as follows:

$${\theta }_j=\mathit{arccos}\left[{\displaystyle \frac{{\sum }_{i=1}^3(x_{i,j}\times x_{i,0})}{{\sum }_{i=1}^3(x_{i.j}^2{)}^{1/2}\times {\sum }_{i=1}^3(x_{i.0}^2{)}^{1/2}}}\right],j=\mathrm{1,2},3$$

(6)

where ${\theta }_j$ is the angle of $x_0$ and the industrial vector $x_j\left(j=\mathrm{1,2},3\right)$. $x_{i.j}$ denotes an element of the industrial-sector vector $x_j$.

Finally, we obtain STRAD from the equation below:

$$STRAD={\sum }_{k=1}^3{\sum }_{j=1}^k{\theta }_j=3{\theta }_1+2{\theta }_2+{\theta }_3$$

(7)

The larger the STRAD is, the higher the level of ISU is.

The results of STRAD are presented in Columns 3–4, Table 5. The parallel trend tests are shown in Appendix B. We still found positive and significant coefficients of MEC on the advanced industrial upgrading, which is in line with the baseline regressions.

4.3. Underlying Economic Channel

The previous literature suggests that green innovation, green total factor, and local government support can be potential underlying economic channels through which the MECDZ policy affects the industrial upgrading of coastal cities (e.g., refs. [27,39,62]). As to potential channels, variables include GTI, GTFP, and GI. GTFP is measured using the Super-SBM method (for details, see Appendix C). This proxy includes inputs, outputs, labor, capital, and energy, which are selected as input indicators [63]. The ratio of municipal fiscal expenditure to provincial fiscal expenditure serves as a proxy for government intervention intensity (GI). Findings are presented in

Table 6. Mechanism analysis.

Variables	(1)	(2)	(3)
	GTI	GTFP	GI
MEC	0.3330 ***	0.0058 *	0.0017 *
	(0.0727)	(0.0030)	(0.0009)
Lnperincome	0.3430	−0.0291 ***	−0.0030
	(0.2200)	(0.0092)	(0.0027)
Lnedu	−0.2360	−0.0164 **	−0.0004
	(0.1550)	(0.0064)	(0.0019)
Lnenterprises	0.1000	0.0068	0.0029 *
	(0.1310)	(0.0054)	(0.0016)
Lnpop	−0.3450	−0.0001	0.0132 ***
	(0.3770)	(0.0156)	(0.0046)
Lnfdi	0.1040 ***	0.0012	−0.0001
	(0.0351)	(0.0015)	(0.0004)
Constant	1.3580	1.2590 ***	0.0016
	(2.546)	(0.1060)	(0.0311)
Observations	689	689	689
R-squared	0.7900	0.0960	0.0730
Time-fixed effect	Yes	Yes	Yes
City-fixed effect	Yes	Yes	Yes

Notes: this table shows estimates of Equation (1). All the variables are defined in Appendix A. The corresponding standard errors are reported in parentheses. ***, **, and * represent statistical significance at the 1%, 5%, and 10% level, respectively.

. Green innovation, green total factor, and local government support are the underlying economic channels behind the main result.

4.4. Environmental Target

The MECDZ policy aims to reduce pollutant emissions, emphasizing the protection of marine ecological environments. The MECDZ policy necessitates the implementation of performance assessments and accountability mechanisms [16]. Additionally, the policy underscores the importance of promoting marine ecological civilization through education and public awareness campaigns, thereby enhancing environmental protection consciousness among both enterprises and residents [15]. This increased awareness is expected to drive efforts towards marine ecological protection and contribute to reducing pollutant releases.

To test the environmental regulation effectiveness, we explored the regulatory effects of GTI, GTFP, and GI. Following ref. [64], this study employed coastal water quality (WQ), the active phosphate concentration taking natural logarithm (LnAP), and the dissolved inorganic nitrogen concentration taking natural logarithm (LnDIN) as indicators of coastal water pollution (Coastal water quality was categorized into five levels: Class I (highest quality), Class II, Class III, Class IV, and Class V (lowest quality). A higher WQ classification indicates more severe pollution and poorer water quality.)

Table 7. Coastal water pollution.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	WQ	WQ	WQ	LnAP	LnAP	LnAP	LnDIN	LnDIN	LnDIN
MEC	0.2490	5.0730	0.1500	−0.2660 ***	0.0090	−0.2550 ***	−0.0775	1.4610	−0.0463
	(0.3510)	(3.7900)	(0.2280)	(0.0760)	(1.4810)	(0.0497)	(0.1240)	(2.4150)	(0.0812)
MEC×GTI	−0.1480 **			0.0168			0.0157
	(0.0737)			(0.0166)			(0.0271)
MEC×GTFP		−5.4550			−0.2060			−1.4730
		(3.7760)			(1.4810)			(2.4140)
MEC×GI			−6.9830 ***			0.7650			0.4450
			(2.2810)			(0.4970)			(0.8120)
Constant	−3.6280	−2.1690	−7.1330	5.8800 ***	5.6830 ***	6.3130 ***	5.9310 ***	5.7340 ***	6.1150 ***
	(5.0290)	(4.9850)	(5.2140)	(1.3370)	(1.3250)	(1.3830)	(2.1820)	(2.1590)	(2.2600)
Observations	689	689	689	517	517	517	517	517	517
R-squared	0.0670	0.0650	0.0750	0.2090	0.2070	0.2110	0.0220	0.0220	0.0220
Control Variable	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Time-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
City-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Notes: this table shows estimates of moderating effects on coastal water pollution. All the variables are defined in Appendix A. The corresponding standard errors are reported in parentheses. *** and ** represent statistical significance at the 1% and 5% level, respectively.

presents the results. The interaction variables in Columns 1 and 3 indicate that GTI and GI had a positive impact on improving coastal water quality. However, the effects of GTI, GTFP, and GI were not significantly pronounced. This is primarily due to the fact that fertilizers and agricultural runoff are the major sources of active phosphate and inorganic nitrogen pollution, particularly in agriculturally dominated regions [65]. Phosphates in fertilizers are highly soluble in water, and during rainfall or irrigation, they can easily runoff into rivers and coastal waters [66].

5. Geographic Location and Environmental Scrutiny

China’s coastal economic circles each possess unique characteristics that contribute to the nation’s economic diversity and growth. To gauge the specific characteristics of each economic area, the entire area was categorized depending on geographical location into the Bohai Economic Circle, the Yangtze River Delta Economic Circle, the Pearl River Delta Economic Circle, and other economic circles (where the coastal cities located in Guangxi Province were classified into the Beibu Gulf Economic Circle, and the coastal cities in Hainan Province were classified into the West Taiwan Strait Economic Circle, collectively referred to as other economic circles).

We re-estimated Equation (1). Results are reported in Columns 1–4, Table 7. The MECDZ policy significantly facilitated industrial structural upgrading in the Pearl River Delta Economic Circle; however, its influence on industrial upgrading in the Bohai Economic Circle and the Yangtze River Delta Economic Circle proved insignificant. This is due to the differences in the modernization of the industrial structures among the three economic circles. The Pearl River Delta Economic Circle is a frontier region in China [67]. In contrast to traditional industries, high-tech sectors and modern service industries attach more importance to clean production and react more vigorously to the MECDZ policy [12,68].

We also took into account that the stringency of environmental policies influences the industrial structure. Thus, this research categorized the samples into key environmental protection cities and non-key counterparts. Findings are presented in Columns 5–6 of

Table 8. Heterogeneity.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	ISU
	Bohai Rim Economic Zone	Yangtze River Delta Economic Zone	Pearl River Delta Economic Zone	Other	EP Priority Cities	Non-Priority Cities
MEC	0.0153	0.0006	0.0417 ***	0.0108	0.0333 **	0.0041
	(0.0101)	(0.0080)	(0.0141)	(0.0144)	(0.0138)	(0.0060)
Constant	0.5320	1.1050 **	0.3870	−2.3870 ***	2.7730 ***	0.1620
	(0.8520)	(0.4450)	(0.3150)	(0.8090)	(0.6280)	(0.2070)
Control Variable	Yes	Yes	Yes	Yes	Yes	Yes
Time-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
City-fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.6950	0.7490	0.4170	0.3610	0.4410	0.6080
Observations	221	143	182	143	286	403

Notes: this table shows estimates of Equation (1). All the variables are defined in Appendix A. The corresponding standard errors are reported in parentheses. *** and ** represent statistical significance at the 1%and 5% level, respectively.

, revealing that the policy has markedly advanced industrial structural upgrading in the demonstration zones aligning with our projections.

6. Conclusions, Discussion, and Policy Implication

6.1. Conclusions

This works adopted a SDID model to estimate the impact of the MECDZ program on industrial structure upgrading. First, the MECDZ has achieved positive results in promoting industrial structure upgrading increased by 1.62%. Secondly, these findings demonstrated robustness when validated through PSM-DID methodology and alternative explained variables. Moreover, green technology innovation, green total factor productivity, and government intervention were underlying channels. Thirdly, substantial heterogeneity emerged across marine economic zones; while the policy drove significant industrial upgrading in the Pearl River Delta demonstration cities, its effects proved statistically insignificant in both the Bohai Rim and Yangtze River Delta regions. Fourthly, environmentally focused cities exhibited more pronounced policy responsiveness than non-priority counterparts. Finally, we confirmed complementary ecological benefits; as such, MECDZ implementation improved coastal water quality, with green technology innovation and government support serving as key mechanisms for this outcome.

6.2. Discussion

As to the methodologies in our work, the combined use of panel data with fixed effects, PSM-DID, heterogeneity tests, and mechanism analysis offers several strengths. First, fixed-effects models control for unobserved time-invariant confounders, improving causal inference. PSM-DID further reduces selection bias by matching treated and control groups on observable characteristics before assessing policy impacts, enhancing robustness. The heterogeneity test reveals how effects vary across regions (e.g., Pearl River Delta vs. Bohai Rim), providing nuanced policy insights. Finally, mechanism tests (e.g., green innovation, government intervention) clarify why the policy works, strengthening theoretical and practical relevance. Together, these methods provide a rigorous, multi-dimensional evaluation of MECDZ’s impact.

The MECDZ policy is a specific branch of the ecological civilization policy targeting China’s coastal cities. Implementing the MECDZ policy nationwide in a top-down manner, progressing from specific points to broader areas, holds significant potential for effectively achieving Sustainable Development Goals (SDGs) in coastal regions [69]. For instance, the successful implementation in China’s eastern regions has driven a unified advancement in China’s ecological civilization development. Moreover, China’s ecological civilization standard is the top-down policy instrument, generally considered dynamic and thus deemed more effective. These indicators come into being via ongoing adjustment and refinement in light of shifts in national strategic demands and practical goals.

The MECDZ framework offers emerging countries a sustainable growth pathway, avoiding the “pollute first, clean up later” trap while creating jobs in eco-tourism, marine biotech, and renewable energy. By leveraging global climate finance, South–South partnerships, and digital monitoring tools, nations can enhance climate resilience (e.g., mangrove restoration in Vietnam), attract ESG investment, and align with UN SDGs. However, the implementation of such policy faces challenges. This may weak enforcement capacity, high upfront costs for infrastructure like wastewater treatment, and resistance from traditional sectors like small-scale fisheries. Corruption, reliance on foreign technology, and conflicts between industrial growth and conservation goals further hinder adoption.

The positive results of the MECDZ, particularly its success in industrial upgrading, green innovation-driven productivity gains, and improved coastal water quality, strongly justify maintaining and expanding the policy. However, the heterogeneous effects across regions suggest that modifications are needed to enhance effectiveness. Authorities should tailor implementation in underperforming regions (Bohai Rim, Yangtze Delta) by adjusting industry-specific incentives (e.g., targeted subsidies for marine tech) and strengthening local enforcement to match the Pearl River Delta’s success. A place-based approach, combined with reinforced green innovation and governance, would maximize its economic and environmental impact.

Our work also suffers limitations. First, while the methodologies we adopt can address the main research question, PSM-DID may not hold if unobserved shocks differentially affect treatment and control groups. Moreover, linear mediation may oversimplify complex pathways, which offers future research directions to analyze the potential channels behind this causal relationship. Second, the mechanisms identified are broadly defined, leaving room for further exploration of specific policy instruments (e.g., subsidies, R&D incentives) that drive industrial upgrading. The study also does not quantify the long-term sustainability of these effects; short-term gains in industrial restructuring may not necessarily translate into persistent pollution reduction or economic resilience. Addressing these limitations could enhance the generalizability and policy relevance of future evaluations.

6.3. Policy Implication

Our work has some policy implications. Firstly, the MECDZ program represents a crucial initiative in promoting ecological civilization and sustainable development, particularly in coastal regions. To boost the efficacy of this policy, the policymakers should gradually expand the range of demonstration zones, with priority given to enhancing sustainable development capacity and optimizing industrial composition. This involves reducing reliance on traditional high-pollution industries and fostering the upgrading of industrial structures.

Second, the government should encourage and support GTI to provide the necessary technical guarantees and impetus for industrial upgrading. This encompasses stepping up investment in green tech projects and their development, building a sound green tech innovation system, and driving the extensive adoption of green technologies in production processes.

Finally, the government should tailor MECDZ policy initiatives in line with the distinct features and developmental requirements of varied cities, guaranteeing the policy’s effectiveness and enforcement in diverse coastal regions. By focusing on environmentally friendly high-tech industries and fostering an attractive investment climate, coastal regions can effectively leverage international resources and markets.