Will “Dual Control of the Amount and Intensity of Energy Consumption (DCEC)” Policy Increase Urban Green Competitiveness?

Abstract

Urban green transformation is an important part of global low-carbon development. Coercive government policies are critical to achieving success, but, as of now, there is no unified concept of “green competitiveness”. To address climate change, it is necessary to clearly define core concepts and apply measures. For this study, “urban green competitiveness” refers to the continuous improvement of total factor productivity by cities while fully emphasizing resource limitations and environmental issues to achieve economic growth. This article focuses on a Chinese policy of “dual control of the amount and intensity of energy consumption” (DCEC), which has not been fully studied. The DCEC policy incorporates a quasi-natural experiment to assess whether urban green competitiveness has been affected. This paper builds on the multi-period DID model and explores the effect based on city-level data from 2003 to 2019. The conclusion shows that DCEC policy can address climate change by restricting energy consumption and intensity. Success is achieved through the promotion of local green patent innovation and by intensifying environmental regulation. The policy should start from a long-term perspective by promoting structural and technological changes in the economy, rather than short-term quick success and instant benefits.

1. Introduction

Climate change is a grave challenge facing the world, and China must make efforts as a major energy consumption and carbon emission country. According to the report of the International Energy Agency (IEA), China’s carbon emissions have been the highest in the world since 2005 [1]. In 2018, China’s per capita carbon emissions exceeded the world average, accounting for 30.8% of the global total. According to Rhodium Group research, China emitted more greenhouse gases in 2019 than all developed countries combined [2]. China is trying to reverse the current situation in various ways. As a country with a unitary hierarchical political system, the policies formulated by China’s central government have extremely high coercive power and play an absolute guiding role in the behavior of local governments [3]. In recent years, China has made saving energy and reducing carbon emissions a basic state policy. Under the top design of energy conservation and emission reduction, the Chinese government launched the “dual control of amount and intensity of energy consumption (DCEC)” policy at the Fifth Plenary Session of the 18th Communist Party of China Central Committee in 2015. Setting the dual goals of reducing energy intensity and controlling total national emissions [4].

From the perspective of green transformation, the DCEC policy is a milestone, setting double goals of reducing energy intensity and total control for the first time nationwide. Particularly, it is different from the previous environmental protection policies, not only strictly setting the quota of energy consumption per unit of Gross Domestic Product (GDP), but also controlling the total amount of energy consumption. Under the constraints of this policy, each local government must subsequently set the segmentation targets at the provincial and city levels. Overachieving goals can earn incentives. On the contrary, if they do not meet the expected targets, the leadership will be punished. Therefore, local governments have strong motivations to improve the competitiveness of their jurisdictions in order to avoid punishment and make political gains [5].

In order to better measure the policy effect, this article chooses “city” as the research object. Firstly, global statistical research has found that the proportion of total carbon emissions in cities exceeds 70% generally, while it was as high as 85% in China [6]. The data at the city level are highly representative. Secondly, there are time differences in the implementation of the DCEC policy among cities, and urban data can better reflect the degree of policy impact. Furthermore, this article focuses its research perspective on urban green competitiveness (GC). The “Man and the Biosphere (MAB)” program of UNESCO proposed the key concept of “ecological city” [7], and facing many pressures such as population loss, the ecological environment is an important aspect of urban competition. Currently, scholars have paid attention to this point and proposed a definition and measurement method for urban green competitiveness [8,9,10,11,12,13,14,15,16,17,18,19,20], providing a good foundation for this study.

With the above, this paper raises two questions:

(1)

What is the impact of the DCEC policy on the urban’s green GC?

(2)

What are the trajectories through which this impact is achieved?

In order to solve the above problems, this paper establishes a multi-phase difference-in-difference model, identifies DCEC policies, and tests the impact of policies on urban-total green productivity with an optimized calculation method. At the same time, the regional heterogeneity analysis was carried out, and obvious differences were found. Subsequently, this paper decomposed the policy effect, and the DCEC policy realized its original intention: to enhance the GC of cities by reducing total energy consumption and energy intensity. The expansion research shows that the government and enterprises will enhance the GC of their cities by increasing the intensity of environmental regulations and strengthening green innovation channels, respectively.

Compared to existing literature, the marginal contribution of this article is as follows:

(1)

The effectiveness of DCEC policies has been measured, enriching current research on China’s environmental policies.

(2)

Enriched relevant research on urban GC and provided ideas for the sustainable development of cities.

(3)

Provide the practical basis for the implementation of environmental protection policies.

The remainder of this paper is organized as follows: Section 2 is the literature review. Section 3 is about research hypotheses. Section 4 is the methodology. Section 5 is data analysis. The final part is the conclusion and policy recommendations.

2. Literature Review

In this part, green competitiveness (GC) is at the core of the discussion.

2.1. Definition of GC

At present, there is no unified concept or definition for “green competitiveness”. According to differences in research directions, green competitiveness appears in multiple fields such as corporate governance, technological innovation, product marketing, and financial investment [8]. At the corporate level, Li et al. believe that the more green patents a company applies for, the stronger its green competitiveness [9]. Wu J. et al. classified green patents and believed that the higher the quality of green patents, the stronger the green competitiveness of enterprises. Ahmed defines a company’s green competitiveness as its level of green productivity [10]. Du et al. defined the green competitiveness of enterprises as their level of green total factor productivity [11]. At the urban level, the definition given in the “China Urban Dual Carbon Index 2021–2022 Research Report” is that the lower the energy consumption and carbon dioxide emissions per unit of GDP, the stronger the green competitiveness of the city. Zhao et al. believe that the level of green innovation in cities can represent their green competitiveness [12]. Some scholars have subdivided urban green competitiveness into five categories: high-quality green growth, resource conservation and emission reduction, urban ecological environment quality, green lifestyle, and green investment competitiveness [13].

2.2. How to Measure GC

Due to the lack of a unified definition, there are also differences in the measurement indicator system. The measurement indicators at the enterprise level include enterprise green patents, enterprise green total factor productivity level, and so on. At the urban level, Li and Deng set the “average annual decline rate of new energy vehicles, building energy efficiency, and energy consumption per 10,000 yuan of GDP” as measurement indicators [14]. Guan and Song established an evaluation system that includes technological level, natural endowment, environmental efficiency, and policy impact [15]. Li et al. evaluated urban green competitiveness from three perspectives: economic development level, social progress level, and ecological resources and environment [16]. Zhou and Liu assigned values to urban green competitiveness by calculating the ratio of urban GDP to industrial wastewater, industrial SO₂, and industrial dust emissions [17]. Managi et al. directly use TFP indicators to represent the level of green competitiveness [18]. In recent years, relevant research has gradually adopted green TFP indicators that consider unexpected outputs to measure the level of green competitiveness, fully considering the impact of technology level, property rights structure, capital structure, and energy structure on green total factor productivity [19,20].

2.3. Relevant Influencing Factors of GC

From the perspective of urban hierarchy, plentiful research focuses on the evaluation of green competitiveness, with only a few literature studies on what factors affect the improvement or decrease of urban green competitiveness. Huang and Zha found through their research on the “river child policy” that this policy reduces the intensity of urban pollution emissions and enhances urban green competitiveness [20]. Zhou et al. believe that foreign direct investment has enhanced urban green competitiveness through channels such as knowledge spillover, human capital optimization, and strengthening environmental regulations [21]. Sun and Lin measured data from 30 provinces and cities in China, believing that green finance can enhance green competitiveness [22].

Enterprises play a crucial role in urban development, and controlling industrial emissions is also an important part of urban green transformation. Expanding the study of influencing factors to the enterprise level provides a useful reference for studying urban green competitiveness. The research on the influencing factors of green competitiveness can be divided into two categories: internal and external. The external influencing factors are mainly environmental regulations, among which Du et al. believe that adopting appropriate environmental regulations will strengthen the innovation ability of enterprises, thereby reducing costs and increasing efficiency, ultimately achieving the goal of shaping green competitiveness. Among the internal influencing factors, entrepreneurial spirit, corporate culture, the industry in which the enterprise is located, enterprise management mode, and enterprise scale can all have a significant impact on the formation of the green competitiveness of the enterprise [11]. Henrique and Sadorsky showed that as the industry in which a company operates changes, its environmental awareness also changes accordingly. Enterprises with higher environmental relevance, such as resource-based enterprises, tend to have higher environmental awareness, while enterprises engaged in the service industry tend to have weaker green concepts compared to them [23].

2.4. Knowledge Gap

The academic community has accumulated rich achievements in green competitiveness, providing a research foundation for this study. (1) It is necessary to define CG and measure it. This article believes that “urban green competitiveness” refers to the continuous improvement of total factor productivity by cities while fully emphasizing resource limitations and environmental issues in order to achieve economic growth. In the indicator measurement, we selected green total factor productivity, considering non-expected output, as the proxy indicator to measure urban green competitiveness. (2) There is still great room for expansion in the research on the impact of urban green competitiveness. From the perspective of China’s political system, research on evaluating the impact of national policies on urban green competitiveness is still far from sufficient. Since its implementation in 2015, the DCEC policy has not received sufficient research. Exploring the impact of this policy on green competitiveness has theoretical value and practical significance.

3. Research Hypotheses

As has been said, the implementation of the “DCEC” policy has increased the attention of local governments to environmental issues, making environmental governance an important standard for evaluating the legitimacy of enterprises within the jurisdiction. This has generated environmental regulatory pressure on enterprises, prompting them to change their environmental protection strategies. Signal theory helps to explain two policy choices for enterprises when facing government pressure: (1) motivation to gain benefit; (2) motivation to avoid punishment [24]. To gain more benefit, after feeling local governments’ efforts in environmental governance, enterprises will increase investment in green innovation. Ayalew and Jeevan believe that, from the perspective of green productivity and green chemistry in enterprises, increasing investment in green production can greatly increase the current development of enterprises [25]. This has led to more enterprises increasing their investment in green innovation, thereby entering a virtuous cycle of green energy consumption. Since the implementation of the “DCEC” policy, especially the strict control of total energy consumption, local governments have had to undertake more severe environmental governance tasks. Enterprises actively engaged in green innovation, reduced environmental compliance costs, and obtained innovation compensation [26]. To avoid punishment, enterprises will consider changing their environmental strategies to cope with environmental institutional pressures from the government [27]. Regardless of the motivation of the enterprise, it can both achieve the goal of reducing environmental pollution and improving the level of regional green development.

Since the 1980s, due to the absolute priority of economic development, local officials have developed a “political championship” system that focuses on economic performance [5]. However, with the implementation of the “DCEC” policy, local governments must reconsider an important issue: how to restore the damaged ecological environment while ensuring economic development and putting environmental protection goals in a more important position [28]. Therefore, after the implementation of the “DCEC” policy, local governments have the motivation to adopt stronger ER measures against enterprises and exert environmental legitimacy pressure. “DCEC” policy, like other environmental policies, puts pressure on enterprises for the implementation of specific standards rather than the policy itself. Pan et al. found that environmental policies can promote the green development of enterprises by strengthening environmental regulation (ER) efforts [29]. Xu et al. found that the level of pollution in enterprises can affect the current development process of enterprises [30]. At the same time, combined with the local environmental pollution level of ER, this will constrain the enterprise, which will slow down the development process of the enterprise. Therefore, the green development of enterprises is an important factor in current development. ER is an important means for the implementation of government environmental policies. After the implementation of the DCEC policy, it will increase the environmental governance pressure on local governments and encourage them to increase their ER intensity. In turn, it forced enterprises facing the hard constraints of ER to adopt a green technology approach, reduce total emissions, and ultimately increase regional GC.

Based on the above analysis, the following theoretical assumptions are proposed in this study:

H1. 

The implementation of the “DCEC” policy will increase urban GC.

Based on the policy connotation of “DCEC”, the following two sub-propositions are proposed in this study:

H1a. 

The implementation of the “DCEC” policy will increase urban GC by controlling total energy consumption.

H1b. 

The implementation of “DCEC” policy will increase urban GC by controlling energy consumption intensity.

Meanwhile, based on the previous analysis, the following two possible mechanisms of action are proposed in this study:

H2. 

“DCEC” policy improves urban GC by promoting local green patent innovation.

H3. 

“DCEC” policy promotes the improvement of urban GC by increasing ER intensity.

4. Methodology

4.1. Metrological Model Setting

The implementation time of “DCEC” policy in different cities is not consistent, so we must adopt a multi-phase difference-in-difference (DID) model [31,32,33] to explore the impact of “DCEC” policy on urban GC.

$${lngc}_{it}={\alpha }_0+{\alpha }_1{Treat}_{it} \ast {Post}_{it}+{\alpha }_2{Treat}_{it}+{\alpha }_3{Post}_{it}+\beta X_{it}+{\mu }_{it}$$

${lngc}_{it}$ represents the GC of city i in year t, and the specific settings of the variables are described later. ${Treat}_{it}$ is a virtual variable representing the processing group. If city i is in the processing group, then ${Treat}_{it}$ = 1, otherwise it is 0. ${Post}_{it}$ is a virtual variable of “DCEC” policy (It must be emphasized here that the DCEC policy was disclosed at the national level in 2015, and the implementation time of each city is not uniform from 2015. Therefore, we sorted out the time when each city started to implement the DCEC policy, which is also the reason for choosing the multi-period DID model). $X_{it}$ is the set of control variables.

4.2. Variable Setting

4.2.1. Green Competitiveness

The measurement of green competitiveness (GC) is the core part of this article. ${gc}_{it}$ represents the GC of the city i in year t. There are two main types of methods used in the existing literature to measure GC [34]. Early studies often directly used TFP indicators to represent GC levels [18]. In recent years, relevant studies have gradually adopted green TFP indicators that consider unexpected outputs to measure GC levels [35,36]. Considering the calculation method comprehensively, this article uses a directional distance function based on a relaxation value calculation and the Luenberger productivity index to calculate urban green total factor productivity (TFP) and uses this index to represent urban GC.

Measurement method: If $x=x_{ij}\in R_M^{+}$ is the factor input vector, as this model includes both expected and unexpected outputs, then the expected output vector included in the model can be assumed to be $y=y_{ij}\in R_N^{+}$ and the unexpected output to be $p\in R_l^{+}$. If the production possibility set $\left(x_{k,t},y_{k,t},p_{k,t}\right)$ in the region during the period is both closed and convex, the data envelopment analysis (DEA) method [37] is used to model it as follows:

$$F(x)=\left\{\begin{array}{l}\left(y^{\prime },p^t\right):{\displaystyle \sum _{k=1}^K{\gamma }_k^t}{y}_{km}^{\prime }⩾y_{km}^{\prime },\forall m;\hfill \\ {\displaystyle \sum _{k=1}^K{\gamma }_k^{\prime }}{p}_{ki}^t⩾p_{ki}^t,\forall i;{\displaystyle \sum _{k=1}^K{\gamma }_k^{\prime }}{y}_{kn}^t⩾x_{kn}^t,\forall n;\hfill \\ {\displaystyle \sum _{k=1}^K{\gamma }_k^t}=1,{\gamma }_k^t⩾0,\forall k\hfill \end{array}\right\}$$

In the above equation, ${\gamma }_k^t$ represents the weight of the observation. ${\sum }_{k=1}^K{\gamma }_k^t=1,{\gamma }_k^t=1,\forall k$ indicates that the production function has the property of variable economies of scale.

The directional distance function based on the relaxation value calculation model considering unexpected output is:

$$\begin{array}{l}{S}_V^{t,k}\left(x_{k,t},y_{k,t}^g,y_{k,t}^p,g^x,g^y,g^p\right)=\\ {\max }_{s^{\prime },s^{\prime },s^b}\left[\frac{1}{N}{\displaystyle \sum _{n=1}^N\frac{s_n^x}{g_n^x}}+\frac{1}{M+I}\left({\displaystyle \sum _{m=1}^M\frac{s_m^y}{g_m^y}}+{\displaystyle \sum _{i=1}^I\frac{s_i^y}{g_i^y}}\right)\right]/2\\ \mathrm{s}.\mathrm{t}. {\displaystyle \sum _{k=1}^K{\gamma }_k^t}{x}_{kn}^t+s_n^x=x_{kn}^t,\forall n;{\displaystyle \sum _{k=1}^K{\gamma }_k^t}{y}_{km}^t-s_m^y\\ =y_{km}^t,\forall m;{\displaystyle \sum _{k=1}^K{\gamma }_k^t}{p}_{ki}^t-s_i^y=p_{ki}^t,\forall i;{\displaystyle \sum _{k=1}^K{\gamma }_k^t}=1,\\ {\gamma }_k^t⩾0,\forall k;s_n^x⩾0,\forall n;s_m^y⩾0,\forall m;s_i^p⩾0,\forall i\end{array}$$

The above equation, $S_V^t$ represents the distance function and includes directionality. As mentioned earlier, it has the property of variable returns to scale. So, the directional distance function of the scale return invariant property (CRS) can be represented by $S_C^t$.

As mentioned earlier, $\left(s_n^x,s_n^y,s_i^p\right)$ is always positive. Its economic meaning is marginal input and decreasing pollution, but it is always positive. So, what $(s_n^x$, $s_n^y,s_i^p)$ represents is the value of excessive use of input factors, excessive pollution emissions, and expected output below the expected level.

Chambers et al. invented the Luenberger productivity index, which can be used to consider increasing output and reducing input from any measurement angle [38]. So, the expression for GTFP and its source between $t$ and $t+1$ periods is:

$${\mathrm{GTFP}}_t^{t+1}=\frac{1}{2}\left\{\begin{array}{l}\left[S_C^t\left(x^t,y^t,b^t;g\right)-S_C^t\left(x^{t+1},y^{t+1},b^{+1};g\right)\right]+\hfill \\ \left[S_C^{t+1}\left(x^t,y^t,b^t;g\right)-S_C^{t+1}\left(x^{t+1},y^{t+1},b^{+1};g\right)\right]\hfill \end{array}\right\}$$

The investment indicators selected in this article include the estimation of urban capital stock and urban labor input using the perpetual inventory method. Among them, the depreciation rate in the measurement of capital stock refers to existing research methods and is taken as 6%. Labor input refers to the number of employees reported in each city’s yearbook.

4.2.2. DCEC Policy Identification

Regarding the identification of DCEC policies, keyword identification was conducted on the legal and regulatory database of the Peking University Treasure website based on policy documents released by various cities from 2003 to 2019 (considering the impact of COVID-19, this paper has limited the time frame to 2019 to avoid disruption). The search keywords include “DCEC”, “dual control of total energy consumption and intensity”, “total energy consumption”, “energy consumption intensity”, etc. Based on this, the “DCEC” policy dummy variable ${Post}_{it}$ was constructed. During the sample observation period, values were assigned to each province based on the introduction time t of energy policies. If the province has issued an “DCEC” policy after t, ${Post}_{it}$ = 1, otherwise, ${Post}_{it}$ = 0.

4.2.3. Control Variable

(1) The total urban economic output (lngdp_it) is expressed as the logarithm of the actual GDP of region i for the year. (2) The regional economic level (lnpgdp_it) is measured as the logarithm of the current per capita GDP of region i. Intuitively, the higher the level of regional economic development, the more stringent environmental policies and stronger GC are. (3) The level of technological research and development (lnrd_it) is measured by the natural logarithm of the amount of research and development expenditure in the city’s public expenditure for the year. (4) According to the practice in Shao’s research [39], opening to the outside world (open_it) is measured by the proportion of FDI to GDP in region i in the current year. (5) Urbanization (Urban_it) is represented by the proportion of urbanization in region i and period t. And the regression also controls for urban and temporal fixed effects.

Table 1. Descriptive Statistics.

Variable	Observed Value	Mean Value	Standard Deviation	Minimum Value	Maximum Value
lngc	4396	0.0091607	0.0287041	−0.0407327	0.0646766
post	4396	0.0552775	0.2285472	0	1
lngdp	4396	15.47772	2.299455	5.442418	19.60485
lnpgdp	4396	10.18896	0.8223695	4.59512	13.05569
lnrd	4396	9.238224	1.919787	−2.040221	15.52928
open	4396	0.0008	0.005	0	0.0856
Urban	4396	0.4859318	0.1711856	0.1117	1

reports the descriptive statistical results of the main variables.

4.3. Data Sources

This article mainly uses unbalanced panel data from 278 cities in China from 2003 to 2019 for analysis. The raw data mainly comes from relevant years such as the China Industrial Statistical Yearbook, China Environmental Yearbook, China Statistical Yearbook, China Energy Statistical Yearbook, China Urban Statistical Yearbook, Peking University’s Treasure website, and provincial statistical yearbooks. A total of 4396 observations were obtained.

5. Empirical Analysis

5.1. Benchmark Regression

Table 2. The Impact of “DCEC” Policy on Urban GC.

Variable	lngcit	lngcit	lngcit
	1	2	3
Treat*post	0.007 ** (2.05)	0.007 ** (2.09)	0.008 ** (2.17)
Treat	0.001 (0.85)	0.001 (1.05)	0.004 (0.49)
post	−0.005 * (−1.71)	−0.005 * (−1.71)	−0.005 * (−1.73)
lngdpit	/	−0.0001 (-0.94)	−0.002 ** (−2.39)
lnpgdpit	/	−0.003 *** (-3.66)	−0.002 * (−1.87)
lnrdit	/	0.001 ** (2.05)	0.002 ** (2.45)
openit	/	−0.182 * (−1.79)	−0.178 * (−1.69)
Urbanit	/	0.012 *** (3.36)	0.014 *** (3.40)
Control year	No	No	Yes
Control area	No	No	Yes
Observed value	4396	4209	4209

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

reports the baseline measurement estimation results. According to the results from columns 1 to 3 in Table 2, the control variable is gradually added, and the fixed effects of the year and region are controlled. The explanatory variable “DCEC” policy implementation (Treat*post) coefficient in each column is significantly positive. This confirms that the “DCEC” policy significantly promotes the improvement of urban GC levels. After gradually adding control variables, the significance and sign of the Treat*post coefficient did not change significantly, indicating that “DCEC” policy significantly promoted the improvement of urban GC levels. Hypothesis 1 has been validated in the previous text.

In the control variable, the economic scale (lngdp) and urban GC change in the opposite direction, which may be because regions with larger economic scales often have earlier economic development starts. These regions introduced many high-emission and energy-consuming industries in their early economic development. In the current economic transformation, these enterprises have not been timely transformed, upgraded, or eliminated, resulting in a significant negative correlation between this variable and the dependent variable. The improvement of economic development level (lnpgdp_it) also inhibits the development of urban GC. Khan et al. found that the improvement of economic development levels is not conducive to the reduction of regional energy consumption and emissions to some extent [40]. In areas with higher economic levels, business operating costs are higher, the relative proportion of energy expenditure in the total cost is reduced, and enterprises may ignore the benefits brought by saving energy.

The coefficient of technological research and development level (lnrd_it) is significantly positive, indicating that the higher the level of urban scientific research, the stronger the urban GC. Zhang et al. found that research and development investment not only has a significant promoting effect on green development in the local area but also, to some extent, drives GC improvement in the surrounding areas [41]. The results in Table 1 indirectly confirm this conclusion. The level of urbanization is significantly negatively correlated with low-carbon economic growth. The study by Ullah Ket al. indicates that further urbanization will have a negative impact on the low-carbon economy when the level of urbanization reaches a certain level [42].

5.2. Regional Heterogeneity Analysis

To investigate the impact of the “DCEC” policy on the GC development of cities in different regions, a classification regression analysis was conducted on all sample cities in the eastern region and the central-western regions of China are listed in

Table 3. The impact of “DCEC” policy on urban GC in different regions.

Variable	lngcit	lngcit
	Eastern	Central-Western
Treat*post	0.014 ** (2.18)	0.010 ** (2.38)
Treat	0.001 (0.18)	0.004 (1.08)
post	−0.002 (−0.26)	−0.005 * (−1.71)
lngdpit	0.001 (1.06)	−0.001 (−1.04)
lnpgdpit	−0.0001 (−0.18)	0.001 (0.79)
lnrdit	−0.001 (−1.13)	−0.0001 (−0.19)
lnopenit	0.084 (0.49)	−0.183 * (−1.66)
Urbanit	−0.001 (−0.21)	−0.004 (−0.89)
Control year	Yes	Yes
Control area	Yes	Yes
Observed value	1794	2415

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

. The eastern region mentioned in this article includes twelve provinces, autonomous regions, and municipalities directly under the central government, including Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Guangxi, and Hainan. Different scholars have adopted different criteria for the division of provinces in eastern China. The standard based on this paper is the division method in the “Seventh Five-Year Plan” adopted by the fourth session of the Sixth National People’s Congress in 1986. The current legal basis for this division is the Urban and Rural Planning Law of the People’s Republic of China. The central-western regions include all regions except the above-mentioned provinces.

In Table 3, the implementation of the “DCEC” policy has a significant positive impact on the improvement of GC in both eastern and central western cities. But the difference is that this impact seems to be greater in eastern cities (0.014 > 0.010). The author believes that the reason for this result may be the early development of eastern cities. Many high-emission and energy-consuming industries were introduced in the early stages of economic development. In the current economic transformation, these enterprises have not been timely transformed, upgraded, or eliminated from production. When facing the newly introduced “DCEC” policy, the adjustment efforts for such cities are greater, which leads to a greater increase in their GC compared to cities in the central and western regions.

5.3. Robustness Test

5.3.1. Placebo Test

To avoid the impact of the “DCEC” policy on urban GC being driven by some unobservable random factors, this article further conducted a placebo test based on the benchmark regression. The core explanatory variables corresponding to each benchmark regression were extracted and disrupted in this study and then randomly matched with the dependent variables for regression, constructing a pseudo-policy time point. After repeating the above process 1000 times, the placebo test results in Figure 1, Figure 2 and Figure 3 were obtained.

In Figure 1, by constructing pseudo-policy time points, the mean of the kernel density estimates of the coefficients after randomizing the core explanatory variable is close to 0 (as shown by the dashed line in Figure 1), which is significantly different from the true value of 0.008. (The true value of 0.008 in the Figure 1 solid line is shown in Table 2, Column 3, the coefficient of variable Treat*post) (as shown by the solid line in Figure 1). In Figure 2, after constructing the pseudo-policy time point, the t-test value approaches 0, which is significantly different from its true value of 2.17. (The true value of the Figure 2 solid line is shown in Table 2, Column 3, the t-value of variable Treat*post, in the bracket). In Figure 3, the coefficient p values obtained by constructing a random core to explain non-acceptance regression through pseudo-policy time points are generally greater than 0.1 (please observe the horizontal dotted line). (The longitudinal solid line in Figure 3 has the same meaning as Figure 1’s solid line), indicating that at the 10% significance level, most coefficients become no longer significant. There are also two points worth noting: First, randomization results in most scatter points approaching zero, indicating that the true coefficient value becomes an outlier in this case. Secondly, at the 10% significance level, these near-zero scatter points are not significant. The above test indicates that the impact of “DCEC” policy on urban GC is not driven by other unobservable factors (or missing variables). The effectiveness of the DCEC policy was confirmed.

5.3.2. Endogeneity Problem

To control for potential endogeneity issues in variables, this article will delay all control variables by one period and conduct regression again in

Table 4. The impact of “DCEC” policy on urban GC.

Variable	lngcit	lngcit	lngcit
	1	2	3
Treat*post	0.007 ** (2.03)	0.007 ** (2.19)	0.008 ** (2.07)
Treat	0.001 (0.74)	0.001 (1.32)	0.004 (0.55)
post	−0.005 * (−1.75)	−0.005 * (−1.75)	−0.005 * (−1.77)
lngdpit-1	/	−0.002 (−0.88)	−0.002 ** (−2.19)
lnpgdpit-1	/	−0.001 *** (−3.83)	−0.007 * (−1.81)
lnrdit-1	/	0.001 ** (2.02)	0.002 ** (2.37)
openit-1	/	−0.088 * (−1.77)	−0.075 * (−1.72)
Urbanit-1	/	0.005 *** (3.73)	0.004 *** (3.45)
Control year	No	No	Yes
Control area	No	No	Yes
Observed value	3813	3813	3813

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

.

After one period of lag in the control variables, there was no significant change in the positive or negative regression coefficients or the significance of each control variable. However, due to a delay of one period, the absolute values of individual control variable coefficients have decreased, which once again verifies the robustness of the conclusions in this paper.

5.4. Mechanism Analysis

To further verify the previous hypothesis and explore the impact path of “DCEC” policy on urban GC, this article first decomposes the effectiveness of “DCEC” policy. And this article examines the possible pathways through which “DCEC” policy affects urban GC by constructing a mediation effect model. The following impact mechanism tests were conducted.

5.4.1. Decomposition of the Effects of “DCEC” Policy

The core connotation of “DCEC” policy is to control the two indicators of EC total quantity and EC intensity simultaneously in economic and social activities. Based on the understanding of the connotation of this policy and the previous analysis, “DCEC” is first decomposed in terms of total energy consumption and EC intensity. According to Wang et al., the total EC amount and intensity of each city within the sample were calculated [43]. Among them, the total amount of EC (lntotal_energy _con_it) is the energy consumption of city i in year t, which is converted into ten thousand tons of standard coal. The EC intensity (lnenergy _con_gdp_it) is the energy consumption per unit of GDP in city i in year t, which is converted into ten thousand tons of standard coal. The following table reports on the impact of the “DCEC” policy on reducing total energy consumption and whether this impact has further impacted urban GC.

In

Table 5. Mechanism analysis—“DCEC” policy affects urban GC by reducing total energy consumption.

Variable	lngcit	lntotal_energy _conit	lngcit
	1	2	3
lntotal_energy _conit	/	/	−0.003 *** (−4.59)
Treat*post	0.007 * (1.74)	−0.070 * (−1.81)	0.003 (0.80)
Treat	0.008 (0.92)	0.919 (1.57)	0.013 * (1.69)
post	−0.005 * (−1.73)	0.042 (1.08)	−0.005 * (−1.88)
lngdpit	0.002 (0.34)	0.636 *** (13.78)	0.001 (0.57)
lnpgdpit	0.001 (0.04)	0.035 (0.85)	−0.002 (−0.04)
lnrdit	−0.001 (−1.02)	−0.014 (−1.07)	−0.001 (−1.04)
lnopenit	−0.138 (−1.29)	−3.542 ** (−2.57)	−0.120 (−1.25)
Urbanit	−0.001 (−0.17)	0.890 *** (7.88)	0.001 (0.20)
Control year	Yes	Yes	Yes
Control area	Yes	Yes	Yes
Observed value	3427	3427	3427

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

, the coefficient of the cross term (Treat*post) in column 2 is significantly negative, indicating that the implementation of the “DCEC” policy significantly suppressed the total urban EC. And from column 3, the coefficient of lntotal_energy _con_it is significantly negative, which indicates that the implementation of the “DCEC” policy has promoted the improvement of urban GC by suppressing the total energy consumption in cities, confirming the hypothesis H1a mentioned earlier.

Furthermore, this study examined whether the implementation of the “DCEC” policy had a significant impact on the energy consumption per unit of urban GDP and whether this impact promoted the improvement of urban GC. The following table reports this result.

In

Table 6. Mechanism analysis—“DCEC” policy affects urban GC by reducing energy consumption intensity.

Variable	lngcit	lnenergy _con_gdpit	lngcit
	1	2	3
lnenergy _con_gdpit	/	/	−0.003 *** (−4.34)
Treat*post	0.007 * (1.74)	−0.073 * (−1.75)	0.003 (0.79)
Treat	0.008 (0.92)	0.476 (0.91)	0.013 * (1.66)
post	−0.005 * (−1.73)	0.038 (1.02)	−0.005 * (−1.88)
lngdpit	0.002 (0.34)	−0.281 *** (−6.63)	−0.002 ** (−2.26)
lnpgdpit	0.001 (0.04)	0.140 *** (3.61)	−0.001 (−0.05)
lnrdit	−0.001 (−1.02)	−0.014 (−1.14)	−0.001 (−0.99)
lnopenit	−0.138 (−1.29)	1.164 (0.89)	−0.105 (−1.09)
Urbanit	−0.001 (−0.17)	0.898 *** (8.44)	0.001 (0.09)
Control year	Yes	Yes	Yes
Control area	Yes	Yes	Yes
Observed value	3427	3427	3427

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

, the coefficient of the cross term (Treat*post) in column 2 is significantly negative, indicating that the implementation of the “DCEC” policy significantly suppresses urban EC intensity and reduces the energy consumption required per unit of GDP production. And according to the results in column 3, the coefficient of EC strength lnenergy _con_gdp_it is significantly negative, indicating that the implementation of the “DCEC” policy promotes the improvement of urban GC by suppressing the urban EC intensity. The Hypothesis H1b mentioned earlier has been verified. Compared with the cross-term “Treat*post” in Table 5 and Table 6, we found that reducing energy consumption intensity has a slightly higher effect (−0.073) on urban green competitiveness than reducing energy consumption quantity (−0.070).

5.4.2. Green Patent Innovation

Based on the theoretical analysis mentioned earlier, this study proposes the Hypothesis H2 that “DCEC” policy will promote the improvement of urban GC by promoting urban green patent innovation. The World Intellectual Property Organization (WIPO) has detailed which patents belong to green patents. This article refers to the classification standard of the World Intellectual Property Organization and calculates the number of green patent applications in sample cities, taking the natural logarithm as lngreen_patent_it.

Table 7. Mechanism analysis—“DCEC” policy affects urban GC by promoting green patents.

Variable	lngcit	lngreen_patentit	lngcit
	1	2	3
lngreen_patentit	/	/	0.003 *** (4.67)
Treat*post	0.008 ** (2.11)	0.250 *** (2.70)	0.004 (1.28)
Treat	−0.002 (−0.22)	0.687 *** (3.56)	0.005 (0.72)
post	−0.005 * (−1.73)	−0.127 * (−1.76)	−0.005 ** (−2.12)
lngdpit	−0.002 * (−1.79)	−0.018 (−0.66)	−0.001 (−0.46)
lnpgdpit	−0.003 ** (−2.26)	−0.020 (−0.63)	0.002 (1.34)
lnrdit	0.002 *** (2.82)	0.018 (1.07)	−0.001 (−0.88)
lnopenit	−0.001 (−0.67)	−0.031 (−0.01)	−0.116 (−1.23)
Urbanit	0.014 *** (3.09)	−0.049 (−0.48)	−0.004 (−1.02)
Control year	Yes	Yes	Yes
Control area	Yes	Yes	Yes
Observed value	2294	2294	2294

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

reports the results of this test.

According to Table 7, both the coefficients of lngreen_patent_it and lngc_it in columns 2 and 3 are significantly positive, indicating that the implementation of the “DCEC” policy has promoted urban green patent innovation, which further promotes the improvement of urban GC. The mediating effect is established. The above conclusion validates Hypothesis 2: “DCEC” policy enhances urban GC by promoting local green patent innovation.

5.4.3. Environmental Regulation (ER)

Based on the previous assumption, local governments have the motivation to impose stronger ER measures against enterprises with the implementation of the “DCEC” policy. This has forced enterprises facing the hard constraints of ER to adopt a green technology approach, reduce total emissions, and ultimately increase regional GC. Referring to the method of Hao et al. [44], keywords related to environmental regulations were extracted from government work reports of various cities, and the frequency and proportion of these keywords were used to measure ER intensity. The ER variable is defined as lner_it, and

Table 8. Mechanism analysis—“DCEC” policy affects urban GC by strengthening ER.

Variable	lngcit	lnerit	lngcit
	1	2	3
lnerit	/	/	0.003 *** (2.64)
Treat*post	0.008 ** (2.12)	0.150 *** (2.82)	0.004 (1.05)
Treat	0.010 (1.27)	−0.146 (−0.79)	0.011 (1.27)
post	−0.005 * (−1.80)	0.033 (0.79)	−0.006 ** (−1.99)
lngdpit	−0.001 (−0.36)	−0.036 * (−1.65)	−0.001 (−0.08)
lnpgdpit	0.001 (0.67)	0.092 *** (3.65)	0.001 (0.58)
lnrdit	−0.001 (−0.87)	0.032 *** (2.73)	−0.001 (−1.11)
lnopenit	−0.138 (−1.30)	0.056 (0.04)	−0.149 (−1.27)
Urbanit	−0.002 (−0.48)	0.107 (1.39)	−0.002 (−0.54)
Control year	Yes	Yes	Yes
Control area	Yes	Yes	Yes
Observed value	3953	3953	3953

Note: ***, **, * respectively represent significant at the 1%, 5%, and 10% levels, with t-statistics in parentheses, the same below.

reports the results of this test:

According to Table 8, the lner_it coefficients in column 2 and lngc_it coefficients in column 3 are significantly positive, indicating that the implementation of “DCEC” policy has strengthened local governments’ ER. The stronger ER has prompted enterprises to implement green transformation, ultimately driving the improvement of urban GC. The mediating effect is established. The above conclusion validates Hypothesis 3: “DCECC” policy promotes the improvement of urban GC by increasing ER intensity.

6. Conclusions and Policy Recommendations

This article takes the proposal of the “DCEC” policy as the background and explores the impact and mechanism of the “DCEC” policy on urban green competitiveness (GC) by measuring the GC levels of 278 cities in China from 2003 to 2019. The main conclusions are as follows:

(1)

The implementation of the “DCEC” policy has significantly improved urban GC, and this improvement effect varies in different regions but remains significant. Specifically, “DCEC” policy has a greater impact on improving GC in cities in the eastern region than in cities in the central and western regions. The differences in this impact are determined by the development stages and paths of cities in different regions.

(2)

Mechanism testing verified that “DCEC” policy has a promoting effect on urban GC by controlling the total and intensity of energy consumption and decomposing the effectiveness of “DCEC” policy. Further testing indicates that the impact of the “DCEC” policy on urban GC was achieved through two trajectories: promoting urban green patent innovation and strengthening local government ER intensity. Specifically, with the implementation of the “DCEC” policy, enterprises have actively carried out green patent innovation, and local governments have implemented stronger ER levels to promote the green transformation of enterprises, ultimately improving urban GC.

This article explores the impact and mechanism of “DCEC” policy on GC, providing empirical evidence for policy implementation. Today, as the economy enters a new stage of development, the implementation of the “DCEC” policy is another innovative exploration of China’s environmental protection policies. This article helps to enrich the relevant literature on the practical effects of “DCEC” policy in theory and provides empirical support to improve the “DCEC” policy in practice. Through this article, the following policy implications have been obtained:

Firstly, the implementation and continuous improvement of “DCEC” policy should start from a long-term perspective, promoting structural and technological changes in the economy rather than short-term “quick success and instant benefits”. The institutional arrangement of “DCEC” policy should pay more attention to the factors of green technology innovation to achieve greater policy effectiveness.

Secondly, local governments should actively provide ER tools that are compatible with “DCEC” policy. This article found that an important way for “DCEC” policy to promote the development of urban GC is through the ER implemented by local governments in this process. Therefore, local governments should actively build a scientific green development system and cooperate with specific regulatory tools to promote the implementation of the “DCEC” policy. A multi-pronged approach can promote the comprehensive improvement of urban GC.

Thirdly, specific measures to implement the “DCEC” policy should be formulated based on the economic development characteristics of different regions and tailored to local conditions. In future institutional arrangements, the central government should consider the different levels of economic development and industrial structure realities in the eastern and central western regions and formulate energy consumption control measures one by one. Ultimately, the comprehensive improvement of GC levels across the country will be achieved.