Promoting “NEVs Pilot Policy” as an Effective Way for Reducing Urban Transport Carbon Emissions: Empirical Evidence from China

Abstract

To reduce urban transport carbon emissions, the Chinese government issued the NEVs Pilot Policy in 2009 to promote NEVs in both the public and private transport sectors. Using panel data from 53 prefecture-level cities in China spanning from 2006 to 2020, this study evaluates the effectiveness of the NEVs Pilot Policy on urban transport carbon emissions based on the difference-in-differences (DID) model. Additionally, it analyzes the influencing mechanism of this policy, as well as the varying policy effect among heterogeneous cities. The empirical results show that the NEVs Pilot Policy has effectively reduced urban transport emissions by an annual average of 29.3%. Annual per capita emissions were lowered by an average of 0.31 t, and the annual emission intensity was also reduced by an average of 2.04 t per unit GDP. We also found that its dynamic effectiveness has lagged but cumulatively increased over time. Furthermore, the mechanism analysis indicates that the policy effect is mainly achieved by adjusting the vehicle structure (VS) and lowering the energy intensity (EI). The heterogeneity analysis also reveals that the effectiveness of NEVs Pilot Policy varies significantly among different cities. The economic level, the political status, and the urban transport development are found to be the key factors that determine its effectiveness. Based on these findings, this study proposes some targeted policy suggestions to promote NEVs in different cities.

1. Introduction

Climate warming caused by increasing concentrations of carbon emission has become a severe global issue [1]. Thus, controlling carbon emissions could be the consensus of most countries in the world [2]. As the world’s largest carbon emitter, China is facing huge pressure to reduce its carbon emissions [3]. Under the framework of The Paris Agreement, the Chinese government put forward its Intended Nationally Determined Contribution (INDC) targets, known as “dual control”: in terms of quantity, to achieve the carbon emission peak by 2030 and strive to peak early; in terms of intensity, to lower carbon emissions per unit of gross domestic product (GDP) by 60% to 65% from the level of 2005 by 2030. The transport sector is the third largest carbon emitter in China after the energy production sector and industrial sector [4]. In 2020, it exceeded 110 million tons, accounting for over 11% of the national total. Particularly, with the dramatic growth of urbanization, China’s urban transport carbon emissions probably continue to rise rapidly in the future [5]. Therefore, reducing China’s urban transport carbon emissions is extremely significant to achieve the carbon peaking and carbon-neutral goals [6].

New energy vehicles (NEVs), as an alternative to conventional internal combustion engine vehicles (ICEVs), provide solutions to reduce transport carbon emissions [7]. NEVs, such as battery electric vehicles (BEVs), hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs), and full cell electric vehicles (FCEVs), are powered by alternative energy sources instead of traditional fossil fuels. These vehicles are considered more environmentally friendly compared with ICEVs. Although NEVs are an important measure to achieve energy conservation and carbon emission reduction, the market share of NEVs is generally low across countries. They still face certain challenges for large-scale promotion [8]. Some previous literature has found that many barriers, such as the much higher initial cost [9,10,11], insufficient charging stations [12,13], long charge times [14,15], technology shortcomings [16,17], and vehicle safety concerns [18,19], hindered consumers (including individuals, households, and corporates) from adopting NEVs. To cope with these barriers, a series of NEVs promotion policies (both in public transport and private sectors) have been promulgated in different countries all over the world [20,21,22]. For example, the direct financial incentives include direct investments in research and development (R&D) or infrastructure [23], subsidies for NEVs purchase and related infrastructure construction, the tax and registration fee exemption [24], free parking, etc. The indirect incentives include high-occupancy vehicle (HOV) access, dense charging station establishment, restrictions on automotive manufacturers (e.g., the emission targets for new vehicles), road priority, having access to restricted traffic zones, etc. [25]. Among these incentives, subsidies for NEV purchases have been widely adopted in most countries and regions, such as in North America, Asia, and Europe [26], which are considered one of the most effective NEVs promotion policies [27]. According to Azarafshar’s research, about 35% of the NEVs sales could be attributed to the purchase subsidy in Canada’s rebating provinces [28]. A survey showed that the subsidies for NEVs purchase in the UK had significantly and positively impacted the new registrations of NEVs [11].

In China, a similar situation is that the Chinese government has promulgated and implemented a series of incentive policies to boost the penetration of NEVs. Among these incentive policies, the NEVs Pilot Policy has the most extensive coverage in China, which has generated the greatest impact on NEVs’ promotion and the development of the new energy automobile industry [29]. The NEVs Pilot Policy consists of two parts: (1) The NEVs Pilot Policy has been implemented in 25 cities from 2009 to 2010, which focused on the public sectors (NEVs_Public). The central government provided subsidies for NEV purchase in the public sectors of these 25 cities, and the local governments mainly provided subsidies for constructing and maintaining the related supporting facilities of NEVs, such as charging stations, and battery repurchases. (2) The NEVs Pilot Policy has been implemented in 6 cities selected from these 25 cities in 2010, which focused on the private transport sector [30,31] (NEVs_Private). The central government provided one-time subsidies for private individuals to purchase, register, and use NEVs, and the local governments also provided subsidies for some related supporting facilities. It is noted that the NEVs_Public has promoted and influenced NEVs_Private. The large-scale promotion of NEVs_Public reflects the Chinese government’s determination to promote NEVs and provides a favorable policy situation for NEVs. At the same time, private NEVs users could also indirectly benefit from the construction and maintenance of related supporting facilities for NEVs. All of this could promote the adoption of NEVs in the private transport sector.

The NEVs Pilot Policy has been implemented for more than 10 years in China. By evaluating the effectiveness of the NEVs Pilot Policy in China, this paper aims to answer the following important questions: (1) Did the NEVs Pilot Policy have a significant effect on reducing urban transport carbon emissions? (2) What were the key factors and mechanisms that influenced the effectiveness of the NEVs Pilot Policy? (3) Was there any difference in the effectiveness of the NEVs Pilot Policy among different cities?” This paper adopted the difference-in-differences (DID) model to quantify the dynamic effectiveness of the NEVs Pilot Policy on reducing urban transport carbon emissions, based on panel data of 53 cities in China from 2006 to 2020. The DID method is a widely used method for policy effect estimation. Compared with the traditional policy evaluation method mainly through the setting of dummy variables for regression analysis, the DID method can evaluate the policy effect more accurately through the establishment of quasi-natural experiment [32]. The DID method considers the policy intervention as independent variables to effectively control the interaction effects between independent and dependent variables and avoid endogeneity problems [33]. On this basis, this paper analyzed the influencing mechanism of the NEVs Pilot Policy using the mediating effect model. Finally, this paper investigated the heterogeneity of the policy’s effectiveness on the urban transport carbon emissions across different cities and proposed policy recommendations for further promotion of NEVs in China.

The potential main contributions of this study are as follows: firstly, this study conducts a detailed empirical investigation based on panel data of 53 prefecture cities in China from 2006 to 2020. The DID model is applied to evaluate the effectiveness of the NEVs Pilot Policy in reducing urban transport carbon emissions. This is more targeted than the comprehensive evaluation of the environmental effects and makes up for the lack of carbon emission reduction efficiency in the NEVs policies research literature.

Secondly, this study evaluates the effectiveness of the NEVs Pilot Policy on urban transport carbon emissions at aggregated, per capita, and intensity levels, allowing for a more comprehensive and accurate measurement of policy effectiveness. Additionally, the time-varying DID model is adopted to evaluate the cumulative and dynamic effectiveness of the NEVs Pilot Policy on urban transport emissions over time, while most previous studies provide static results rather than possible dynamic effects of NEVs policies. This estimated policy effect exhibited lagging but cumulative increasing patterns, providing valuable insights for policymakers to adjust and formulate more targeted NEVs policies accordingly.

Thirdly, this study analyzes the heterogeneity of the cities from the perspective of the economic level, the political status, and the urban transport development and investigates the cause of the heterogeneous policy effect. Previous literature has lacked intensive study on the heterogeneity among cities. Moreover, based on the heterogeneity results, the analysis of this study is more objective and in-depth and proposes targeted policy suggestions to promote NEVs in different cities.

Finally, to show the robustness and validity of this study, we compared the effect of this NEVs Pilot Policy with China’s other policies (including the Emission Trading Scheme (ETS) and the New Energy Demonstration City (NEDC) Policy).

The remainder of this paper is organized as follows: Section 2 reviews the relevant literature; Section 3 describes the data and methodology including the sample data and econometric model; Section 4 is the analysis and discussion of the empirical results; Section 5 is the conclusion and policy suggestions of this paper.

2. Literature Review

In recent years, NEVs, with their effects on resource conservation and environmental protection, have attracted increasing attention from various countries. The world’s major automobile production countries, including China, the United States, Japan, Germany, and others, have elevated the new energy automobile industry to a strategic position and increased investment in technology R&D of NEVs, as well as implemented various NEV policies. Consequently, the effectiveness of NEVs policies in promoting the adoption of NEVs and their impact on environmental protection have become major topics in literature.

2.1. Promotion Effectiveness of NEVs Policies

In order to reveal the promoting effectiveness of NEVs policies, there is a wealth of research focused on subsidy policies aimed at promoting NEVs, such as in Canada [28,34], United States [35,36], China [37,38], and so on. The NEVs policies implemented in various countries and their corresponding policy effects are summarized in

Table 1. NEVs policies implemented in several countries and their policy effect.

Categories	NEVs Polices	Countries	Policy Effect	References
Direct Financial Incentives	Waiver of toll roads	Norway	Increase NEVs sales	Zhang et al., 2016 [24]
	Purchase subsidy	Canada	Increase NEVs market share	Azarafshar and Vermeulen, 2020 [28]
	Concessional goods and service tax	India	Increase NEVs sales	Das and Bhat, 2022 [45]
	Import duty exemption	Brazil	Increase NEVs market share	Rietmann and Lieven, 2019 [46]
Indirect Incentives	HOV lane access	United Sates	Increase NEVs registrations	Wee et al., 2018 [36]
	Dense charging stations establishment	Sweden	Increase NEVs registrations	Egnér and Trosvik, 2018 [22]
	Purchase Restrictions on ICEVs	China	Increase NEVs sales	Chi et al., 2021 [47]
	Improvement in vehicle quality	Indonesia	Increase NEVs sales	Setiawan et al., 2022 [48]

. However, the conclusions of the existing literature are not consistent. Some literature has concluded that the subsidy policies could significantly increase the total production and sales of NEVs from both the supply side and demand-side. For instance, Azarafshar et al. [28] estimated that NEVs market shares increased by an average of over 5% in Canada when fiscal subsidy increased by CAD 1000. Gass et al. [39] proposed that the subsidy policies have made NEVs more competitive than traditional ICEVs in Austria. Gu et al. [40] found that subsidy policies could promote the optimal output and expected utility of NEVs manufacturers, which is more conducive to expanding the NEVs market. Huang et al. [41] showed that the subsidy of purchase was one of the important factors in determining consumers’ preference for NEVs. On the contrary, some other literature argued that subsidy policies had a negative impact on the financial burden of local governments [42,43]. Li et al. [44] indicated that subsidy policies have induced frequent fraud and weakened the competitiveness of the NEVs industry in China.

2.2. Environmental Effectiveness of NEVs Policies

To investigate the environmental effectiveness of NEVs policies, many existing studies have mainly discussed it at the city level. Kong et al. [7] argued that compared with traditional ICEVs, NEVs emit fewer air pollutants during the driving phase, and promoting NEVs policies have become an effective means to solve severe urban environmental problems. Qiao et al. [49] found that even from a life cycle perspective, NEVs emit less greenhouse gases than traditional ICEVs. Tan et al. [50] investigated that promoting NEVs policies could reduce the nitrogen dioxide concentrations in urban air. Zhang et al. [51] proposed that promoting some specific NEVs (e.g., new energy taxis and Didi) has great potential to alleviate the greenhouse effect and even reduce the haze weather in Beijing, China.

In summary, while much literature has constructed indicators for comprehensive evaluation of the environmental effects of NEVs policies, there are few in-depth studies on carbon emission. Moreover, the recent research has mainly focused on the effect of NEVs policies on the whole cities but omits their effects on specific transport industries or sectors. Actually, the NEVs policies are primarily aimed to mitigate carbon emissions from conventional fuel vehicles, which are the main source in the urban transport sector. The NEVs Pilot Policy, covering the public and private transport sectors, is an important measure implemented by the Chinese government to address the contradiction between increasing traffic and energy demand. However, its effectiveness in carbon emission reduction is still unclear. Therefore, compared with a comprehensive evaluation, it is worth paying more attention to the carbon emission efficiency of the NEVs Pilot Policy in the urban transport sector. This study seeks to fill such a research gap by exploring whether urban transport carbon emission is reduced significantly and effectively after implementing the NEVs Pilot Policy, which is also a major contribution of this paper.

3. Materials and Methods

3.1. Samples and Data

The NEVs Pilot Policy has been implemented in 25 cities (25 cities for only NEVs_Public, 6 cities for both NEVs_Public and NEVs_Private) from 2009 to 2010, which is the largest scale of NEV policies promoted in China. All cities belonged to a 53 first-tier and second-tier city list (The Ranking of Cities’ Business Attractiveness, 2022). Therefore, we selected these 25 cities as the treatment group in our study. And, the other 28 cities that belonged to the 53 first-tier and second-tier city list were considered as the control group (as shown in Figure 1). To account for the possibility of a time lag between the introduction and implementation of the first batch of the NEVs Pilot Policy, this paper sets the policy time point as 2010. In order to study the relatively long-term effectiveness of the NEVs Pilot Policy in China, the study period is from 2006 to 2020, which includes the 5 years before the NEVs Pilot Policy was implemented and 10 years after. Referencing “2006 IPCC (Intergovernmental Panel on Climate Change) Guidelines for National Greenhouse Gas Inventories” [52], the calculation methods of energy-related CO₂ from mobile sources can be divided into two categories: top-down and bottom-up. The former calculates the emissions based on the fuel consumption of transport vehicles (i.e., fuel-based), and the latter based on the data of different vehicle types, vehicle ownership, mileage, and energy intensity (i.e., mileage-based). The statistical scope of energy consumption in China’s transportation industry only includes the energy consumption of socially operating vehicles, while the fuel consumption of a large number of non-operating traffic (mainly from private cars and motorcycles) is not covered in the statistics. Thus, considering the statistical characteristics of China’s urban transport energy consumption to quantify the carbon emissions of urban transport more accurately, this paper adopts the top-down method to calculate the carbon emissions of operating vehicles and the bottom-up method to calculate the emissions of private traffic. These datasets are primarily derived from the China City Statistical Yearbook, the Prefecture-level City Statistical Yearbook, the official website of the provincial and municipal Statistics Bureau, statistical bulletins of cities, and the transport industry.

3.2. Research Hypotheses

The NEVs Pilot Policy provides strong support for the development of new energy vehicles, promotes the rapid growth of production and sales of new energy vehicles, and also improves the penetration rate of new energy vehicles. On the one hand, the NEVs Pilot Policy provides subsidies for individual purchase, registration, and use of new energy vehicles in pilot cities, prompting more people to purchase new energy vehicles to replace fuel private cars, reducing the proportion of traditional fuel vehicles, and promoting the reduction of urban traffic carbon emissions. On the other hand, this policy promotes the adoption of new energy vehicles in the public sector through financial subsidies in the field of public transportation and expands the scale of environmentally friendly public transport vehicles. At the same time, the renewal and upgrading of public transport vehicles have also attracted more people to use public transport including attracting private car owners to turn to public transit to avoid traffic congestion and protect the environment [53], thereby improving the carbon emission efficiency of the urban transportation sector. In addition, the promotion of this policy in the public domain also stimulates the adoption of NEVs by private car users because the city’s related infrastructure (e.g., more charging stations and special parking space for NEVs) becomes better under this policy, the private users can also indirectly benefit from the promotion.

As the largest greenhouse gas (GHG) emitter in the world, China emits most of the GHG through fossil fuel burning. The transport industry of China is highly dependent on oil consumption. NEVs provide a substitute for traditional internal combustion engine vehicles, which presents way an effective for the transport sector to achieve the low-carbon transition by shifting oil consumption to electricity consumption. Under the promotion of the NEVs Pilot Policy, the operating cost and related risks of new energy vehicle manufacturers have been significantly reduced, thus creating more space for new energy vehicle manufacturers to carry out more technological research and development work including the manufacturing of key components of the electric drive system, the improvement of battery energy storage, and driving mileage, and the guarantee of safe driving [29]. This further promotes the replacement of traditional fuel vehicles by new energy vehicles, which reduces fossil fuel consumption and the energy intensity of the urban transport sector. In addition, the NEVs Pilot Policy also encourages consumers to buy new energy vehicles through purchase subsidies and related infrastructure construction. This also accelerates the replacement of fuel vehicles by new energy vehicles and thus lowers the energy intensity of the urban transport sector.

Based on the above theoretical analysis, this paper proposes the following research hypotheses:

Hypothesis 1.

The implementation of the NEVs Pilot Policy can significantly reduce the urban transport carbon emissions.

Hypothesis 2.

The NEVs Pilot Policy can reduce urban transport carbon emissions by adjusting the vehicle structure (VS) of the urban transport sector.

Hypothesis 3.

The NEVs Pilot Policy can reduce urban transport carbon emissions by lowering the energy intensity (EI) of the urban transport sector.

3.3. Model Establishment and Variable Description

The difference-in-differences (DID) method was first introduced to economics research by Ashenfelter and Card in 1985 when quantifying the causal effect of a government training program [54]. In the field of economics, the application of the DID method mainly considers policy intervention as “quasi-natural experiments” [55,56]. The basic idea of the DID method is to construct a difference-in-differences statistic that reflects the policy effect by comparing the differences between the treatment group and control group before and after policy implementation. The main steps of the DID method are as follows: (1) Introducing the “control group” that is not affected by the policy, and the “treatment group” as the target of policy intervention; (2) Determining the outcome variables such as income level, emissions amount, etc.; (3) Calculating the differences in outcomes before and after the policy implementation for the treatment group and control group, respectively; (4) Calculating the differences in outcomes between the treatment group and control group. The DID model has been widely used for policy effectiveness evaluation. For example, Qiu et al. [57] employed the DID model to identify the positive effect of China’s Low-carbon city policy on the city’s green total factor productivity. Xie et al. [58] conducted an empirical study on the relationship between NEVs subsidy program and urban air pollution.

In this study, we built the difference-in-differences (DID) model to evaluate the effectiveness of NEVs Pilot Policy on urban transport carbon emissions by comparing their differences in 25 cities (as the treatment group) and 28 cities (as the control group) before and after the policy implemented. Three indicators were used to measure the urban transport carbon emissions at aggregated, per capita, and intensity levels: the total of urban transport carbon emissions (TCE), the urban transport carbon emissions per capita (PCE), and the urban transport carbon emission intensity (CEI). To obtain the treatment effectiveness of NEVs Pilot Policy on urban transport carbon emissions, the benchmark DID model was set as follows:

$$Y_{\mathit{it}}=\alpha +\beta \left({\mathit{Treat}}_i\times {\mathit{Post}}_t\right)+\gamma X_{\mathit{it}}+{\mu }_i+{\lambda }_t+{\epsilon }_{\mathit{it}},$$

(1)

in which i represents the i city (i = 1, 2, …, 53), and t represents the t year (t = 2006, 2007,…, 2020). Y_it is the dependent variable representing the urban transport carbon emissions of the i city in the t year. The coefficient α is the constant. β is the core coefficient indicating the net effectiveness of the NEVs Pilot Policy on urban transport carbon emissions. Treat_i represents whether the NEVs Pilot Policy is implemented in the i city. If the i city belongs to the treatment group, Treat_i equals 1, otherwise equals 0. Post_t indicates whether the NEVs Pilot Policy is implemented in the t year. Post_t equals 1 if the t year equals or larger than 2010, otherwise equals 0. Therefore, Treat_i × Post_t as one of the independent variables, is the core explanatory dummy variable in this study. Treat_i × Post_t equals 1 if the i city belongs to the treatment group and in 2010 or after that. Otherwise, Treat_i × Post_t equals 0. X_it is the control variable that may affect the urban transport emissions. In this paper, the urban economic development level (GDP per capita, PGDP), the urban transport industry development level (GDP of transport industry, TrGDP), the transport intensity (total converted turnover/GDP, TI), the transport structure (road converted turnover/total converted turnover, TS), the road passenger turnover (RPT), and the road freight turnover (RFT) are selected according to the previous literature investigated the influencing factors of urban transport emissions [59,60,61,62]. μ_i is the city-related fixed effect, which controls all the city-related factors that do not change over time. λ_t is the time-related fixed effect, which controls all the time related factors that do not change with city changes. ε_it represents the random error term. The descriptive statistics of the variables for the sample cities are summarized in

Table A1. Descriptive statistics of the variables for the sample cities.

Variables	Total (Obs. = 795)		Treatment Group (Obs. = 375)		Control Group (Obs. = 420)
	Mean	Std. Dev	Mean	Std. Dev	Mean	Std. Dev
lnTCE	6.8299	0.7511	7.0774	0.6681	6.5956	0.7489
PCE	1.3938	0.7781	1.5086	0.6544	1.2912	0.8617
CEI	4.9107	2.7302	4.2329	1.6933	5.5159	3.2845
lnPGDP	10.989	0.5309	11.081	0.5010	10.906	0.5437
lnTrGDP	5.3553	0.8203	5.7106	0.7710	5.0381	0.7285
TI	0.3337	2.3685	0.4895	3.4396	0.1946	0.1718
TS	0.3608	0.2276	0.3134	0.2116	0.4030	0.2332
lnRPT	4.2783	0.7692	4.3673	0.8412	4.1988	0.6899
lnRFT	5.1874	0.9807	5.3756	0.9353	5.0194	0.9909

.

A two-stage mediating effect model is adopted to analyze the influencing mechanism of the NEVs Pilot Policy on urban transport carbon emission reduction [3,63]. In this model, the basic assumption is that the dependent variables could be affected by the independent variables through the mediating variables. The analysis of mediating variables includes both the direct effect of independent variables on dependent variables and the indirect effect of independent variables on dependent variables through mediating variables. In this study, we use model (1) to examine the direct effect. Then, model (2) and (3) are used to verify potential mediating effects. These models are set as follows:

$$M_{\mathit{it}}={\alpha }_1+{\beta }_1\left({\mathit{Treat}}_i\times {\mathit{Post}}_t\right)+{\gamma }_1{X}_{\mathit{it}}+{\mu }_i+{\lambda }_t+{\epsilon }_{\mathit{it}},$$

(2)

$$Y_{\mathit{it}}={\alpha }_2+{\beta }_2\left({\mathit{Treat}}_i\times {\mathit{Post}}_t\right)+{\delta M}_{\mathit{it}}+{\gamma }_2{X}_{\mathit{it}}+{\mu }_i+{\lambda }_t+{\epsilon }_{\mathit{it}}.$$

(3)

In model (2), if the coefficient β₁ is not significant, it suggests that the mediating variable M_it has no significant effect, and the analysis stops. Otherwise, the variable M_it is considered in model (3). If the coefficient β₂ and δ are both significant, it indicates that the corresponding variable M_it has a mediating effect, that is, the effect of NEVs Pilot Policy is achieved through the mediating variable M_it.

4. Results

4.1. Baseline Regression Results

In this section, the effectiveness of NEVs Pilot Policy on urban transport carbon emissions is estimated based on Equation (1), where the total urban transport carbon emissions (TCE, t), the urban transport carbon emissions per capita (PCE, t), and the urban transport carbon emission intensity (CEI, t per unit GDP (CNY 100 million)) are used as the dependent variables to measure urban transport carbon emissions. The estimated results are presented in columns 1–3 of

Table 2. The effectiveness of NEVs Pilot Policy on urban transport carbon emissions.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3465 ***	−0.3056 **	−2.0441 ***	−0.4029 ***	−0.3782 ***	−2.1922 ***
	(−5.08)	(−2.43)	(−4.47)	(−7.63)	(−3.38)	(−6.72)
lnPGDP				−0.0073	−0.3412	−0.7714
				(−0.05)	(−1.21)	(−1.36)
lnTrGDP				0.0532	−0.0123	−3.7209 ***
				(1.34)	(−0.11)	(−7.92)
TI				0.0017	0.0033	0.0062
				(0.49)	(0.61)	(0.38)
TS				−0.2438 **	−0.1745	−0.5423
				(−2.48)	(−0.74)	(−0.91)
lnRPT				−0.0120	−0.0247	−0.4920
				(−0.34)	(−0.31)	(−1.61)
lnRFT				0.4682 ***	0.6569 ***	2.5657 ***
				(12.38)	(6.22)	(5.94)
Constant	6.9427 ***	1.4995 ***	5.6178 ***	4.4673 ***	2.0996	23.0609 ***
	(294.15)	(34.54)	(35.55)	(3.19)	(0.71)	(3.67)
Observation	795	795	795	795	795	795
R-squared	0.930	0.581	0.725	0.966	0.647	0.878
Controls	N	N	N	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. ** and *** represent significant at the level of 5% and 1%, respectively.

. Models (1)–(3) indicate that the NEVs Pilot Policy has a significantly negative impact on urban traffic carbon emissions, which further suggesting that this NEVs Pilot Policy has significantly reduced the transport carbon emissions of cities in the treatment group. Specifically, compared to the cities in the control group, the NEVs Pilot Policy has significantly reduced the total urban transport carbon emissions (TCE) by an annual average of 29.3% (According to the semi-elastic model, the net effect of Treat_i × Post_t on the dependent variable should be exp(−0.3645) − 1 = 29.3%.) during the period from 2010 to 2020. The annual carbon emissions per capita (PCE) of urban transport have been lowered by an average of 0.31 t, and the annual carbon emission intensity (CEI) have also been reduced by 2.04 t per unit GDP. These findings are sensible and also consistent with a previous study [53], but the carbon emission reduction obtained in this paper are slightly higher. This is because this previous study only considered the NEV policy in the public sector and did not account for the carbon emissions from the private transport sector. As a major measure for developing the new energy automobile industry in China, the NEVs Pilot Policy has promoted the diffusion of NEVs through purchase subsidy and related infrastructure investments, providing alternatives to traditional ICEVs [64]. Moreover, the implementation of this NEVs Pilot Policy has reduced urban transport carbon emissions, which offering a solution to alleviate both the environmental problems and energy crisis faced by the transport sector in China.

After that, we considered the impact of control variables (including PGDP, TrGDP, TI, TS, RPT, and RFT), which may affect the effectiveness of the policy on urban transport carbon emissions. The results are shown in columns 4–6 of Table 2. The regression coefficient is still negative and has passed the significance test. These results indicate that the NEVs Pilot Policy can effectively reduce the level of urban transport carbon emissions, further supporting the conclusions drawn previously.

4.2. Parallel Trend Test and Dynamic Treatment Effect

The vital prerequisite to ensure unbiased results of the difference-in-difference (DID) estimation is that the treatment group and the control group satisfy the parallel trend hypothesis [65]. In other words, in the absence of NEVs Pilot Policy intervention, the variation trend of urban transport carbon emissions in cities of the treatment group and of the control group should be the same.

Additionally, the above benchmark results reflect the average effectiveness of the NEVs Pilot Policy on urban transport carbon emissions but do not capture its dynamic effectiveness. It is noted that there may be a possible time lag between the introduction of the NEVs Pilot Policy and its implementation. Also, the policy effectiveness might not work immediately due to the following reasons. First, NEVs, as a new type of automobile product, requires time for consumers to understand and accept. Second, the production of NEVs as well as the construction and installation of related supporting facilities, also takes time. Therefore, it requires a certain period to replace traditional ICEVs with NEVs. Third, in order to make the NEVs market’s sustainable development, the policy is also constantly adjusted with the development of NEVs industry, which may lead to changes in the policy effectiveness [53]. In China, the central government has begun to gradually reduce the subsidies of NEVs since 2017, which aimed at the transformation of NEVs industry from government intervention into market dominance.

Consequently, following Liu et al. [3] and Wing et al. [66], this study makes the parallel trend test and the dynamic effectiveness analysis by extending the baseline DID model to the time-varying DID model as follows:

$$Y_{\mathit{it}}={\alpha }_3+{\sum }_{j=2006}^{2020}{\beta }_j\left({\mathit{Treat}}_i{\times \mathit{Post}}_t\times {\mathit{Year}}_{\mathit{it}}^j\right)+{\gamma }_3{X}_{\mathit{it}}+{\mu }_i+{\lambda }_t+{\epsilon }_{\mathit{it}}.$$

(4)

where ${\mathit{Year}}_{\mathit{it}}^j$ is the dummy variable, j = 2006, 2007,…, 2020. Referring to Liu [67], we set the first year in the study period, namely 2006, as the base year to run the regression based on (4). β_j represents the different estimated coefficients of the j year. If the estimated coefficient β_j is not significant before the year (2010) that the NEVs Pilot Policy implemented, it proves that the parallel trend hypothesis is true. Furthermore, this model is used to measure the dynamic effectiveness of NEVs Pilot Policy in each year after its implementation.

Figure 2 shows the estimation results of β_j under the 95% confidence intervals for the dependent variables of TCE, PCE, and CEI, respectively. In Figure 2, β_j was not statistically significant before 2010, which indicates that there was no significant difference between the treatment group and the control group before the NEVs Pilot Policy implemented. Therefore, the parallel trend hypothesis is true.

The results of the dynamic effectiveness analysis (in

Table A2. The dynamic effectiveness of NEVs Pilot Policy on urban transport carbon emissions.

Year (Time)	(1)	(2)	(3)
	lnTCE	PCE	CEI
2009 (−1)	−0.0315	0.1558	−0.7444
	(−0.39)	(1.67)	(−1.23)
2010 (0)	−0.1139	0.0892	−1.3967 **
	(−1.35)	(0.82)	(−2.29)
2011 (1)	−0.1694 *	0.0456	−1.7256 **
	(−1.89)	(0.37)	(−2.46)
2012 (2)	−0.2106 **	−0.3461	−2.0456 ***
	(−2.15)	(−0.82)	(−2.78)
2013 (3)	−0.2486 **	0.0192	−2.2249 ***
	(−2.45)	(0.12)	(−3.00)
2014 (4)	−0.2691 ***	−0.0284	−1.9787 ***
	(−2.71)	(−0.19)	(−3.14)
2015 (5)	−0.3860 ***	−0.2097	−2.7893 ***
	(−3.51)	(−1.28)	(−3.23)
2016 (6)	−0.4615 ***	−0.3382 *	−2.5174 ***
	(−4.51)	(−1.95)	(−3.73)
2017 (7)	−0.4815 ***	−0.3575 *	−2.5288 ***
	(−4.55)	(−1.85)	(−3.50)
2018 (8)	−0.5095 ***	−0.4138 **	−2.6342 ***
	(−4.74)	(−2.04)	(−3.61)
2019 (9)	−0.5291 ***	−0.5035 **	−3.2710 ***
	(−4.92)	(−2.61)	(−4.28)
2020 (10)	−0.5691 ***	−0.5732 ***	−3.1443 ***
	(−5.62)	(−3.45)	(−4.41)
Constant	6.9486 ***	1.4675 ***	5.7795 ***
	(195.73)	(27.57)	(23.57)
Observation	795	795	795
R-squared	0.938	0.596	0.734
Controls	Y	Y	Y
City FE	Y	Y	Y
Year FE	Y	Y	Y

Notes: Robust t-statistics in parentheses. *, ** and *** represent significant at the level of 10%, 5% and 1%, respectively.

) show that the coefficients of TCE and PCE began to be significant from 2011 and 2016, respectively, implying a certain lagging effect on urban transport carbon emission. In particular, the coefficient of TCE showed a significantly negative and increasing trend from 2011 to 2020, which illustrates that the effectiveness of NEVs Pilot Policy accumulated and increased over time, although the subsidies declined. However, the coefficient of CEI became statistically significant from 2010, which is consistent with the Chinese government’s commitment to controlling carbon emission intensity, supplemented by total carbon emission control. Since the announcement of the NEVs Pilot Policy, the specific subsidy policy has been continuously adjusted in the past decade, which could be divided into three stages: the first stage (2009–2012) is “promoting the energy conservation, reducing the carbon emissions, and expanding the NEVs industry”, the second stage (2013–2015) is “accelerating the development of NEVs industry”, and the third stage (2016-present) is “developing the healthy and high-quality NEVs industry” [68]. Furthermore, the Chinese government has adopted a modest declining subsidy policy, which did not have a serious negative impact on the development of the NEVs industry, and even still has a positive effect on the reduction of urban transport carbon emissions. Meanwhile, with the implementation of NEVs Pilot Policy, more residents became aware of the important roles of NEVs in environmental protection. The market share of NEVs continues to expand, improving the effectiveness of the NEVs Pilot Policy [69]. These above findings support Hypothesis 1.

4.3. Robustness Analysis

4.3.1. Placebo Test

In order to exclude some unobserved random factors changing over time or possible missing variables that may affect the policy effect, this study uses the counterfactual method to conduct a placebo test by randomly selecting the samples as the treatment group [70]. The specific experimental contents are as follows: (1) Using Stata software 16.0 to run 500 random experiments on the samples. In each experiment, 25 cities are randomly selected as the treatment group, and the remaining samples are used as the control group. (2) The policy interaction dummy variable “treat*period” is generated and then regressed using Equation (1) to obtain the estimated coefficient of random experiments. (3) Drawing the kernel density function diagram of the estimated coefficient of random experiments and its’ p-value distribution diagram. The results are shown in Figure 3, and the estimated coefficients of “treat*period” under the random experiments are mainly concentrated near 0. The actual estimated coefficients are significantly different from the placebo results, and most estimated coefficients have p-values greater than 0.1, indicating that the estimation effect under random treatment is not significant, and the core conclusions of this study are robust.

4.3.2. PSM-DID Test

Although the DID model enables the estimation of the average effectiveness of the NEVs Pilot Policy, it is not a strictly natural experiment, and there may be potential sample selection bias. Therefore, to select a suitable control group for comparison, this study further uses propensity score matching (PSM) [71,72]. By using the PSM test, we select the covariates which are highly correlated with the dependent variables to be the matching variables and use the Logit model to calculate the propensity score of each sample. Then, according to the kernel matching criterion, we choose matching samples from the control group that is closest to the propensity scores of the samples in the treatment group. To ensure the reliability of the matching results, this paper conducts a balance test, that is, after matching, there should be no significant systematic difference in the covariates between the treatment group and the control group, except in TCE, PCE, and CEI. Figure 4 shows that most samples of the treatment group and the control group are within the common value range, and the other samples which have the extreme scores are concentrated near 0 or 1. The results of the balance test are shown in Figure 5. The standardized deviation of the matched variables between the treatment group and the control group is significantly reduced, which indicates that the PSM test is effective. Finally, the matched samples are used to run the DID model, and the estimated results of PSM-DID are shown in

Table A3. Estimation results of PSM-DID.

	(1)	(2)	(3)
	lnTCE	PCE	CEI
Treat_Post	−0.3721 ***	−0.3634 ***	−1.8415 ***
	(−10.30)	(−3.51)	(−7.32)
Observation	732	732	732
R-squared	0.973	0.654	0.903
Controls	Y	Y	Y
Year FE	Y	Y	Y
City FE	Y	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

, which are not significantly different from the previous benchmark regression results. Therefore, the NEVs Pilot Policy still significantly reduced the level of urban traffic carbon emissions, which further supports the conclusions of this study.

4.3.3. The Impact of Other Pilot Policy

The Chinese government has adopted various policies to reduce urban carbon emissions, which may lead to overestimation or underestimation of the effectiveness of the NEVs Pilot Policy. To ensure the robustness of our results, we searched for other policies implemented during the period of 2010–2020 that aimed at reducing urban carbon emissions. Among these policies, the most direct and significant ones are the Emission Trading Scheme (ETS) and the New Energy Demonstration City (NEDC) Policy, which was announced and implemented by the Chinese government to achieve urban carbon emissions reduction targets [3,73,74]. In 2013, the Chinese government launched the ETS in seven pilot regions, which aimed to realize energy saving and emission reduction through market mechanisms. And, the NEDC Policy has been implemented in 81 pilot cities since 2014 to promote the advancement of new energy technologies for low-carbon urban development. Some existing research showed that the ETS and the NEDC Policy have significantly reduced the carbon emissions in these pilot regions [75,76]. Therefore, to rule out the possible confusion caused by the ETS and the NEDC Policy, the dummy variables represented the ETS and the NEDC Policy, respectively, are included in the baseline regression model. The results (shown in

Table A4. Robustness test considering ETS and NEDC.

	ETS			NEDC
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3971 ***	−0.3569 ***	−2.1890 ***	−0.4027 ***	−0.3763 ***	−2.1871 ***
	(−8.30)	(−3.39)	(−6.98)	(−7.52)	(−3.32)	(−6.69)
ETS	−0.0527	−0.1908	−0.0283
	(−0.65)	(−1.35)	(−0.08)
NEDC				−0.0068	−0.0630	−0.1656
				(−0.13)	(−0.58)	(−0.66)
Observation	795	795	795	795	795	795
R-squared	0.966	0.650	0.878	0.966	0.647	0.878
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

) indicate that the coefficients of both the ETS and NEDC Policy are not statistically significant. This suggests that these policies had no significant impact on urban transport carbon emissions. This is reasonable since the ETS is mainly aimed at the manufacturing sectors but not the transport sectors, while the NEDC Policy is committed to developing the new energy industry rather than directly targeting the transport industry. Moreover, the NEDC Policy has been implemented in many cities in both the treatment group and control group. Thus, after excluding the confounding effects of the ETS and the NEDC Policy, the results still support the significant reduction of urban transport carbon emissions due to the NEVs Pilot Policy, demonstrating its robustness.

4.4. Mechanism Analysis

The empirical results above demonstrate that the NEVs Pilot Policy has significantly reduced urban transport carbon emissions. In this section, we further investigate the mechanism through which the NEVs Pilot Policy influences urban transport carbon emissions in the following two aspects: the vehicle structure (VS) and the energy intensity (EI). Specifically, we used the ratio of the number of buses to the number of private cars as a proxy variable for the vehicle structure. The energy intensity is measured by the ratio of total urban transport energy consumption to the GDP of the transport industry.

The estimated results of Equations (2) and (3) are given in

Table 3. Results of the mediating variables test.

	(1)	(2)
	VS	EI
Treat_Post	0.0059 ***	−1.0381 ***
	(6.35)	(−6.49)
Observations	795	795
R-squared	0.635	0.878
Controls	Y	Y
Year FE	Y	Y
City FE	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

and

Table 4. Results of the mediating effect test.

	(1)	(2)
	lnTCE	lnTCE
Treat_Post	−0.3422 ***	−0.1918 ***
	(−7.28)	(−4.28)
VS	−10.3438 ***
	(−2.81)
EI		0.2034 ***
		(4.41)
Observations	795	795
R-squared	0.967	0.982
Controls	Y	Y
Year FE	Y	Y
City FE	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

, respectively. The dependent variables in Table 3 are the mediating variables of the NEVs Pilot Policy. The results (shown in Table 3) indicate that the NEVs Pilot Policy has significantly promoted the adjustment of vehicle structure, which has further led to lower urban transport carbon emissions, as evidenced by the significantly negative coefficients of VS and Treat_Post (shown in Table 4). This is due to the fact that higher penetration of NEVs in the public sector increased carbon emission efficiency and promoted urban transport carbon emission abatement. Additionally, the energy intensity (EI) is also found to be significantly affected by the NEVs Pilot Policy (shown in column (2) of Table 3), and lower EI results in less urban transport carbon emissions (shown in column (2) of Table 4). Similar results were also found by Wang [29], suggesting that reducing energy consumption is an effective way for the urban transport sector to achieve emission reduction targets. Thus, the two theoretical mechanisms of this paper (Hypotheses 2 and 3) are verified. The NEVs Pilot Policy has vigorously and successfully encouraged NEVs in the public sector, with most public vehicles now being replaced by NEVs. In contrast, the penetration rate of NEVs is relatively low in the private sector. Therefore, to further reduce urban transport carbon emissions, promoting NEVs in the private sector and reducing the energy intensity of the transport industry is necessary for the future.

The results above confirm the significant role of the vehicle structure (VS) and energy intensity (EI) in the effectiveness of the NEVs Pilot Policy. The transport sector is one of the biggest carbon emitters in China, and the annual growth rate of its energy consumption exceeds the national average level [77]. By adjusting the vehicle structure and reducing energy intensity, it is helpful to promote the transformation of urban transport energy consumption, which further reduces urban transport carbon emissions. The Chinese government should fully utilize the NEVs Pilot Policy to promote NEVs as urban public service vehicles, and gradually decrease the proportion of traditional ICEVs in automobile production and sales. Furthermore, it is recommended that the Chinese government encourages automobile companies to research and develop low-carbon technology for reducing vehicle energy consumption.

4.5. Heterogeneity Analysis

The estimated results above have proven that the NEVs Pilot Policy could effectively reduce the urban transport carbon emissions of cities in the treatment group. However, the differences in the economic level, the political status, and the urban transport development of these cities may lead to a certain degree of heterogeneity of the policy effect. In this section, we analyze the heterogeneity in the policy effect by sample grouping regression, which is used to explore how the NEVs Pilot Policy can have different effectiveness on urban transport carbon emissions in different cities. According to the city’s per capita GDP, this study divides the economic development of cities into two levels: “High level” and “Low level”. We set a dummy variable “Economic” to indicate city’ economic development level. When the city belongs to the “High level” group, “Economic” equals 1, otherwise it equals 0. Moreover, we set a dummy variable “Political” to signify whether the city is a municipality or provincial capital. Similarly, this study chooses the urban population, the number of buses, and the GDP of the urban transport industry to represent the urban transport development level of cities, which are set as the corresponding dummy variables “POP”, “BUS” and “TrGDP”, respectively. The dependent variables are the same as those in the baseline regression model, and the regression results are shown in

Table A5. Heterogeneity analysis based on the economic level.

	Economic_High			Economic_Low
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3273 ***	−0.2340 *	−1.6359 ***	−0.2053 ***	−0.0941	−1.4056 ***
	(−4.17)	(−1.79)	(−5.85)	(−4.96)	(−0.84)	(−3.82)
Observations	396	396	396	394	394	394
R-squared	0.970	0.880	0.826	0.987	0.617	0.902
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. * and *** represent significant at the level of 10% and 1%, respectively.

,

Table A6. Heterogeneity analysis based on the political status.

	Political_Muni/Cap			Political_Other
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3927 ***	−0.5537 ***	−1.6823 ***	−0.3534 ***	−0.1399	−2.7012 ***
	(−11.95)	(−3.49)	(−7.59)	(−5.70)	(−0.86)	(−5.13)
Observations	405	405	405	390	390	390
R-squared	0.973	0.539	0.890	0.977	0.911	0.895
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

,

Table A7. Heterogeneity analysis based on the urban population.

	POP_High			POP_Low
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3968 ***	−0.5342 ***	−2.4385 ***	−0.2781 ***	0.0133	−1.4396 ***
	(−8.01)	(−4.55)	(−5.53)	(−4.69)	(0.10)	(−4.30)
Observations	396	396	396	395	395	395
R-squared	0.969	0.868	0.912	0.971	0.609	0.905
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. *** represents significant at the level of 1%.

,

Table A8. Heterogeneity analysis based on the urban bus development.

	BUS_High			BUS_low
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3789 ***	−0.4431 ***	−1.5654 ***	−0.1918 ***	0.1082	−1.7641 ***
	(−3.85)	(−2.85)	(−3.76)	(−3.08)	(0.76)	(−2.79)
Constant	7.5332 ***	5.1228	34.8508 ***	4.0445 ***	1.9483	23.0396 **
	(3.24)	(0.98)	(3.54)	(3.27)	(0.68)	(2.30)
Observations	395	395	395	397	397	397
R-squared	0.933	0.812	0.888	0.982	0.589	0.902
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. ** and *** represent significant at the level of 5% and 1%, respectively.

and

Table A9. Heterogeneity analysis based on the urban transport industry development.

	TrGDP_High			TrGDP_Low
	(1)	(2)	(3)	(4)	(5)	(6)
	lnTCE	PCE	CEI	lnTCE	PCE	CEI
Treat_Post	−0.3747 ***	−0.6156 **	−2.0323 ***	−0.2371 ***	−0.1296	−1.5844 ***
	(−2.85)	(−2.05)	(−3.08)	(−4.26)	(−1.12)	(−3.70)
Observations	395	395	395	396	396	396
R-squared	0.921	0.815	0.815	0.980	0.501	0.912
Controls	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
City FE	Y	Y	Y	Y	Y	Y

Notes: Robust t-statistics in parentheses. ** and *** represent significant at the level of 5% and 1%, respectively.

.

In terms of the economic level, the regression results (shown in Table A5) indicate that the NEVs Pilot Policy has a negative effect on urban transport emissions both in the two groups with different economic development levels. Moreover, the NEVs Pilot Policy has stronger effectiveness on cities with higher economic development levels. There could be several factors that contribute to the better emission reduction effects in cities with a higher level of economic development: (1) Infrastructure investment: Higher levels of economic development often translate into greater financial resources available for infrastructure investment. Cities with more advanced economies may allocate more funds for developing charging infrastructure and other supporting facilities for NEVs. This is also consistent with the findings of the study conducted by Silvia and Krause [78]. (2) Market competition: The presence of a more competitive market in economically developed cities can drive technological innovation and accelerate product improvement (e.g., enhance NEVs performance, increase range, and reduce charging time), which increases the attractiveness of NEVs to potential users. This further validates the findings of the previous study [79]. (3) Consumer affordability: Higher income levels in economically developed cities can make NEVs more affordable for residents. The cost of purchasing and maintaining NEVs may be relatively lower in comparison to their income, leading to a higher demand for NEVs. Notably, the coefficient of PCE in the Economic_Low group is not statistically significant though negative, which indicates that emission efficiency should be enhanced in the future.

In terms of the political status, the regression results (presented in Table A6) suggest that the NEVs Pilot Policy has more obvious effectiveness on TCE and PCE in municipalities and provincial capitals, as compared to other cities. This may be due to better policy support and resource advantages in cities with higher political status. On the one hand, cities with higher political status may provide stronger policy support for the NEVs promotion, such as more financial incentives and subsidies, and preferential treatment in terms of parking and traffic restrictions. Such supportive policies can effectively encourage individuals and businesses to adopt NEVs. On the other hand, cities with higher political status generally possess abundant resources including financial resources, human capital, and technological capabilities. These resources can support the development of NEVs industry, thereby enhancing the efficiency of promotion and emissions reduction. However, the policy effect on CEI in these cities is relatively lower. This could be attributed to the stricter environmental protection policies in these cities, resulting in lower carbon intensity and making further reduction more challenging. This finding is consistent with the earlier research by Wen et al. [80].

In terms of urban transport development, the regression results (as shown in Table A7) reveal that the NEVs Pilot Policy has a notably stronger impact on the urban transport emissions in the POP_High group. The promotion of policy effectiveness by population effect can be explained from the following two aspects: (1) Market potential: cities with larger populations offer a larger market scale for NEVs. This attracts more manufacturers and suppliers to provide a wide range of vehicle models and charging infrastructure options. The availability of diverse choices and services makes it more convenient for residents to adopt and use NEVs. (2) Transportation infrastructure: cities with larger populations generally have higher population densities. Such cities typically have more developed transportation infrastructure, including road networks and public transportation systems. This can facilitate NEVs to be more efficient, making it easier for residents to adopt NEVs and utilize NEVs charging infrastructure. This provides further support for the research conducted by Liu et al. [81], suggesting that the Chinese government should target larger cities with more populations to promote NEVs. Furthermore, the regression results (shown in Table A8 and Table A9) indicate that the effectiveness of the NEVs Pilot Policy is more obvious on TCE and PCE in the cities where urban transport development is higher (including the urban bus development and the urban transport industry development). In these cities, the developed transport system facilitates the efficient flow of labor, capital, and high-quality resources across various production activities, which is a crucial measure for reducing urban transport carbon emissions. Furthermore, the developed urban transportation industry is more conducive to attracting more capital investment in the NEVs industry, which could further enhance the policy effect. However, it is surprising to find that the policy effect on the CEI is more significant in cities where the urban transport development is lower. The reasons for this may be as follows: The promotion effect of the NEVs Pilot Policy is more significant in the public transport sector of the cities with higher urban transport development. In these cities, most public vehicles have already been replaced by NEVs, resulting in a lower potential for energy intensity reduction compared to cities with lower urban transport development.

Therefore, the economic level, political status, and urban transport development are the significant factors that determine the effectiveness of the NEVs Pilot Policy on urban transport emissions. In the future, the Chinese government should target cities with higher economic development levels and larger populations to promote NEVs. In addition, for cities with higher economic and urban transport development levels or the provincial capitals and municipalities, the Chinese government should focus on promoting NEVs in the private transport sector, which could further improve the overall carbon emission efficiency. In the meantime, for developing cities with lower urban transport development levels, promoting the NEVs in the public transport sector should be a priority.

5. Conclusions and Policy Implications

As a vital and strategic policy in achieving the goal of emission peak and carbon neutralization, the NEVs Pilot Policy implemented in China has attracted worldwide attention. This study evaluated its dynamic effectiveness and the influencing mechanism in promoting China to reduce urban transport carbon emissions.

By using the panel data of 53 prefecture-level cities in China from 2006 to 2020, this study performed the difference-in-differences (DID) model to evaluate the effectiveness of the NEVs Pilot Policy on urban transport carbon emissions. Overall, the NEVs Pilot Policy has a significantly negative effect on urban transport carbon emissions. Specifically, the findings indicate that the NEVs Pilot Policy is effective in reducing the total urban transport carbon emissions (TCE) by an annual average of 29.3% during the period from 2010 to 2020. The annual carbon emissions per capita (PCE) of urban transport have been lowered by an average of 0.31 t, and the annual carbon emission intensity (CEI) has also been reduced by 2.04 t per unit GDP. These findings are similar to those of Zhang et al. [53], who conclude that enhancing the adoption of NEVs in the public sector leads to a significant reduction in emissions from urban transportation. Furthermore, our study makes a valuable contribution to validating the crucial role of the private sector’s NEVs promotion in facilitating the reduction of carbon emissions in urban transportation.

In addition to the accumulative policy effect, this study also adopted the time-varying DID model to estimate the dynamic effectiveness of the NEVs Pilot Policy over time. The findings show that the estimated policy effect on TCE and PCE has exhibited certain lagging but enhancement patterns, but the policy effect on CEI has been consistent with the timing of the Chinese government’s commitment. In particular, despite the Chinese government having adopted a modest declined subsidy policy, the NEVs Pilot Policy continues to have a positive effect on reducing urban transport carbon emissions. This can be attributed to the higher penetration of NEVs in the urban transport sector of the cities, and the development of the NEVs industry has not been seriously affected.

A series of robustness tests were conducted, and the results show that our conclusions have sufficient robustness. To rule out the confounding impacts of other policies, robustness checks have also been conducted. We especially considered and compared the effectiveness of this NEVs Pilot Policy with China’s other policies (including the Emission Trading Scheme (ETS) and the New Energy Demonstration City (NEDC) Policy) to reduce urban transport carbon emissions. The ETS and NEDC Policy both do not contribute to significant urban transport carbon emissions because the ETS is mainly aimed at the manufacturing sector and the NEDC Policy is committed to developing the new energy industry. These two policies are not directly targeted at the transport industry. Thus, our empirical evidence supports that the NEVs Pilot Policy has still significantly reduced urban transport carbon emissions after excluding the confounding effects of the ETS and the NEDC Policy.

In addition to the above estimation of the policy effect, we investigated the influencing mechanism through which the urban transport carbon emissions have been reduced. There are two aspects, namely the vehicle structure (VS) and the energy intensity (EI), the higher carbon emission efficiency of higher NEVs penetration in urban public transport and the lower energy consumption of the urban transport sector. The NEVs Pilot Policy has been found to significantly promote the adjustment of VS and EI, then the urban transport carbon emissions have been reduced as a result. These findings suggest that adjusting the vehicle structure and reducing the energy intensity could promote the transformation of urban transport energy consumption, which further reduces urban transport carbon emissions. That is, the mechanism of the NEVs Pilot Policy to reduce urban traffic carbon emissions by adjusting vehicle structure and reducing energy intensity has been confirmed. Furthermore, the Chinese government should promote the NEVs as urban public service vehicles for reducing the proportion of traditional ICEVs in automobile production and sales and encourage automobile companies to research and develop low-carbon technology for reducing vehicle energy consumption.

Moreover, when formulating and implementing the NEVs Pilot Policy in different cities, it is necessary to adapt it to local conditions in accordance with local economic, political, and transport characteristics. This study examined the varying policy effect among heterogeneous cities. And, the heterogeneity analysis shows that the effectiveness of the NEVs Pilot Policy varies significantly in different cities due to their different economic level, political status, and urban transport development. This outcome is consistent with the results obtained by Liu et al. [81] who argue that the variations in economic development, infrastructure development, population density, and market environment across different regions contribute to the disparities in the effectiveness of NEVs promotion policies. Our findings indicate that this policy’s effect on urban transport carbon emissions is stronger in municipalities and provincial capitals, as well as in cities with higher economic development levels and larger populations. This is due to the fact that these cities have well-developed transportation infrastructure, favorable market conditions, strong policy support, high population density, and abundant resources, enabling the NEVs Pilot Policy to fully realize its emission reduction potential. These factors have also been identified in previous literature as crucial determinants for promoting NEVs adoption in different counties. For instance, studies conducted by Lemphers et al. [82] on the promotion of transportation electrification in Norway, California, and Quebec have highlighted the significance of government-supportive policies, infrastructure construction, and the encouragement of NEVs manufacturer development in NEVs promotion. Additionally, research conducted by Chakraborty et al. [83] on factors affecting NEVs acceptance in India has identified service infrastructure improvement, financial incentives, and consumers’ environmental awareness as key drivers for NEV adoption in India. Conversely, the NEVs Pilot Policy has a relatively weakened effectiveness in other cities, particularly those with lower economic development levels and smaller populations. In these cities, current policies should be maintained and optimized to improve emission efficiency in the future. These cities should continue to intensify their efforts in promoting NEVs by considering their unique characteristics. They should formulate targeted NEVs promotion policies that specifically address infrastructure development, NEVs manufacturers, and NEVs consumers, aiming to enhance the effectiveness of NEVs promotion policies and achieve significant emissions reduction in urban transportation. In terms of urban transport development, the findings show that the policy effect is more obvious on TCE and PCE in the cities where the urban transport development is higher, while it is more significant on the CEI in the cities where the urban transport development is lower. Therefore, to ensure the sustainability of resource development and account for the heterogeneity of the NEVs Pilot Policy in different cities, the Chinese government should prioritize targeting cities with higher levels of economic development and larger populations. To further enhance the effectiveness of NEVs Pilot Policy on reducing urban transport carbon emissions, the focus should be on promoting NEVs in the private transport sector of the provincial capitals and municipalities, and the cities with higher economic and urban transport development levels. Meanwhile, it is more practical to promote the NEVs in the public transport sector of developing cities with lower urban transport development levels, which could facilitate the construction of low-carbon transport systems.

The study is also subject to some limitations but also opens avenues for future research. First, due to data availability, this paper only measured the direct urban transport carbon emissions caused by the burning of fossil fuels, which did not consider the indirect urban transport carbon emissions such as electricity use. China’s current power structure is still dominated by thermal power generation, leading to corresponding carbon emissions during the power production stage. When more detailed data on the proportion of coal power in charging amount is available, the effectiveness of the NEVs Pilot Policy on urban transport carbon emissions can be estimated again to generate more accurate and convincing results. Second, this study is based on the Chinese market, thus the extension of the findings to other developing countries or emerging markets needs to be further verified. Future research on other markets worldwide could be compared with our results from China. Third, there may be other factors that could influence policy effectiveness, such as public awareness, infrastructure development, and consumer behavior. However, these factors were not considered in this study. We will further explore the relationship between these factors and policy effectiveness in future research.