Study on the Theme Evolution and Synergy Assessment of China’s New Energy Vehicle Policy Texts

Abstract

Drawing on data from 133 Chinese New Energy Vehicle (NEV) policy documents from 2007 to 2023, this study utilizes Dynamic Topic Modelling (DTM), social network analysis and a quantitative model to investigate the evolutionary path of policy themes and the coordination effects. The following results were obtained. (1) A thematic cross-sectional analysis identified six core policy themes, namely, coordinated promotion of technology and finance, industry development and safety standardisation, market service and technical support systems, promotion strategy and urban cluster development, industrial capital and safety supervision mechanisms, and policy support and market expansion. The analysis also mapped the distribution of hot spots within these themes. (2) The keyword co-occurrence network of the NEV policy indicated that the network structure evolved from an initial ‘overall dispersion–theme concentration’, comprising 16 policy themes, to an ‘overall stability–theme coordination’, consisting of 14 policy themes. (3) The coordination degrees across the three types of policies exhibited a consistent upward spiral, with the comprehensive coordination index surging from 30 in 2007 to 951 in 2023, underscoring the complementary effects among policy instruments. These conclusions offer valuable insights for government departments to understand NEV development trends and dynamically adjust policy themes accordingly.

1. Introduction

Amid the growing concerns over global greenhouse gas emissions and energy security, a key challenge lies in leveraging energy policies to decouple economic growth from environmental degradation and to establish a clean, low-carbon, safe and efficient energy system [1]. China has begun to critically reassess the pitfalls of its long-standing model of high energy consumption and heavy pollution [2]. Consequently, energy policies oriented towards green and low-carbon development have gradually evolved into a core component of the national strategy, with a focus on the steady and proactive advancement of the ‘dual carbon’ targets, grounded in the country’s energy resources [3,4]. Owing to their clean, intelligent, and low-carbon footprint [5,6], new energy vehicles have emerged as a pivotal tool in achieving the dual carbon objectives and driving energy conservation and emissions reduction. In response, government bodies have introduced a series of policies, including the ‘New Energy Vehicle Industry Development Plan (2021–2035)’ and the ‘Notice on Launching the 2024 New Energy Vehicle Rural Campaign’ [7]. Despite the numerous policies on the NEV, scholarly analyse remains relatively sparse, often focusing on policy design and formulation [8], supervision and control during implementation [9], and assessments of policy outcomes [10], while neglecting the quantitative analysis of policy themes over time, the dynamic evolution of thematic network structures, and the synergies between policies. Therefore, objectively evaluating the goals, thematic focus, efficiency and synergy of the NEV policies is crucial for enhancing policy frameworks and promoting the high-quality development of the new energy vehicle sector.

Currently, the formulation and evaluation of the NEV policies have garnered notable attention. Research in this area primarily focuses on three levels: The first level is analysing NEV policy texts to achieve sustainable development goals. By utilising text mining and content analysis methods, policy keywords can be accurately extracted, and core themes can be summarised from the external structure and internal semantics of policy texts. NEV policies encompass various domains such as transportation, energy and the environment, making the policy content diverse and complex [11]. Considering the main actors, regulatory goals and policy tools of China’s NEV policies is essential [7,12] whilst learning from the experiences of the United States and Europe [13,14]. This scenario helps in constructing a comprehensive policy text analysis framework that includes policy tools, actors and themes [7,15]. The second level is analysing the policy implementation process from a dynamic perspective. The use of policy networks offers key insights into policy effectiveness [16]. Based on policy network theory, exploring the network interactions during the formulation and implementation stages of NEV policies reveals the dynamic relationships and influences among stakeholders [17,18]. It also identifies deviations and influencing factors in the policy implementation process, such as differences in policy understanding, uneven enforcement and local characteristics [19,20]. The third level is evaluating the implementation effects of NEV policies, considering both macro and micro-level impacts. Some scholars analyse the overall effects of NEV policies and the differences in impacts among various types of policies [21,22]. Effective policy evaluation requires a comprehensive consideration of design goals and the rational use of policy tools to ensure alignment with market demand and technological development stages. Additionally, examining the social and public opinion impacts of NEV policies is crucial. The responses of different resident groups reveal the differentiated impacts of policies among various social groups [23,24]. Public opinion on online and social media highlights key public concerns, emphasising the need for timely adjustments to manage public expectations [20,25]. The success of policies depends not only on economic incentives or technical support but also substantially on public perception and opinion guidance.

In this study, 133 energy policy texts issued in China from 2007 to 2023 serve as the research sample. By employing text analysis, social network analysis, and DTM, the study segments the policy texts, identifies key themes in the NEV policies, and maps the co-occurrence of these themes across different development stages. It then analyses the evolutionary patterns and projects future trends of these policy themes. Additionally, a coordination evaluation model is also employed to analyse the synergistic effects of policy measures.

The innovations of this study compared to previous research on the NEV are fourfold: First, it integrates policy text coding techniques with machine learning methods to construct a comprehensive panel dataset of China’s NEV policies from 2007 to 2023. This dataset, with its extensive coverage and large volume, overcomes the limitations of previous studies that relied solely on comparative analysis [26] and case studies [27] to explore the underlying logic and evolution of these policies. Second, DTM is used for topic modelling and natural language processing, allowing for the precise identification and extraction of high-frequency terms within policy texts and enabling the calculation of trends in topic consistency and the generation of topic-word matrices. This approach overcomes the limitations of Latent Dirichlet Allocation (LDA) in dynamically analysing topic evolution over time [28], providing a comprehensive and dynamic analysis of the evolution of policy themes in the NEV sector. Third, social network analysis is utilized to construct a co-occurrence matrix of NEV policy themes in China, resulting in detailed network diagrams that illustrate the coordinated evolution of themes and the changing roles of core themes at different stages. Fourth, the study establishes a three-dimensional coordination evaluation model that links “policy intensity—policy objectives—policy measures”, expanding upon traditional methods that focused primarily on external structure and internal semantics [29]. This model offers a more scientific and holistic framework for evaluating policy coordination, thereby increasing the validity and rigor of the study’s findings. Finally, the study’s results provide valuable insights into the current optimization of goals and the coordination among various policy tools within China’s NEV policy landscape. These insights will help identify strengths and weaknesses in the policy implementation process, offering empirical guidance for improving policy coordination, enhancing the policy framework, and supporting the implementation of a green and low-carbon development strategy.

2. Materials and Methods

2.1. Data Acquisition and Preprocessing

2.1.1. Policy Text Data Acquisition

The Peking University Magic Weapon Database (https://www.pkulaw.com/; accessed on 16 May 2024) is an authoritative Chinese legal research platform that offers convenient, accurate, and comprehensive legal information services [16,30]. The database contains roughly 760,000 documents, including laws, departmental regulations and local regulations from 1949 to the present. This study began with the retrieval of the ‘New Energy Vehicle Production Access Management Rules’ issued in 2007, and by the end of 2023, 298 relevant policy documents were identified from the database. After cross-referencing these documents with energy policies from the Wanfang Data platform [31] and expanding the keywords related to the strength, objectives, and measures of new energy policies based on the policy synergy evolution evaluation model [32], we excluded duplicate entries from the vehicle model directories exempt from purchase tax, policies with low relevance for evaluation, and texts with limited legal authority, such as announcements and approvals. Ultimately, we compiled a dataset of 133 policy documents. The distribution of these texts peaked in 2016 with 16 texts, whilst 2007 and 2013 had the least, with only one text each. Figure 1 illustrates the annual issuances of the NEV policies.

2.1.2. Text Data Preprocessing

Topic modelling is a widely used statistical tool for identifying latent variables within large volumes of unstructured text data [33]. Prior to constructing the topic model, the text data need to undergo programmatic preprocessing, as shown in Figure 2.

(1) Acquisition and Cleaning of Policy Texts: Using Python3.11.4, we scraped policy texts related to new energy vehicles from the Pkulaw website, including those with keywords such as ‘new energy vehicles’, ‘electric vehicles’ and ‘clean energy vehicles’. We also matched policies from other websites such as the National Energy Administration and the Ministry of Industry and Information Technology. Irrelevant policies were manually removed to obtain cleaned and relevant policy text data. (2) Further Processing and Vectorisation: The cleaned texts were standardised on the basis of title, publication date, issuing department and main content. We then performed tokenisation, part-of-speech tagging, word frequency filtering and temporal slicing to generate timestamped keywords. Using the gensim library’s corpora module, the tokenised texts were converted into word frequency vectors based on the bag-of-words model. (3) Topic Identification Using DTM: A Python programme was used to traverse the corpus and keywords for each topic, creating a keyword matrix. DTM was used for temporal topic analysis and modelling, incorporating word2vec embeddings and the softmax function from neural networks for topic semantic calculations. This process allowed us to build topic models and identify the topics within the texts. (4) Identifying Topic Evolution Patterns: Gephi0.10.1, a specialized tool for knowledge mapping, effectively visualizes and analyses the relationships and evolution of topics within policy texts, offering powerful network visualization capabilities [34]. While other data visualization tools, such as Tableau and Power BI, and text analysis software like NVivo12.0 can process similar data, they are less effective than Gephi0.10.1 in conducting network analysis and deep topic exploration. In this study, we used a knowledge graph software tool such as Gephi0.10.1 to create co-occurrence network graphs of topics at different stages, enabling a visual analysis of the evolution of policy topics.

2.2. Research Methods

2.2.1. DTM Model

The DTM extends the LDA model and is an unsupervised machine learning approach designed specifically to track and analyse how topics evolve over time within a document collection [35]. In practical terms, policy text data are first discretized into chronological time slices, with the assumption that both the distribution and content of topics evolve gradually between adjacent time slices [36], allowing for the identification of topic chains across sequential datasets. The topic evolution model is described by Equation (1).

$${\beta }_{t,k}\mid {\beta }_{t-1,k}\sim \mathcal{N}\left({\beta }_{t-1,k},{\theta }^{2}I\right)$$

(1)

In this equation, K denotes the number of topics, N represents the number of words in a document, β is the word distribution for each topic, and α is the topic distribution for each document. When α is applied to a document, θ represents the document’s topic model, which follows a multinomial distribution, with α serving as the conjugate prior for θ. For each word N in the document, the topic l is selected through θ’s multinomial distribution, and the corresponding words are selected through the multinomial distribution of l and β. Subsequently, appropriate model parameters—such as the number of topics and iterations—are configured in Python, and the text data are segmented into various time periods. For each time slice t, the document topic distribution αt and the associated word distribution βt,k are contingent on αt − 1 and βt − 1,k from the previous time slice t − 1, with βt,k for the current slice being derived from the previous slice’s βt − 1,k.

To determine the optimal number of topics, we apply the “coherence” measure as introduced by Mimno et al. [37], which evaluates the semantic similarity of frequently co-occurring words within topics. A higher coherence score indicates a more interpretable model, calculated using the following formula:

$$\phi \left(t,{\lambda }^{\left(t\right)}\right)={\sum }_{m=2}^M{\sum }_{l=1}^{M-1}log{\displaystyle \frac{Z\left({\lambda }_m^{\left(t\right)},{\lambda }_l^{\left(t\right)}\right)+1}{Z\left({\lambda }_l^{\left(t\right)}\right)}}$$

(2)

Here, $Z\left(\lambda \right)$ represents the number of documents containing the word λ, $Z\left(\lambda ,{\lambda }^{\prime }\right)$ represents the co-occurrence frequency of the words λ and λ′, and λ_(t) refers to the list of words most likely associated with topic t.

Finally, the DTM model produces a “Document-Topic” matrix that estimates the probability of each document belonging to a specific topic. By comparing these probabilities against a predetermined threshold, we can identify the supporting documents for each topic. Building on the work of McCallum et al. [38], we set a 10% threshold and calculate topic strength using the support index, as shown in the following formula:

$${MI}_{\left(t\right)}^{\lambda }={\displaystyle \frac{{Doc}_{\left(t\right)}^{\lambda }}{{Doc}_{\left(t\right)}}}$$

(3)

In this formula, ${MI}_{\left(t\right)}^{\lambda }$ denotes the strength of topic l in the current time slice t, with the numerator representing the number of documents under topic l in time slice t, and the denominator representing the total number of documents in the same time slice.

2.2.2. Policy Effectiveness and Policy Synergy Evaluation Model

In the current field of policy effectiveness research, the three-dimensional policy text quantification evaluation model proposed by Peng et al. [32] is highly representative. This model emphasizes that policy strength is mainly determined by the policy-issuing body and type, while policy goals and measures depend on the clarity of the policy content. This study draws on the method of Peng et al. [32], constructing a three-dimensional policy effectiveness quantification evaluation model based on “policy strength—policy goals—policy measures”. Using the scoring criteria from

Table 1. Scoring Criteria for Policy Strength and Policy Goals.

Evaluation Index	Score	Scoring Standard
Policy Intensity	5	Laws enacted by the National People’s Congress and its Standing Committee are assigned a score of 5.
	4	Regulations, directives, and rules issued by the State Council, as well as orders from ministries, receive a score of 4.
	3	Provisional regulations, rules, and other decision-making documents like plans, decisions, opinions, measures, and standards from the State Council, as well as regulations, rules, and decisions from ministries, are scored as 3.
	2	Opinions, measures, plans, guidelines, provisional regulations, detailed rules, conditions, and standards from ministries are assigned a score of 2.
	1	Notices, announcements, and plans issued by sub-ministerial units are scored as 1.
Policy Goals	5	Clear policy goals with quantifiable metrics, including specific targets like project scale, electricity generation, and promotion numbers, are scored 5 when all metrics are present.
	3	Clear goals but with incomplete quantifiable metrics are scored 3, based on the sum of available numerical indicators.
	1	If the policy only outlines broad visions or expectations without specific targets, it receives a score of 1.

and

Table 2. Scoring Standards for Policy Measures.

Type of Policy Measure	Score	Scoring Standard
Command and Control	5	(1) Enforces industry entry thresholds, specific conditions, and evaluation standards for the new energy vehicle sector; (2) Sets the process for project inspection, assessment criteria, and supervisory methods; (3) Strictly enforces environmental impact assessments for new energy vehicle projects; (4) Defines the use and management of special funds for new energy vehicles; (5) Implements detailed mandatory management measures for these projects. Each criterion scores 1 point, with a total score of 5.
	3	(1) Proposes the need to establish entry thresholds, conditions, evaluation standards, and oversight plans; (2) Outlines the management of special funds and environmental assessments but lacks detailed implementation or follow-up; (3) Suggests policies to support new energy vehicle projects. Each criterion scores 1 point, with a total score of 3.
	1	The document only references the need for command-and-control measures, such as entry thresholds, without detailing implementation or specific plans, scoring 1 point.
	0	No command-and-control measures are mentioned, resulting in a score of 0.
Economic Incentives	5	(1) Offers comprehensive financial support, including detailed budgets, subsidies, and incentives; (2) Specifies exact amounts and management methods for subsidies, investments, and rewards; (3) Promotes new energy vehicle projects through pricing, fees, and metering strategies; (4) Encourages vehicle replacement programs via price and fee adjustments; (5) Supports the development of charging infrastructure and technological innovation. Each criterion scores 1 point, with a total score of 5.
	3	(1) Proposes financial support to promote new energy vehicle adoption; (2) Suggests price adjustments to encourage vehicle purchases; (3) Develops plans for charging infrastructure and innovation but without detailed management strategies. Each criterion scores 1 point, with a total score of 3.
	1	Mentions financial support and price adjustments as incentives but lacks detail on specific measures, scoring 1 point.
	0	No economic incentives are mentioned, resulting in a score of 0.
Guidance and Demonstration	5	(1) Establishes clear implementation methods guiding local governments or enterprises in project development; (2) Develops management plans for demonstration projects or pilots; (3) Creates technology promotion catalogs for new energy vehicles; (4) Establishes a comprehensive policy system; (5) Develops industry development catalogs. Each criterion scores 1 point, with a total score of 5.
	3	(1) Advocates for the promotion of new energy vehicle projects; (2) Suggests the development of industry catalogs; (3) Proposes guidance policies related to new energy vehicles. Each criterion scores 1 point, with a total score of 3.
	1	Only mentions the need for demonstration policies and technologies without detailed measures, scoring 1 point.
	0	No guidance measures are mentioned, resulting in a score of 0.

, the policy measures in the three dimensions are scored. To more clearly demonstrate the changing trends in the effectiveness of the NEV policies, this paper divides policy effectiveness into overall effectiveness and average effectiveness, and uses Formulas (4) and (5) to calculate the policy synergy for the NEV policies in each year, which is used to evaluate the overall effectiveness and average effectiveness of China’s NEV policies.

$$PEi={\sum }_{j=1}^N{(g}_j+m_j)p_j,i=[2007,2023],$$

(4)

$$APEi={\displaystyle \frac{{\sum }_{j=1}^N{(g}_j+m_j)p_j}{N}},i=[2007,2023].$$

(5)

where i represents the year in which the energy policy was implemented, N represents the total number of energy policies in year i, j represents the j-th policy implemented in year i, g_j represents the goal score of the j-th policy, m_j represents the measure score of the j-th policy, p_j represents the strength score of the j-th policy, PEi represents the overall effectiveness score of all policies in year i, and APEi represents the average effectiveness score of policies in year i.

In addition, this study employs a policy synergy evaluation model, which assesses the overall execution efficiency and synergy of policies by quantifying the interaction between different policy measures. The model is set according to the evaluation framework for the synergy of China’s NEV policies, calculating the synergy of policy measure combinations, including command and control, economic incentives and guidance and demonstration types. Policy synergy generally refers to the variety of policy tools used within a single policy; the more diverse the tools, the higher the policy synergy [39]. Through Formulas (6) and (7), this study calculates the policy synergy of new energy vehicle policies for each year. The entire research process emphasizes the systematicity and scientific rigor of the method, ensuring a deep understanding of policymakers’ intentions, policy goal adjustments, and policy tool selection from multiple dimensions, thereby providing empirical support and a theoretical basis for the formulation and optimization of the NEV policies.

$${PC}_{ilk}={\sum }_{j=1}^N{p}_j\times m_{jl}\times m_{jk} l\ne k,i=[2007,2023],$$

(6)

$${PC}_{ilkh}={\sum }_{j=1}^N{p}_j\times m_{jl}\times m_{jk}\times m_{jh} l\ne k\ne h,i=[2007,2023].$$

(7)

In Formula (6), PC_ilk represents the synergy between the l-type and k-type policy measures within the NEV policies in year i. Here, N denotes the total number of such policies in that year, p_j refers to the policy strength score for the jth policy, and m_jl and m_jk are the scores for the l-type and k-type measures within the jth policy. The variables l and k can represent any two of the following policy types: command-and-control, economic incentives, or guidance and demonstration. In Formula (7), PC_ilkh represents the synergy among all three policy types in year i, with N again representing the number of policies, p_j the strength score for the j-th policy, and m_jl, m_jk, and m_jh being the respective scores for the command-and-control, economic incentives, and guidance and demonstration measures within that policy.

Policy strength is primarily concerned with assessing the administrative impact and legal authority of NEV policies. Typically, policies issued by higher-ranking bodies have greater influence. According to the “Regulations on the Procedures for the Formulation of Rules” by the State Council and the evaluation criteria provided by Lan [31], policy strength is categorized into five levels, with scores ranging from 1 to 5. Policy goals assess the clarity and specificity of the objectives set out in the policy texts. Generally, more clearly defined and quantifiable goals receive higher scores. Given that these policies are focused exclusively on new energy vehicles, and the criteria for scoring policy goals are relatively simple, with only three levels, the scoring is designed to ensure clear distinctions in the evaluation outcomes. Accordingly, policy goals are assigned scores of 1, 3, or 5, which differs from the scale used for policy strength. Table 1 details the specific scoring criteria.

Policy measures are the tools used by government departments to achieve set goals, aiming to assess the actual effectiveness, operability and efficiency of new energy vehicle policies. This study classifies the NEV policies into three categories―command and control, economic incentives and guidance and demonstration―based on the environmental policy measures classification criteria defined by the Organization for Economic Cooperation and Development (OECD) in 1996. Each policy category is assigned values according to its executability, level of detail and implementation strength. Table 2 details the specific assignment standards.

2.2.3. Social Network Analysis

Social Network Analysis (SNA) is a tool used to study social structures by analysing the relationships (edges) between entities (nodes) to construct a network. This method quantifies and describes these relationships to reveal the characteristics of social structures [40]. SNA is used primarily to predict the decisions and behaviours of nodes within the network as well as the formation and evolution of interactions between nodes. It also uncovers interaction patterns among individuals or groups within the system, analyses the causes of these patterns and attempts to identify their impacts and outcomes [41].

In this study, we primarily use Python3.11.4 for DTM modelling to identify and analyse themes in new energy vehicle policies, constructing a theme co-occurrence matrix. We then use Gephi0.10.1 to visualise the collaborative network graph of policy themes.

2.3. Sample Stage Division

As China optimises its national energy structure and advances a green, low-carbon circular economy, the NEV policies must adapt to changes in market mechanisms and the creation of value synergies [42]. The demand for energy structures varies across different periods of national economic and social development, leading to differences in specific policy directions and functions [43]. Therefore, based on the adjustment of national strategic development policies and key time nodes, the evolution of China’s NEV policies is divided into the following four stages:

Policy Initiation and Exploration Stage (2007–2012): The promulgation of the ‘New Energy Vehicle Production Access Management Rules’ in 2007 marked the beginning of China’s focus on technological advancements in automotive products, promoting energy conservation and sustainable development and encouraging the research, development and production of NEVs. In 2009, the ‘Automobile Industry Adjustment and Revitalization Plan’ set a goal of achieving a production capacity of 500,000 NEVs, successfully establishing the electric vehicle as a strategic emerging industry. The 2012 ‘Energy-Saving and New Energy Vehicle Industry Development Plan (2012–2020)’ established a primary focus on pure electric drive technology, marking a new development stage for the NEV industry. During this stage, policy implementation bodies were relatively few, and policy agenda and goals were relatively simple.

Policy Support and Incentive Stage (2013–2016): During this period, supportive industry policies substantially accelerated the development of NEVs with notable success in demonstration operations. In 2013, the government increased its support for NEVs, launching pilot projects in several cities through R&D investments and direct subsidies and prioritising their deployment in the public sector. The release of the ‘Catalog of New Energy Vehicle Models Exempt from Vehicle Purchase Tax’ stimulated consumer interest, with NEV sales exceeding 330,000 units, accounting for nearly 60% of global NEV sales. This figure positioned China as the world’s largest NEV market. In this stage, the diversity of policy implementation bodies increased, and policy agenda and goals became increasingly varied and comprehensive.

Policy Transformation and Internationalisation Guidance Stage (2017–2020): Starting in 2017, China’s NEV industry accelerated its transition to a market-driven approach with increased openness to foreign investment and a gradually emerging competitive landscape. Policies such as the ‘Parallel Management of Average Fuel Consumption and New Energy Vehicle Credits for Passenger Car Enterprises’ and the ‘Special Administrative Measures for Foreign Investment Access (Negative List) (2018 Edition)’ were introduced to support this transition and openness. By the end of 2020, China’s NEV market featured a competitive landscape with four main camps: traditional independent brands, new car-making forces, joint ventures and luxury brands. This stage was marked by a noticeable diversification of policy implementation entities and enhanced integration and combination of policy agenda and goals.

Policy Systematisation and Synergisation Stage (2021–2023): To further support the high-quality development of the NEV industry and promote NEV adoption, various government departments jointly introduced a series of policies from the perspective of the NEV industry chain. These policies include the ‘Notice on Improving the Financial Subsidy Policy for the Promotion and Application of New Energy Vehicles’ and the ‘Comprehensive Utilization Industry Specification Conditions for the Recycling of New Energy Vehicle Power Batteries’. These policies not only met the requirements for optimising and upgrading the national automotive industry but also enhanced the synergy effects among NEV enterprises and boosted the industry’s international competitiveness. During this stage, the synergy and networking of policy implementation entities became prominent with a focus on NEV safety, the civilised consumption of new energy and ecological civilisation.

3. Results

3.1. Theme Analysis

Determining the number of themes is crucial for extracting themes from new energy vehicle policies. This study calculates the average theme coherence for different numbers of themes across the four stages to determine the optimal number of themes for the DTM. Higher theme coherence values indicate better performance of the thematic language model [44]. As shown in Figure 3, when the number of themes is six, the average theme coherence across the four stages is the highest. Therefore, we set the number of themes to six.

Using DTM, we then obtain the ‘theme-word matrix and word weights’ for each stage of energy policies. The top 10 high-probability characteristic words for each theme are organised to identify the theme labels that best match these high-probability words, as shown in

Table 3. Energy Policy Theme Keywords Table.

Theme	Theme Keywords
Coordinated Promotion of Technology and Finance	Financial subsidy, implementation plan, interim regulation, technological innovation, lithium-ion, financial fund, charging station, design standard, commission office, fixed price
Industry Development and Safety Standardisation	Standardisation, industry regulations, comprehensive utilisation, subsidy funds, bidding, after-sales service, design standard, vehicle and vessel tax, electric vehicle safety, public consultation
Market Service and Technical Support System	Import volume, earnest implementation, parking space, financial services, product quality, electronic information, consultation phone, Southern Power Grid, technical regulations, engine
Promotion Strategy and Urban Cluster Development	Urban cluster, bidding, fuel cell, power battery, compliance, implementation plan, development plan, establishment and improvement, promotion quantity, parking lot
Industrial Capital and Safety Supervision Mechanism	Capital funds, target value, ecological environment, engineering technology, open type, limited liability, safety supervision, tax exemption, network security, plug-in hybrid
Policy Support and Market Expansion	Product certification, vehicle and vessel tax, technological innovation, power battery, financial subsidy, production volume, motor vehicle, interim measures, import volume, transportation

.

(1) Coordinated Promotion of Technology and Finance. This theme examines the interplay between innovation and subsidy policies for new energy vehicles. Technological innovation encompasses the advancements in lithium-ion battery technology and the design standards for charging station construction, which provide essential technical support for NEVs. Financial subsidies represent government economic support, including funding to promote the widespread adoption of these vehicles. Moreover, government agencies must ensure price transparency of vehicle testing institutions and conduct targeted inspections of subsidy funds through the financial supervision office to guarantee the transparency and effective use of these funds. Overall, this theme investigates how effectively coordinating technological advancements and financial support strategies can accelerate the commercialisation and market penetration of new energy vehicles.

(2) Industry Development and Safety Standardisation. This theme addresses aspects such as the protection of electronic information and vehicle safety design specifications. In the new energy vehicle industry, standardisation and industry norms are essential for ensuring vehicle safety and promoting healthy industry development. Standardisation aids in the efficient use of industry resources, ensures the effective utilisation of subsidy funds and maintains fairness in the bidding process. The integration of industry development and safety standardisation not only enhances product quality but also boosts consumer confidence.

(3) Market Service and Technical Support System. This theme focuses on the service quality and technical support system of the new energy vehicle market. It includes the availability of financial services, such as loans and insurance for NEVs, and guarantees product quality, especially the reliability of charging and swapping systems. Additionally, the presence of after-sales consultation hotlines and support from the Southern Power Grid demonstrates the completeness of the technical support system, ensuring that users receive timely and effective assistance when using new energy vehicles.

(4) Promotion Strategy and Urban Cluster Development. This theme focuses on the promotion strategies for new energy vehicles within urban clusters and the corresponding development plans. It covers technological advancements in fuel cells and power batteries as well as comprehensive transportation development planning for urban clusters, including the construction of parking lots and charging facilities. Promotion strategies must be closely integrated with urban development plans to ensure the sustainability and effectiveness of new energy vehicle promotion.

(5) Industrial Capital and Safety Supervision Mechanism. Effective capital management and safety supervision are crucial for the sustainable development of the NEV industry. Proper capital management impacts both the operational efficiency of companies and the quality of project implementation. Safety supervision, covering aspects such as network security and vehicle safety, is essential for protecting consumer interests and enhancing the industry’s reputation. Additionally, open-ended investment funds and tax exemption policies play a crucial role in attracting investment and supporting industry growth.

(6) Policy Support and Market Expansion. This theme explores how policies facilitate the expansion of the new energy vehicle market. Key points include technological innovation and financial subsidies, particularly advancements in power battery technology and the efficient allocation of financial resources. Additionally, preferential policies on vehicle and vessel taxes, the implementation of interim measures and product certification for motor vehicles are crucial policy tools for promoting market growth.

After analysing the themes in the energy sector, it becomes clear that there are significant differences in the number of policy texts dedicated to each theme, as shown in Figure 4. The theme of ‘Promotion Strategy and Urban Cluster Development’ has the highest number of policy texts, totalling 31. This reflects policymakers’ strong emphasis on integrating new energy vehicle policies with urban development plans, as well as their focus on building essential urban infrastructure such as charging stations and parking facilities. Close behind are the themes of ‘Coordinated Promotion of Technology and Finance’ and ‘Market Service and Technical Support System’, with 29 and 27 policy texts, respectively. These numbers highlight the central role of combining technological innovation with financial incentives, along with enhancing market service quality and technical support, in the policy-making process for new energy vehicles. On the other hand, the themes of ‘Industrial Capital and Safety Supervision Mechanism’ and ‘Policy Support and Market Expansion’ are less prominent, with 20 and 14 texts, respectively, suggesting that these areas may receive relatively less attention and resources in practical policy implementation. The least covered theme is ‘Industry Development and Safety Standardisation’, with just 12 texts, indicating that this crucial area, despite its importance for ensuring vehicle safety and fostering the industry’s healthy growth, has been relatively underemphasized in policy discussions.

3.2. Analysis of Theme Evolution Paths

To dynamically display the evolution paths of each theme, we used Gephi0.10.1 to create co-occurrence maps of new energy vehicle policy keywords across four periods, as shown in Figure 5. In these maps, each node represents a keyword from the policy themes, and the node size reflects its weighted degree centrality. The lines between nodes indicate connections between keywords within the same policy text, with line width representing the frequency of these connections. A connection strength threshold of 1 was set, meaning two keywords must co-occur in more than one policy text to establish a connection. Thicker lines indicate more frequent co-occurrence of the two keywords in the same policy text [45].

The Policy Initiation and Exploration Stage (2007–2012) accumulated a total of 18 policy texts and 16 policy themes, characterised by an ‘overall dispersed—theme concentrated’ network structure. The close association between fuel cells and plug-in hybrids indicates that these technologies were viewed as key drivers of industry growth. The link between after-sales service and plug-in hybrids highlights the importance of quality service in enhancing consumer acceptance of new technologies. The connections of power batteries with multiple themes underscore their central role in the industry, directly impacting consumer experience and vehicle performance. Policy support during this stage was concentrated on technological innovation and subsidy funds, with the relationship between subsidies and implementation plans emphasising the critical role of financial support in advancing the industry. The co-occurrence of development planning and standardisation reflects the foundational role of technical standards in ensuring healthy industry growth and protecting consumer interests.

The Policy Support and Incentive Stage (2013–2016) accumulated 41 policy texts and 27 policy themes, characterised by an ‘overall expansion—multi-centre compact themes’ network structure. The strong connection between power battery technology and themes such as after-sales service and product quality underscores the importance of technology reliability for market acceptance. The development of fuel cell technology also depends on market support, as shown by its high co-occurrence with after-sales service and product quality. The close link between bidding and energy conservation highlights the importance of efficient standards. The substantial co-occurrence of financial subsidies with the establishment of market rules (e.g., parking lot construction and implementation plans) emphasises the role of policy support in shaping the market.

The Policy Transformation and Internationalisation Guidance Stage (2017–2020) accumulated 45 policy texts and 27 policy themes, characterised by an ‘overall efficient—balanced themes’ network structure. Collaboration between Southern Power Grid and State Grid highlighted the importance of power infrastructure in promoting new energy vehicles, ensuring the electricity supply for charging facilities. The integration of charging stations with parking lots emphasised effective urban planning, enhancing the convenience of NEV use. The connection between Southern Power Grid and limited liability companies indicated strategies for capital operations and risk management, particularly in financing and managing grid expansions or charging station construction. In terms of technological innovation, the relationship between fuel cells and plug-in hybrids underscored the crucial role of innovation in improving vehicle performance and market expansion. Policies on establishing and improving transportation infrastructure were essential for the successful promotion of the NEV market. Infrastructure projects required optimised capital management and regulatory compliance, necessitating strong collaboration between limited liability companies and the State Grid.

The Policy Systematisation and Synergisation Stage (2021–2023) accumulated 29 policy texts and 14 policy themes, with the network structure showing ‘overall stability—coordinated themes’. Under the ‘dual carbon’ strategy, close interaction and interdependence existed between NEV policies, industry support and market behaviour. The link between fuel cells and plug-in hybrids indicated a trend towards technological integration, reflecting efforts to promote diversified energy solutions. The prevalence of fuel cells and vehicle standardisation showed the close relationship between policy support and market expansion, with certification systems emphasising strict quality and safety standards. High-frequency co-occurrence of keywords related to power battery technology, fuel cells, plug-in hybrids and after-sales service pointed to the complexity of technological integration and increased market service demand. Connections between product certification, import volumes, the automotive industry and mandatory policies underscored the role of policymaking in ensuring product quality and safety.

3.3. Policy Effectiveness and Synergy Analysis

Figure 6 illustrates the significant changes in the scores for policy objectives, strength and policy measures from 2007 to 2023. The score for policy objectives began at 2.0 in 2007 and, after a period of adjustment, gradually stabilized, peaking at 2.4 in 2017. This trend indicates that over time, policy objectives became more precise and quantifiable. Regarding policy strength, the score peaked at 5.0 in 2013, signalling the government’s stronger measures to advance the NEV industry, such as boosting financial investment and providing regulatory support to facilitate widespread adoption and technological innovation. The score for policy measures also reached its highest point at 5.0 in 2013, highlighting a substantial improvement in the detail and enforcement of policy tools, including subsidies for new energy vehicle purchases and tax incentives. The progression of these scores not only reflects the growing emphasis on NEV policies but also demonstrates a continuous refinement of policy content and implementation strategies.

When evaluating the effectiveness of NEV policies from 2007 to 2023, the descriptive statistics across four stages reveal varying levels of implementation and effectiveness across different policy measures (command and control, economic incentives and guidance and demonstration), the specific data results are shown in

Table 4. Descriptive Statistical Results of Energy Policy Effectiveness.

Development Stage	Statistic	Command and Control	Economic Incentives	Guidance and Demonstration
Policy Initiation and Exploration Stage	mean	9.3	6.5	8.3
	std	3.9	2.1	3.1
	min	6.5	4.0	5.3
	max	16.0	9.3	12.0
Policy Support and Incentive Stage	mean	11.0	9.1	10.0
	std	0.7	0.8	0.6
	min	10.0	8.0	9.5
	max	11.5	10.0	10.9
Policy Transformation and International Guidance Stage	mean	12.0	8.8	9.0
	std	5.2	3.4	3.4
	min	7.8	5.0	5.5
	max	18.6	12.5	13.1
Policy Systematisation and Synergisation Stage	mean	9.5	8.5	8.5
	std	2.8	1.4	2.9
	min	7.5	7.5	6.5
	max	11.4	9.4	10.6

. During the Policy Initiation and Exploration Stage, there was significant variability in the implementation of command and control, economic incentives and guidance and demonstration policies, with command-and-control policies showing the most pronounced disparity between maximum and minimum values. This suggests early instability and inconsistency in regional enforcement. As the Policy Support and Incentive Stage began, the average scores for all policy measures increased, while standard deviations decreased significantly, indicating a more focused and targeted approach to policy implementation, leading to greater consistency and stability. In the Policy Transformation and Internationalization Guidance Stage, although command-and-control policies had the highest average score, the increased standard deviations across all policy types reflected a growing adaptability and flexibility in policy execution, likely driven by international competition and the need to respond to emerging technologies and markets, while, in the Policy Systematisation and Synergisation Stage, policy implementation became more systematic and coordinated, with further reductions in standard deviations, indicating a high level of maturity and efficiency. Overall, the implementation of new energy vehicle policies evolved from early instability into a mature, systematic approach, underscoring the importance of ongoing policy adjustments and the proactive role of policymakers in responding to market and technological shifts.

Figure 7 illustrates the evolution of the synergy among command-and-control, economic-incentive, and guidance-and-demonstration policies for new energy vehicles from 2007 to 2023. Despite some fluctuations, the overall synergy of these policies exhibited an upward spiral trend. The combined synergy value of the three policy types surged from 30 in 2007 to 951 in 2023, highlighting the profound collaborative effect achieved by integrating command-and-control, economic-incentive and guidance-and-demonstration policies in driving policy goals.

The synergy between command-and-control and economic-incentive policies increased from 10 in 2007 to 291 in 2023, indicating a growing integration of economic incentives, such as tax benefits and subsidies, with regulatory requirements, such as emission standards. The synergy between command-and-control and guidance-and-demonstration policies rose from 30 in 2007 to 281 in 2023, emphasising the importance of strategies that promote new technologies and practices through demonstration projects. The synergy between economic-incentive and guidance-and-demonstration policies also showed notable growth, with the synergy value rising from 6 in 2007 to 199 in 2023, underlining the trend of enhancing policy effectiveness through joint research and technology sharing.

4. Discussion and Conclusions

4.1. Discussion

In the face of global climate change and the ‘Dual Carbon’ strategy, China’s energy governance is confronting significant challenges in achieving a green and low-carbon transition. Promoting the widespread adoption of the NEV has therefore become a key strategy in the nation’s ecological civilization efforts. This study, drawing on 133 policy texts related to NEVs from 2007 to 2023, utilizes DTM, synergy evaluation, and social network analysis to investigate the dynamic evolution of policy themes and their coordination.

Firstly, this study departs from traditional approaches, such as theoretical analysis, literature reviews, and knowledge mapping [12,26,27], which focus on exploring the conceptual framework, current status, and optimization paths of discourse power. Instead, it employs quantitative methods like content analysis, text mining, and social network analysis to delve into the structural logic of discourse power, identifying high-frequency themes and their associated keywords. Moreover, the study explores the interconnections among the six core policy themes. For instance, the technical support system serves as the core driving force, whilst industrial capital and safety supervision mechanisms and industry development and safety standardisation provide foundational support. Policy support and market expansion offer institutional guarantees. The shift in policy focus from solely technical support and financial incentives to a broader inclusion of market services and safety standardisation reflects policymakers’ responsiveness to industry dynamics and market maturity. Overall, the six major themes primarily align with two broad areas: ‘low-carbon greening’ and ‘technological intelligence’, which resonate with national strategies to enhance innovation in intelligent internet technologies within the new energy vehicle sector and to promote green development. This dynamic adjustment underscores the logical and hierarchical nature of China’s strategic planning for new energy vehicles.

Secondly, the evolution of thematic networks shows a clear transition from the ‘overall dispersion–theme concentration’ pattern in the initial stages to an ‘overall stability–theme coordination’ structure by the fourth stage. This progression supports the conclusions of [2,6,12] that new energy vehicle policy themes have been consistently adjusted in response to temporal milestones and developmental needs, demonstrating a strong sense of continuity and advancement within the policy framework. In the first stage, the theme network is characterised by ‘overall dispersion–theme concentration’, indicating a small number of policy themes with weak cohesion. Policies during this period focused on technological innovation, financial support, service quality and market promotion strategies to foster technological maturity and increase market share, creating a favourable environment for sustainable industry development. In the second stage, the theme network exhibits an ‘overall expansion-multi-centre compactness’ structure, showing a larger number of policy themes with poor synergy. This stage required continuous technological innovation, high standards in market services and improved regulation to boost consumer confidence and market penetration. The third stage also shows an ‘overall expansion-multi-centre compactness’ structure, reflecting the interaction of various factors during the market transition. Adapting policies and industrial strategies to local conditions is needed; they must also be reasonably planned. In the fourth stage, the theme network demonstrates ‘overall stability-theme coordination’, indicating remarkable diversification and strong coordination of policy themes. The policy themes offer valuable insights into advancing new energy vehicle policies under the ‘Dual Carbon’ strategy, while also stressing the importance of further enhancing the synergy between technological innovation and financial support, strengthening industry standards and safety regulations, and improving service quality and technical support systems.

Finally, by employing a three-dimensional ‘Policy Strength—Policy Objectives—Policy Measures’ synergy evaluation model to assess the coordination of new energy vehicle policy measures, this study goes beyond the traditional focus on external structure and internal semantics [29]. This model proves to be more effective in addressing the coordinated implementation of various new energy vehicle industry policies (command and control, economic incentives and guidance and demonstration) by adapting them to local conditions. The evaluation and analysis of the synergy of new energy vehicle policy measures reveal that synergistic policies are more effective in addressing the current challenges in the new energy vehicle industry than single-category measures. From 2012 to 2017, the rapid increase in policy synergy was closely related to the government’s focused support, technological advancements and higher market acceptance. The government implemented various comprehensive policies, such as tax reductions, direct subsidies and stricter emission standards, which highly boosted the development of NEVs. However, the decline in synergy afterwards was due to the reduction in subsidies and technological bottlenecks. As the market matured and subsidy policies were adjusted, the industry experienced a brief adjustment period, reducing the interaction between policy and market. After 2019, with the adjustment of new policies and the recovery of market demand, including increased support for new technologies (e.g., battery and autonomous driving technologies) and improvements in the infrastructure for new energy vehicles (e.g., charging stations), the synergy increased again. The changes in the synergy values between command-and-control and economic-incentive policies stem from the rapid development of the NEV industry, which has revealed issues such as uneven regional development and inadequate power infrastructure. In response, the government has adopted command-and-control policies to strictly control market entry, enhance the approval and management of NEV development projects and actively provide financial support and policy incentives to approved projects. The synergy between command-and-control and guidance-and-demonstration policies highlights the growing importance of promoting new technologies and practices through demonstration projects. These projects require strong governmental oversight to ensure their authenticity and prevent the misuse of funds. Additionally, policies need to be aligned with national strategic development goals and ecological sustainability requirements to improve the guidance system for the NEV industry. The synergy between economic-incentive and guidance-and-demonstration policies underscores the trend of enhancing policy effectiveness through joint research and technology sharing. The NEV industry’s advantages in low-carbon energy savings, combined with government financial subsidies and price incentives, have accelerated the establishment of demonstration bases and the promotion of NEVs. This approach encourages local governments to attract investment, build manufacturing facilities and speed up the implementation of NEV projects.

4.2. Conclusions

Examining the spatial and temporal dimensions, the NEV policy themes in China from 2007 to 2023 have evolved through four distinct stages, each reflecting six major themes: coordinated promotion of technology and finance, industry development and safety standardisation, market service and technical support systems, promotion strategy and urban cluster development, industrial capital and safety supervision mechanisms, and policy support and market expansion. This progression underscores China’s comprehensive and systematic approach to the development and deployment of new energy vehicles. From the perspective of dynamic evolution, the keyword co-occurrence analysis across these four periods shows a notable increase in the number of policy themes, moving towards a structure characterized by ‘overall stability–theme coordination’. Policy coordination evaluations reveal a consistent upward spiral in the degree of coordination, indicating that policy tools have become increasingly complementary over time.

4.3. Limitations and Avenues for Future Research

This study focuses on new energy vehicle policies, employing text analysis, social network analysis, and DTM to identify core themes across different periods, trace the dynamic evolution of policy themes, and assess the coordination among various policy types. These insights provide valuable data and theoretical references for future innovation and deeper research in this field. However, certain limitations persist in the current research on the NEV policies, leading to the following suggestions for further development.

Firstly, research themes should be more systematically integrated. Analysis of high-frequency terms and thematic identification reveals a broad yet somewhat fragmented scope in the NEV policy research, lacking coherence. Therefore, there is a need to develop a more systematic theoretical and methodological framework tailored to the specificities of the NEV policies, integrating relevant theories and methods from related disciplines to guide and enhance ongoing research.

Secondly, future studies should broaden the scope to include diverse policy implementation actors. The complexity of policy formulation and implementation, involving policy communities, intergovernmental networks, and other actors with varying roles in the governance landscape, suggests that addressing challenges in the NEV policies requires multi-actor collaboration, with government, market, and societal stakeholders working together under specific institutional frameworks.

Finally, there is significant potential for a deeper empirical analysis of the NEV policies. The current study’s evaluation of policy coordination lacks a detailed analysis of its economic and technological impacts. Future research should explore the interplay between policy coordination and factors such as economic performance and environmental impact, advancing the understanding of multi-factor interactions and contributing to research on collective entity building, ecological optimization, and high-quality economic development in the context of new energy vehicle policies.

5. Countermeasures and Suggestions

Firstly, focusing on the evolution trends of NEV policy themes and dynamically adjusting policy content and priorities are crucial. The evolution of China’s NEV policy themes reflects both the internal development patterns of the NEV industry and changes in the external environment. This evolution is vital for guiding effective policy formulation and implementation. Policymakers should closely monitor technological advancements and shifts in market demand, flexibly adjusting policy themes and focus areas to keep pace with new industry stages. Policies should also be forward-looking, using predictive technologies and market trends to anticipate future needs and set long-term goals. As technology matures and market acceptance grows, policies should transition from initial technical and financial support to comprehensive management of market operations and safety regulation. This approach not only promotes healthy industry development but also reduces market uncertainty and boosts business confidence.

The second approach comprises developing a synergistic policy framework and optimising policy combinations. By integrating and coordinating various policy measures (command and control, economic incentives and guidance and demonstration), a multi-level and multi-tool framework can be established. This framework can effectively address complex issues such as balancing environmental protection with economic development. Considering that the roles of different policy tools are essential at various stages of NEV industry development, regular evaluations of existing policies should be conducted to assess their effectiveness, and adjustments should be made based on these evaluations. Through data and case analysis, the most effective policy combinations can be identified, ensuring optimal policy configuration and maximising efficiency.

The third approach entails establishing a long-term cooperation mechanism to enhance China’s influence in the NEV sector. Information exchange and resource sharing should be strengthened between various Chinese government departments and other countries by establishing regular communication channels. This approach will improve policy coordination and execution efficiency, particularly in the technology-intensive and market-sensitive NEV field. Creating a sustainable cooperation mechanism ensures that relevant departments can effectively collaborate during the formulation and implementation of NEV policies. This cooperation is essential for addressing complex cross-sectoral issues and providing Chinese solutions for global NEV technology standards and norms. Such efforts will play a constructive role in promoting the healthy development and equitable utilisation of NEV technology.