The Impact of Environmental Incentive Policies on the Value of New Energy Enterprises—Evidence from China’s New Energy Demonstration Cities

Abstract

The world is facing increasingly severe environmental challenges, making the development of new energy a crucial trend for the future. The corporate value of new energy enterprises plays a vital role in their sustainable growth. Current environmental regulations predominantly rely on punitive measures, with limited use of incentive-based policies. This study examines China’s New Energy Demonstration City (NEDC) policy, employing panel data from listed new energy firms (2010–2023) and a difference-in-differences (DID) approach to quantify incentive-based policies effects. The results demonstrate that the NEDC policy significantly enhances the corporate value of new energy enterprises, the findings are robust to multiple tests. The policy’s impact exhibits notable heterogeneity: state-owned enterprises (SOEs), large firms and firms in regions with stringent environmental regulations benefit more. Mechanism analysis reveals that the policy alleviates financing constraints and encourages green transformation, thereby boosting corporate value. This study provides empirical evidence supporting incentive-based environmental policies.

1. Introduction

The world is facing increasingly severe environmental challenges, making the development of new energy a critical trend for the future. The corporate value of new energy enterprises plays a vital role in their growth. In 2023, greenhouse gas concentrations, along with global land and ocean temperatures, reached record highs [1]. Climate change has significantly exacerbated the incidence and intensity of energy poverty [2], with energy pollution, energy poverty, and climate change emerging as pressing issues demanding urgent solutions [3]. The evolution of the global energy system is a key factor in mitigating climate change and enhancing energy security [4], while the new energy industry plays a vital role in addressing energy crises and alleviating emission reduction pressures [5].

As new energy has become a key driver of high-quality economic growth and global sustainable development, corporate value—a critical measure of the industry’s progress—has gained heightened importance. Serving as the foundation for enterprise growth, it directly determines the potential and expansion capacity of new energy companies, thereby influencing the widespread adoption of clean energy and global greenhouse gas emissions. Existing literature on the corporate value of new energy firms has primarily focused on areas such as ESG ratings and financial performance [6], green finance and operational efficiency [7], as well as individual behavior and corporate outcomes [8].

Current environmental regulation policies are predominantly punitive, with relatively few incentive-based measures. However, many punitive environmental policies are ineffective and may even produce unintended adverse effects. On the one hand, such policies often lead to the pollution haven effect, where heavily polluting enterprises relocate from regions with stricter environmental regulations to those with weaker ones [9,10,11,12]. On the other hand, punitive environmental regulations require substantial enforcement costs. As early as 1995, Palmer et al. [13] pointed out that environmental oversight entails significant expenses for emission reduction and production technology control. Beyond these issues, punitive policies can also generate additional unintended consequences. For example, China’s Key Cities Policy for air pollution control and strict wastewater discharge standards significantly reduced labor demand [14,15]. Similarly, in the U.S., allowing oil refineries flexibility in meeting gasoline content standards failed to improve air quality while increasing consumer costs [16]. Canada’s air quality regulations led to a sharp decline in exports [17], and Mexico’s driving restrictions inadvertently increased the total number of vehicles in circulation—particularly older, more polluting models—worsening air quality [18]. In Europe, emission standards for the automotive market were undermined by corporate non-compliance and strategic manipulation of technical requirements, harming both consumers and manufacturers [19]. In contrast, incentive-based policies not only achieve environmental objectives [20,21] but also avoid these unintended negative economic impacts.

China, as the world’s largest producer and consumer of raw materials, is driving profound global transformations in resource efficiency, energy transition, and economic resilience through its policies and technologies [22]. The New Energy Demonstration City (NEDC) policy represents a typical government-led incentive-based environmental regulation [23]. By fostering regional industrial ecosystems, this policy stimulates value growth for new energy enterprises from both supply and demand sides. It significantly reduces early-stage operational risks while enhancing corporate reputation, offering a valuable reference for developing countries seeking to promote comprehensive enterprise development.

To increase urban clean energy adoption and promote resource-efficient development, China’s National Energy Administration designated 81 New Energy Demonstration Cities in 2014 [24], which initiative is a national policy launched by China’s National Energy Administration (NEA). Its primary goal is to promote the large-scale adoption of renewable energy in urban areas, reduce dependence on fossil fuels, and foster sustainable urban development through integrated planning and technological innovation. Encourage cities to utilize local renewable resources—such as solar, wind, geothermal, biomass, and waste-to-energy—to achieve a significant proportion of renewable energy in their total energy consumption.

This study investigates the impact of China’s New Energy Demonstration City (NEDC) policy on the value of new energy enterprises through a micro-level analysis, expanding the research perspective on incentive-based command-and-control environmental regulations. Using panel data from listed new energy companies (2010–2023) and applying a difference-in-differences (DID) model with the NEDC policy as a quasi-natural experiment, we find that the policy significantly enhances the value of new energy enterprises. This positive effect remains robust after undergoing multiple tests, including propensity score matching (PSM), parallel trend tests, placebo tests, alternative dependent variables, exclusion of other policy shocks, sample selection controls, and one-period lag analyses. The policy’s impact exhibits notable heterogeneity: enterprises in regions with stringent environmental regulations benefit significantly more than those in medium- or low-regulation areas; state-owned enterprises (SOEs) show stronger responses than non-SOEs; large firms outperform small ones; and medium capital-intensive enterprises are more positively affected than those with low or high capital intensity. Further mechanism analysis reveals that the NEDC policy effectively alleviates financing constraints for new energy firms while increasing their focus on green transformation, ultimately boosting the overall value of listed companies.

This study contributes to the existing literature in three key aspects. First, it enriches research on the impact of incentive-based policies on enterprises, an area that has received less attention compared to punitive environmental regulations. Punitive environmental policies often incur substantial regulatory costs. For instance, amendments to the U.S. Clean Air Act resulted in USD 810 million to USD 3.2 billion in lost market surplus for the cement industry [25]. These costs extend beyond economics—clean energy regulations have reduced political support for local government officials [26], while mandatory household electrification in the U.S. created high social net costs as families weighed compliance expenses against environmental benefits [27]. Moreover, punitive measures frequently trigger pollution haven effects. In water pollution control, high-polluting firms relocate upstream, spreading contamination more widely [28]. Air pollution regulations have led to strategic downwind siting (“pollute my neighbor” behavior) that harms downwind residents [29], while environmental supervision in Beijing-Tianjin-Hebei’s “2 + 26” cities caused pollution leakage to unregulated areas [30]. In contrast, Cao and Ma [20] found economic incentives for straw collection that are more effective than penalties in reducing agricultural burning pollution.

These findings suggest the potential superiority of incentive-based environmental policies. Previous studies have shown that environmental subsidies and tax incentives enhance corporate environmental performance [31], while proactive environmental responsibility strengthens market competitiveness [32]. Building on this literature, our study examines the NEDC policy’s ultimate economic consequence—enterprise value—testing whether incentive-based regulation can simultaneously improve environmental outcomes and economic performance. The results demonstrate a significant positive impact of the NEDC policy on new energy enterprise value, validating the effectiveness of incentive-based environmental regulation and providing a comprehensive micro-level foundation for policy evaluation.

As a representative incentive-based environmental policy, this study expands research on the NEDC policy. Internationally, studies emphasize that well-designed policies balancing technical feasibility and economic efficiency can enhance urban renewable energy self-sufficiency, reducing both energy costs and pollution [33,34]. Byrne et al. [35] reveal the interaction between policy tools and market mechanisms: demand-side policies (e.g., subsidies, tax incentives) boost short-term market activation, while supply side policies (e.g., carbon pricing, renewable quotas) foster long-term market vitality. Domestically, scholars find that NEDC policy improves urban energy efficiency through innovation, industrial, structural, service, and regulatory effects [36]. Wang et al. [37] use nighttime light data to demonstrate the policy’s spatial spillover effects in reducing local and adjacent regional carbon emissions. Ding et al. [23] focus specifically on NEDC’s environmental benefits. At the micro level, studies primarily examine corporate green innovation [38,39], though Lin and Xie [40] find that Xi’an’s subsidies reduced renewable firms’ total factor productivity. Pathway analyses indicate NEDC policies enhance urban green innovation via R&D, industrial innovation, and environmental performance [41]. Yang et al. [42] compare mechanisms, showing technological innovation has the strongest environmental impact, followed by resource allocation and industrial restructuring. Zhang et al. (2024) [43] confirm that urban innovation capacity, government support, and industrial upgrading indirectly boost green productivity. Green innovation is a key driver of China’s green growth [44], and NEDC policy elevate carbon efficiency by optimizing industrial structures, stimulating green technology, and reducing energy intensity [45].

Distinct from macro-level studies, this research adopts a micro perspective by innovatively examining corporate market value and financial indicators. Using Tobin’s Q, we analyze the policy’s impact on new energy enterprise value and pioneer the investigation of three firm-level mechanisms: financing constraint alleviation, green innovation incentives, and green transformation. This systematically reveals the internal channels through which the NEDC policies influences enterprise value. Although the policy demands higher capabilities, innovation, and structural adjustments from firms with associated costs, it ultimately stimulates long-term innovation among new energy enterprises—a finding consistent with the Porter Hypothesis. Our results demonstrate that the policy significantly enhances new energy enterprise value, validating its effectiveness and underscoring its potential to drive China’s comprehensive green economic transition.

Second, we fulfill a series of literature regarding pro-environment initiatives that can enhance firm value or investor appeal. Flammer (2021) [46] demonstrates that investors exhibit a positive market reaction to firms undertaking credible environmental initiatives, such as green bond issuances. Similarly, Cheng, Kim, and Ryu (2024) [47] provide evidence that firms with superior environmental, social, and governance (ESG) performance achieve higher market valuations in the Chinese context; Government green procurement serves both direct incentive and supervisory roles, while also generating indirect signaling effects that motivate firms to proactively engage in green investments and attract specific investors, thereby enhancing corporate ESG performance [48]. Green credit policies guide enterprises to improve ESG performance, which in turn reduces the weighted average cost of capital and significantly enhances financial performance [49]. Hao et al. (2024) [50] find that China’s Direct Reporting for Environmental Statistics (DRES) policy significantly improves corporate ESG performance. Similarly, the Environmental Protection Tax (EPT) policy and government green advocacy have spurred the emergence of green investors as a distinct institutional investor type, which further strengthens corporate ESG performance [51]. These findings are consistent with the theoretical proposition that markets reward environmental commitment and transparency. By explicitly engaging with this body of literature, the authors can more effectively situate their research within contemporary financial discourse. For example, they could contend that their findings offer causal support for the view that state-incentivized sustainability programs contribute to shareholder value, thereby extending prior research that established predominantly correlational relationships.

This study complements existing research on corporate value. Since Friedman’s (1970) seminal work established that profit maximization within legal boundaries constitutes corporate value, research has expanded to diverse value determinants [52]. Recent studies show ESG ratings, environmental disclosures, and regulatory factors significantly influence firm value. Specifically, ESG disclosure—particularly environmental components—enhances corporate valuation, with high-ESG firms commanding market premiums due to improved risk management and stakeholder confidence [47,53]. Corporate governance improvements, such as increased major shareholder ownership, boost management efficiency and thereby firm value [54], while media oversight curbs executive self-interest, creating shareholder value [55]. Firms also build value through social responsibility initiatives and digital transformation [56,57]. However, the impacts are not uniformly positive. Policy uncertainty typically reduces market valuation [58], excessive ESG disclosure can diminish value creation [59], while knowledge spillovers [60], climate risks [61], and CEO tenure under varying conditions [62] all demonstrate complex valuation effects.

From the perspective of the Porter Hypothesis, Wang et al. (2024) [63] demonstrate that carbon market efficiency enhances green technology innovation by increasing corporate R&D intensity and government subsidies, confirming the existence of the Porter Effect in carbon markets. Wang et al. (2023) [64] analyze carbon trading policies and argue that implementation not only directly reduces CO₂ emissions through production cuts but also triggers the Porter Hypothesis by “forcing” firms to pursue green innovation, thereby curbing disorderly capacity expansion. Liu et al. (2025) [65] find that carbon trading promotes corporate green transformation by improving ESG performance, further validating the Porter Hypothesis. Cui et al. (2022) [66] test the “weak” version of the Porter Hypothesis, empirically showing that China’s Clean Production Audit (CPA) program stimulates green innovation in firms. Above all, the Porter Hypothesis and the concept that “carrot” policies can drive efficiency and innovation to enhance profitability together provide a theoretical basis for our findings [13].

This study demonstrates that the New Energy Demonstration City policy—an incentive-based environmental regulation—effectively enhances enterprise value, revealing how centrally driven administrative policies can create tangible economic benefits for firms. Specifically, the new energy enterprise incentives increase firm value by alleviating financing constraints, strengthening green innovation incentives, and accelerating sustainable transformation—collectively boosting listed companies’ market valuation.

Third, this study enriches both principal-agent theory and information asymmetry theory. Under the principal-agent framework, local governments’ implementation of central directives may yield unintended consequences—solving one problem while creating another. For instance, performance evaluations targeting SO₂ emission reductions have led local officials to prioritize pollution control at the expense of economic growth [67]. Environmental decentralization has similarly resulted in pollution spillovers during local enforcement [68]. China’s unique context of environmental federalism further exacerbates this tension, as central environmental priorities often conflict with local governments’ economic development incentives [69]. From the perspective of information asymmetry theory, deregulation in the U.S. electricity market led to significant price volatility in the opaque coal-fired power sector [70]. China’s “one permit” environmental policy reform primarily affected smaller firms that had fewer interactions with regulators and possessed less information [71]. When it comes to government-monitored pollutants, state-owned enterprises (SOEs) demonstrate better environmental performance than their private counterparts [72]. Benefiting from preferential treatment, SOEs’ exports remained largely unaffected by SO₂ regulations [73].

This study demonstrates that incentive-based environmental policies can enhance the effectiveness of principal-agent relationships. Our findings reveal that state-owned enterprises (SOEs) outperform their non-state counterparts in policy implementation, while large firms achieve better results than smaller ones. Both SOEs and large corporations not only comply more effectively with central government policies but also enjoy greater access to information about various incentive programs.

The structure of this paper is organized as follows: Section 2 develops the research hypotheses based on theoretical analysis. Section 3 describes the dataset and research methodology. Section 4 presents the main empirical results, including robustness checks and heterogeneity analysis. Section 5 examines the underlying mechanisms; and Section 6 discusses policy implications.

2. Theoretical Analysis and Hypotheses

2.1. Principal-Agent Theory

Berle and Means (1932) proposed the proposition of “separation of ownership and control” in The Modern Corporation and Private Property, advocating for the division of ownership and management rights, which laid the foundation for the development of principal-agent theory [74]. Then Jensen and Meckling (1976) formally introduced the principal-agent theory for the first time in their paper “Theory of the firm: Managerial behavior, Agency cost and Ownership structure” [75]. The core of this theory lies in analyzing the conflicts of interest and information asymmetry between principals and agents, exploring how institutional designs (e.g., contracts, incentives, monitoring, policies) can reduce agency costs in order to achieve alignment of interests between both parties. It provides a crucial framework for understanding power distribution and decision-making mechanisms in modern organizations, while also advancing the development of corporate governance practices.

As a crucial instrument for local governments to drive green transition, the new energy demonstration city policy, inherently constitutes a complex principal-agent relationship. The central government act as the “principal,” establishing macro-level emission reduction targets and industrial upgrading objectives; local governments serve as both “sub-principals” and policy implementers, tasked with translating abstract goals into localized practices; while new energy enterprises ultimately function as “agents,” responsible for executing technological R&D, project deployment, and operations. The core tension in this chain stems from divergent objective functions: higher-level governments prioritize global environmental benefits and long-term energy security, local governments focus more on short-term economic growth, fiscal revenue, and policy evaluation compliance, while profit maximization remains the fundamental objective of enterprises. Such goal misalignment often leads to “selective incentive” distortions in policy implementation—local governments, aiming to demonstrate quick political achievements, so they may preferentially support projects with high maturity and short-term payoffs rather than cutting-edge technologies requiring long-term investment, then it results in resource misallocation and suppressed corporate innovation momentum. Environmental governance challenges within multi-tier government systems urgently require attention [76]. The principal-agent relationship poses significant challenges in effectively compelling enterprises to deliver substantive ecological improvements, while also complicating the accurate assessment of polluting firms’ environmental and social responsibility performance. He and Wang (2020) argue that due to the spatial discontinuity in the implementation of environmental regulation policies, well-designed central incentive mechanisms may become distorted and lead to unintended consequences [77]. Therefore, it is crucial to examine the effectiveness of environmental regulation policies from the perspective of principal-agent relationships. Hence, we put forward our main hypothesis.

Hypothesis 1: 

The implementation of the New Energy Demonstration City policy can enhance the value of new energy enterprises.

2.2. Information Asymmetry Theory

The theory of information asymmetry emerged in the 1970s, proposed by American economists Akerlof, Spence, and Rotschild [78,79,80]. It posits that in market transactions there exists an imbalance of information between buyers and sellers. The party with superior information typically holds a strategic advantage, while the other not, ultimately impairing overall market functionality. In such financial markets, disparities in information often exist between enterprises and external investors, as well as between enterprises and financial institutions. Consequently, adverse selection and moral hazard problems tend to emerge, prompting investors and financial institutions to adopt cautious attitudes toward corporate investment and lending. This may lead to increased external financing costs and heightened financing constraints for enterprises, ultimately restraining their investment activities and growth opportunities.

Most new energy enterprises are technology-intensive and innovation-driven, with their development heavily reliant on technological R&D and innovation activities. However, this inherently accompanied by significant uncertainty, making accurate predictions of future cash flows particularly challenging for enterprises. Moreover, the asset structure of these enterprises predominantly consists of intangible assets, which creates significant obstacles when seeking loans from financial institutions, thereby exacerbating their financing constraints. Meanwhile, lenders tend to favor entrepreneurs with more observable signals due to information asymmetry [81]. Consequently, in capital markets, investors have doubts regarding the stability of returns from new energy enterprises, compelling these firms to incur higher costs in equity financing to attract investment. This bilateral information asymmetry directly constrains corporate funding sources through actual financing constraints, making it difficult for enterprises to meet the capital requirements for large-scale technological R&D and business expansion. Consequently, numerous valuable investment projects may be suspended or delayed, severely impeding both the growth pace and development scale of enterprises. The implementation of new energy demonstration city policies provides an effective mechanism to alleviate financing constraints for new energy enterprises. The establishment of demonstration cities and zones enables new energy enterprises within these areas to gain market visibility and investor recognition, thereby mitigating information asymmetry between firms and investors. This alleviation of financing constraints consequently fosters green innovation outputs [82]. Moreover, as the policy directly provides resources from the government to enterprises, the government’s credit endorsement can directly and indirectly alleviate corporate financing constraints. However, this effect varies depending on the ownership structure and scale of the enterprises [83]. Dang et al. (2022) argue that financing constraints constitute a critical factor in how environmental regulations influence corporate investment behavior, so policy discussions should take this factor into account [84]. Therefore, clarifying the formulation process of corporate environmental policies and their relationship with enterprise financing will facilitate the advancement of environmental protection efforts [85]. In all, green finance policies can effectively alleviate financing constraints for green innovation [86], broaden corporate financing channels and reduce funding costs, help enterprises overcome financing bottlenecks until enhance their market value. Hence, we put forward the following hypothesis.

Hypothesis 2: 

The New Energy Demonstration City policy can alleviate financing constraints for new energy enterprises, thereby influencing their corporate value.

2.3. Porter Hypothesis

Michael Porter (1991) argues that well-designed environmental regulations can stimulate firms’ innovation incentives. In the long run, it enables enterprises to benefit from innovation—not only offsetting compliance costs but also achieving cost reductions and competitiveness gains through technological advancements, thereby realizing a win-win outcome of improved environmental quality and enhanced corporate economic performance [87].

As a significant environmental and industrial policy, the New Energy Demonstration City policy significantly promotes corporate green innovation and green transformation. First, the policy incentivizes enterprises to increase investment in R&D, which may raise costs. However, Wu and Lin (2022) contend that while environmental regulations serve as an effective pollution control tool, they can stimulate corporate innovation in the long term, thereby offsetting cost increases and enhancing competitiveness. Secondly, this policy drives corporate green transformation, reflected in its compensation effect on firms’ technological innovation capabilities. As profit-maximizing entities, enterprises facing environmental regulations, they leverage technological innovation to optimize production processes, thereby enhancing operational efficiency [88]. Cui et al. (2022) further found that the policy’s regulatory impact proves particularly significant in promoting radical green innovations such as environmental invention patents [66]. Furthermore, green innovation and green transformation can enhance a company’s social image and brand value. The increase in both the quantity and quality of corporate green innovation can attract on-site research by institutional investors and greater media attention, among other channels [89]. Under the policy, enterprises can establish a positive corporate image, strengthen consumer trust, and gain higher market share and brand premium, thereby significantly improving profitability and social value. Hence, we put forward the following hypothesis.

Hypothesis 3: 

The New Energy Demonstration City Policy exerts influence on the value of new energy enterprises through mechanisms of green innovation and green transformation.

2.4. The Pollution Haven Hypothesis

Copeland (1994) proposed the Pollution Haven Hypothesis, which posits that in the context of global economic integration, due to disparities in environmental regulation stringency across nations or regions, pollution-intensive industries tend to relocate production activities to countries or regions with relatively lower environmental standards and more lenient regulations. These areas effectively function as “havens” for pollution, reflecting firms’ propensity to locate production in jurisdictions with weaker environmental standards [90].

Although the new energy industry is generally environmentally friendly, certain production processes still carry pollution risks. When implementing the New Energy Demonstration City policy, variations in environmental regulation stringency may partially influence policy effectiveness. Researchers found that in regions with high environmental regulation stringency, intelligent transformation has a significant incentive effect on green innovation, whereas in areas with low regulatory intensity, this impact is not statistically significant [89]. Pollution-intensive firms tend to locate in countries or regions with relatively lower environmental standards [91]. This occurs because areas with weaker environmental regulations attract enterprises through lower land costs and more lenient initial environmental requirements, allowing firms to leverage local resource advantages to reduce costs and enhance efficiency through industrial chain innovation. In contrast, regions with stringent environmental regulations impose stricter standards and entry barriers, while these may require substantial initial environmental investments from firms, in the long term, it compel innovation, spur green technologies, and not only ensure compliance but also open new markets and strengthen competitiveness, thereby generating lasting benefits for enterprises. Generally, new energy enterprises belong to the category of environmentally capable firms, and such firms are more likely to locate in regions with stricter regulations [92]. This demonstrates that the intensity of environmental regulations significantly impacts corporate costs, thereby influencing both enterprise and industrial location choices [93], which ultimately leads to differential effects on firm value. Hence, we put forward the following hypothesis.

Hypothesis 4: 

The impact of the New Energy Demonstration City policy on the value of new energy enterprises exhibits heterogeneity across regions with varying environmental regulation intensities.

2.5. Property Rights Theory and Economies of Scale Theory

Coase laid the foundation for modern property rights theory through his seminal papers “The Nature of the Firm” (1937) and “The Problem of Social Cost” (1960) [94,95]. This theory primarily examines the relationship between property rights, incentives, and economic behavior, emphasizing how the nature of property rights affects firms’ resource acquisition and decision-making processes [96,97].

SOEs typically possess well-established governance structures and substantial financial capabilities, granting them comparative advantages in accessing government support. Under policy incentives, SOEs tend to increase the quantity of green innovations [98]. They leverage existing resource and scale advantages to rapidly expand market share in new energy sectors, and environmental regulations exhibit more pronounced positive effects in promoting green transformation among SOEs [99]. However, due to their large organizational scale and complex internal systems, SOEs are prone to path dependency and often lack endogenous drivers for continuous transformation. When confronting the rapidly evolving market dynamics and technological innovation demands of the new energy industry, their long decision-making processes may result in delayed responses to market changes. In contrast, non-state-owned enterprises typically possess more flexible operational mechanisms and sharper market acumen. With policy support, they can swiftly adjust business strategies and increase investments in R&D and market expansion to gain competitive advantages. However, due to their relatively smaller scale and lower credit ratings, these enterprises may face challenges in accessing policy resources and financial support.

Marshall (1890) formally introduced the concept of “economies of scale” [100]. This theory posits that, within a given period, as a firm expands its production scale and increases output, the unit production cost will progressively decline, thereby enhancing economic efficiency.

Large enterprises can effectively leverage policy incentives by capitalizing on their scale advantages, gaining easier access to government subsidies and preferential loans for expanding production capacity, upgrading technological equipment, or developing new energy projects. Furthermore, they generally possess stronger capabilities in R&D and innovation. Research indicates that China’s green finance reforms have significantly enhanced corporate patent quality, with this effect being particularly pronounced among large enterprises [101]. Small and medium-sized enterprises (SMEs), while potentially constrained by capital and technological limitations, exhibit higher operational flexibility and market responsiveness. Research by Zwick and Mahon (2017) found that tax incentive policies demonstrate more pronounced effects for small businesses [102]. Similarly, Liu and Mao (2019) revealed that tax reforms exert greater impact on firms facing high financing constraints, indicating that reduced capital costs and improved cash flows generated by the reforms disproportionately benefit these financially constrained enterprises [103]. The NEDC policy creates development opportunities for SMEs, which enables them to access financing and technological innovation resources, thereby enhancing competitiveness. Meanwhile, they can rapidly adjust business strategies to focus on specialized sectors, developing differentiated new energy products and services. However, they may also face greater risks and uncertainties in responding to policy changes due to relatively limited resources and weaker risk-bearing capacity. Therefore, governments need to fully account for firm-size heterogeneity when formulating and implementing the environmental policy. Hence, we put forward the following hypothesis.

Hypothesis 5: 

The implementation of the New Energy Demonstration City policy generates heterogeneous impacts on the value of new energy enterprises across different ownership types and firm sizes.

3. Research Design

3.1. Data

The data used in this paper is mainly the matched data, including (1) the official Notice on the First Batch of New Energy Demonstration Cities (Industrial Parks) published by China’s National Energy Administration in 2014, which designated 81 cities and 8 industrial parks as pilot areas; (2) the city-level data obtained from the China City Statistical Yearbook; and (3) detailed firm-level financial data from the CSMAR database. To comprehensively evaluate the policy effects before and after the implementation of the NEDC initiative, this study analyzes 2010–2023 data from A-share new energy firms to evaluate the policy’s long-term effects, applying rigorous sample selection criteria to ensure data quality.

The data processing process is as follows: First, to ensure financial stability, we exclude ST/*ST firms and companies with debt-to-asset ratios exceeding 100%. Second, observations with missing critical data are excluded to ensure statistical validity, we also winsorize all continuous variables at the 1st and 99th percentiles, and apply logarithmic transformations to selected variables, ultimately yielding 1091 valid observations.

3.2. Model Specification

This study treats the NEDC pilot policy as a quasi-natural experiment and employs a difference-in-differences (DID) approach to examine the policy effects. Controlling for various potential factors influencing new energy enterprise value, the baseline regression model is specified as follows.

$${TobinQ}_{i,t}=\alpha +\beta {Treat}_i\times {Post}_t+{\gamma Controls}_{i,t}+{\theta }_i+{\lambda }_t+{\epsilon }_{i,t},$$

(1)

where ${TobinQ}_{i,t}$ is the New Energy Enterprise Value of firm i at year t; ${Treat}_i\times {Post}_t$ represents the core explanatory variable which refers to the policy dummy variable, The interaction term between the NEDC dummy variable (Treat) and the policy implementation time dummy variable (Post); ${Controls}_{i,t}$ is a set of control variables at the firm level and the city level; ${\theta }_i$,${\lambda }_t$ represent firm and year fixed effects, respectively. ${\epsilon }_{i,t}$ refers to the error term cluster. The coefficient $\beta$ is the primary focus of this study, will demonstrate a positive impact of the NEDC policy on enterprise value if it is statistically significant and positive. Following Mugerman et al. (2022) [104], all empirical results are clustered at the firm level. Subsequent robustness checks with alternative clustering approaches yield consistent findings, confirming the reliability of our conclusions.

3.3. Variables Description

(1)

Dependent Variable: New Energy Enterprise Value (TobinQ)

Following the Nanda et al. and Cremers and Ferrell method [105,106], we in this paper estimate the new energy enterprise value by Tobin’s Q, which uses year-end corporate valuation to reflect the value and achievements accumulated over the period, indicating the driving force behind enterprise value growth. The calculation formula is (1) (Variable data comes from the China’ A-listed company database).

$$Tobin\prime s Q=(Market Value of Outstanding Shares+Non-tradable Shares+Net Debt) / Total Assets at Period End$$

(2)

(2)

Explanatory Variable: New Energy Demonstration City Policy (Treat × Post)

The key treatment variable (Treat × Post) is constructed by interacting a dummy for firms located in New Energy Demonstration Cities (Treat = 1) with a post-policy period indicator (Post = 1 for 2014 onward). This DID specification compares outcomes between treatment and control groups before and after policy implementation while controlling for city and year fixed effects, with robustness checks verifying parallel pre-trends.

(3)

Control Variables

To more comprehensively analyze the impact of new energy demonstration city construction on the value of new energy enterprises, we controlled for a series of firm-level and city-level variables. Firm-level controls include company size (Size) measured as the natural logarithm of total assets, leverage ratio (Lev) calculated as year-end total liabilities divided by total assets, return on assets (ROA) computed as net profit divided by average total assets, growth capability (Growth) measured as current year operating revenue divided by previous year revenue minus one, listing age (ListAge) calculated as current year minus IPO year, asset turnover ratio (ATO) measured as operating revenue divided by average total assets, and cash holdings (Cashflow) computed as cash and equivalents divided by total assets. City-level controls include industrial structure (GDPct3) measured as tertiary industry value-added divided by GDP and registered population (POP).

(4)

Mechanism Variables

Following the Whited and Wu (2006), Hao et al. (2024), and Loughran and McDonald (2011) [107,108,109], this paper examines financing constraints (WW), green innovation (EnvrPat), and green transformation (Green) as mechanism variables to elucidate the underlying channels through which new energy policies affect enterprise value. The financing constraints (WW) are measured using the WW index, which assesses the degree of financing limitations based on a firm’s capital structure decisions between debt and equity. Green innovation (EnvrPat) is quantified as the natural logarithm of one plus the sum of green invention patent applications and green utility model patent applications. For subsequent mechanism analysis, green innovation is further categorized into green invention innovation (EnvrInvPat), measured as the natural logarithm of one plus green invention patent applications, and green utility innovation (EnvrUtyPat), measured as the natural logarithm of one plus green utility model patent applications. The green transformation (Green) is evaluated through textual analysis of keywords related to green practices in corporate annual reports, reflecting the extent of a firm’s green transition. Detailed variable definitions are provided in

Table 1. Description of variables (The units of Monetary Amount is CNY).

Variable	Definition (unit)	Mean	Std. Dev	Time Span
	Panel A: Dependent variable
TobinQ	Tobin’s Q = (Market Value of Outstanding Shares + Market Value of Non-tradable Shares + Market Value of Net Debt)/Total Assets (%)	1.3975	0.528	2010–2023
	Panel B: Independent variable
Treat × Post	If a city i applies to be a demonstration city at time t, it is 1; otherwise 0	0.2159	0.412	2010–2023
	Panel C: Controls
Size	Natural logarithm of total assets (CNY)	23.4060	1.312	2010–2023
Lev	Year-end total liabilities/Year-end total assets (%)	0.5724	0.165	2010–2023
ROA	Net profit/Average total assets (%)	0.0295	0.052	2010–2023
Cashflow	Cash and equivalents/Total assets (%)	0.0484	0.064	2010–2023
Growth	(Current year operating revenue/Previous year revenue)-1 (%)	0.1820	0.845	2010–2023
ListAge	Current year-IPO year (year)	2.7021	0.667	2010–2023
ATO	Operating revenue/Average total assets (%)	0.6225	0.438	2010–2023
GDPct3	Tertiary industry value-added/GDP (%)	0.5427	0.131	2010–2023
POP	Registered population (in 10,000 s) (people)	875.011	670.67	2010–2023
	Panel D: Mechanism Variables
WW	Whited-Wu index	−1.0615	0.080	2010–2023
EnvrPat	ln(1 + Green invention patents + Green utility patents)	0.7284	1.292	2010–2023
Green	ln(1 + Frequency of 113 green transformation keywords in annual reports)	3.2593	0.889	2010–2023

.

Panel A shows our dependent variable TobinQ, which exhibits a mean of 1.398 (SD = 0.528). This distribution indicates moderate dispersion. Panel B presents the independent variable, Treat × Post, shows that 21.6% of the sample firms were subject to the policy during the study period. Controls introduced in Panel C, The log-transformed firm size (Size) exhibits a mean of 23.406 (SD = 1.312), indicating moderate variation in scale across firms. The leverage ratio (Lev) averages 0.572, suggesting moderate debt levels. With a mean return on assets (ROA) of 0.030, the sample demonstrates relatively weak profitability. Growth capability (Growth) shows a mean of 0.182 but extreme values ranging from −0.86 to 22.10, reflecting pronounced heterogeneity. The logged listing age (ListAge) averages 2.702, indicating generally recent IPOs. Asset turnover ratio (ATO) has a mean of 0.623 with substantial efficiency dispersion. Industrial structure (GDPct3) averages 0.543 across cities, while registered population (POP) ranges dramatically from 104.74 to 3416. In Panel D of mechanism variables, the average financing constraint (WW) of −1.062 demonstrates widespread financing difficulties among sample firms. Green innovation (EnvrPat), measured by the logarithmic count of green patent applications, exhibits a modest mean of 0.728, indicating relatively limited green innovation outputs. The green transformation index (Green) achieves a higher mean of 3.259, reflecting substantial textual evidence of sustainability focus on corporate disclosures.

3.4. Descriptive Statistics

After processing the data, we obtained a balanced panel of 1283 firm-year observations, comprising 277 treatment group firms located in New Energy Demonstration Cities and 1006 control group firms from non-pilot cities.

Table 2. Descriptive statistics.

	All Firms			Treat Firms			Control Firms
	Mean	Std. Dev	N	Mean	Std. Dev	N	Mean	Std. Dev	N
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Panel A: Dependent variable
TobinQ	1.3975	0.528	1283	1.4589	0.589	277	1.3806	0.509	1006
	Panel B: Independent variable
Treat × Post	0.2159	0.412	1283	1.0000	0.000	277	0.0000	0.000	1006
	Panel C: Controls
Size	23.4060	1.312	1283	23.5988	1.199	277	23.3529	1.337	1006
Lev	0.5724	0.165	1283	0.5383	0.180	277	0.5818	0.159	1006
ROA	0.0295	0.052	1283	0.0274	0.049	277	0.0301	0.053	1006
Cashflow	0.0484	0.064	1283	0.0513	0.064	277	0.0476	0.065	1006
Growth	0.1820	0.845	1282	0.1295	0.291	276	0.1964	0.941	1006
ListAge	2.7021	0.667	1283	2.8712	0.591	277	2.6555	0.680	1006
ATO	0.6225	0.438	1283	0.6072	0.463	277	0.6267	0.431	1006
GDPct3	0.5427	0.131	1192	0.6060	0.145	277	0.5235	0.121	915
POP	875.011	670.67	1092	793.7636	377.812	242	898.142	731.494	850
	Panel D: Mechanism Variables
WW	−1.0615	0.080	1256	−1.0726	0.068	266	−1.0585	0.082	990
EnvrPat	0.7284	1.292	1090	0.8513	1.452	242	0.6934	1.241	848
Green	3.2593	0.889	1146	3.8642	0.679	238	3.1007	0.869	908

presents comprehensive summary statistics for all variables, with the sample strategically limited to new energy enterprises to ensure comparability given the city-level policy implementation. The results demonstrate remarkably similar variable means between groups across most indicators: the dependent variable TobinQ shows comparable values (1.4589 vs. 1.3806), key controls like Size (23.5988 vs. 23.3529) and ROA (0.0274 vs. 0.0301) exhibit minimal differences, and even mechanism variables such as the WW index (−1.0726 vs. −1.0585) remain closely aligned. This statistical balance confirms that our empirical strategy successfully addresses selection bias concerns, enabling us to attribute observed effects to the policy intervention rather than pre-existing differences between groups, thereby strengthening the validity of our difference-in-differences approach for causal identification.

4. Empirical Analysis

4.1. Baseline Regression

In the baseline regression, we test the effect of the reform on the value of new energy enterprises; the report is shown in

Table 3. Baseline regressions.

	Dependent Variable: TobinQ
	(1)	(2)	(3)	(4)
Treat × Post	0.078 **	0.130 ***	0.131 ***	0.085 **
	(2.017)	(3.418)	(3.345)	(2.313)
Size		−0.233 ***	−0.228 ***	−0.228 ***
		(−16.278)	(−14.675)	(−15.272)
Lev		−0.150	−0.050	−0.105
		(−1.419)	(−0.448)	(−1.010)
ROA		1.015 **	1.288 ***	1.041 ***
		(2.322)	(3.217)	(2.660)
Cashflow		0.581 **	0.854 ***	0.841 ***
		(2.567)	(3.522)	(3.657)
Growth		0.021	0.035	0.003
		(0.677)	(1.030)	(0.095)
ListAge		0.021	0.022	0.011
		(0.951)	(0.958)	(0.471)
ATO		0.142 ***	0.015	0.029
		(4.615)	(0.331)	(0.635)
GDPct3		0.250 *	0.487 ***	0.464 ***
		(1.866)	(3.496)	(3.441)
POP		0.000 **	0.000 **	0.000 **
		(2.483)	(2.174)	(2.154)
YearFE	No	No	No	Yes
IdFE	No	No	Yes	Yes
R2	0.004	0.317	0.439	0.528
Observations	1282	1091	1091	1091

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

. Column (1) presents regression results without control variables, this result means that after the reform, firms’ average TobinQ decreases by almost 7.8% at 1% significance level, indicating a positive policy effect. In Column (2) with controls but no fixed effects, it leads to 13% higher market valuation than non-policy counterparts. Column (3) incorporates control variables and firm fixed effects, with the policy remaining significant at 0.131 at a 1% significance level which confirms the policy’s positive effect after accounting for firm heterogeneity. Column (4) further adds year fixed effects, yielding a still-significant coefficient of 0.085. These robust results demonstrate the policy’s stable and universal value-enhancing effects across both terms.

The control variables reveal three key patterns: First, firm size (Size) shows significantly negative coefficients (−0.233 in Column (2) and −0.228 in Column (4), both at 1% level, suggesting a market valuation discount for larger firms, so we will further explore in heterogeneity analysis. Second, profitability (ROA) consistently enhances firm value, reflecting market recognition of financial performance. Third, other controls exhibit varying significance across specifications, collectively supporting the robustness of our policy effect analysis.

4.2. Robustness Tests

4.2.1. PSM-DID

In practice, the parallel trends assumption between treatment and control groups may be violated by unobserved confounders, potentially introducing selection bias. PSM addresses this by matching treated and control units on observable covariates, thereby improving estimation robustness. This study applies PSM with: (1) New energy firms in cities implementing the policy since 2014 as treatment group; (2) New energy firms in non-policy cities from 2010 to 2023 as control group; (3) All control variables from the baseline regression as covariates.

Figure 1 presents the nearest-neighbor matching results for treated and control groups. The post-matching differences in control variables show substantial reduction compared to pre-matching, confirming the appropriateness of our covariate selection and matching approach, which supports the reliability of baseline regression estimates.

Then we use the matched sample and re-estimate the model which shows in

Table 4. Robustness test: PSM-DID results.

	Dependent Variable: TobinQ
	Nearest Neighbor Matching	Kernel Matching
	(1)	(2)
Treat × Post	−0.251 ***	−0.031 **
	(0.039)	(0.014)
Controls	Yes	Yes
Year FE	Yes	Yes
Id FE	Yes	Yes
R2	0.554	0.546
Observations	638	987

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. ** and *** indicate statistical significance at the 5% and 1% level, respectively.

. The Treat × Post coefficients remain positive and statistically significant at the 1% level, consistent with baseline results. This confirms the policy’s robust positive effect on enterprise value and rules out self-selection bias.

4.2.2. Parallel Trend Test

To increase the credibility of our findings, we rigorously assessed the parallel trend assumption during the pre-treatment period. Referring to Beck et al. (2010) and Nunn and Qian (2011), we use the event study approach and regress our dependent variable on a set of year dummies that represent the reform years [110,111]. The model is as follows:

$$\begin{array}{c}{Y}_{it}={\alpha }_i+{\beta }_{-3}{D}_{it}^{-3}\times {treat}_i+{\beta }_{-2}{D}_{it}^{-2}\times {treat}_i+{\beta }_0{D}_{it}^0\times {treat}_i+{\beta }_1{D}_{it}^1\times {treat}_i,\\ +{\beta }_2{D}_{it}^2\times {treat}_i+{\beta }_3{D}_{it}^3\times {treat}_i+{\gamma Controls}_{i,t}+{\theta }_i+{\lambda }_t+{\epsilon }_{i,t},\end{array}$$

(3)

where $D_{it}^{-j}$ is a dummy variable indicating the j-th year before the reform, while Dj indicates the j-th year after the reform. We exclude the year before the reform year, D⁻¹, as the reference group, to avoid multicollinearity. The estimated β vectors represent the relationship between $Y_{it}$ and the year dummies. Specifically, we expect ${\beta }_{-3}$ and ${\beta }_{-2}$ to be constant over time for all years before the implementation of the reform, indicating that there are no significant differences between the treatment and control groups before the reform. In contrast, ${\beta }_0, {\beta }_1$, ${\beta }_2,$ and ${\beta }_3$ are expected to be significant and positive, indicating that after the reform, the TobinQ of the treatment group increases. The results of Model (3) are shown in Figure 2.

The pre-policy estimates consistently approximate zero with confidence intervals crossing the null line, confirming the parallel trends assumption by demonstrating no systematic differences between treatment and control groups prior to policy enactment. Following implementation, the treatment effect exhibits a significant immediate increase, maintaining elevated levels in subsequent periods before gradually moderating, Therefore, the parallel trend test passed.

In summary, the implementation of the New Energy Demonstration City policy has exerted a statistically significant positive impact on the value of new energy enterprises, thereby confirming Hypothesis 1.

4.2.3. Placebo Test

To ensure robustness, we conduct placebo tests to rule out pre-existing trends or external shocks. Figure 3 plots the distribution of estimators under falsified policy timings, showing a near-normal distribution centered at zero and no systematic deviation from the null. This confirms our baseline estimates reflect true policy effects rather than model misspecification or sampling bias.

4.2.4. IPW-DID

Furthermore, we use Inverse Probability Weighting Difference-in-Differences (IPW-DID) to prove the robustness of our results. IPW-DID is a causal inference method that mainly solves the selection bias of treatment group and control group, which combines propensity score weighting with the DID framework. It involves two steps: first, estimating treatment probabilities (propensity scores) using pre-treatment control variables (e.g., city GDP, industrial structure, size, ROA) and calculating inverse probability weights to balance covariates between treatment and control groups; second, applying these weights to a standard DID model for weighted estimation. IPW-DID directly mitigates selection bias from observables through weighting and offers double robustness (yielding consistent estimates if either the propensity score or outcome model is correctly specified).

We present the density distribution of propensity scores generated during the Inverse Probability Weighting (IPW) process (Figure 4), comparing the treatment and control groups. The overlapping density curves indicate that the propensity score weighting effectively balances the observed covariates between the two groups, as both distributions cover a similar range of propensity scores with comparable density peaks. This alignment demonstrates that the IPW method successfully mitigates selection bias by creating a comparable counterfactual group, satisfying the balancing condition required for robust causal inference. We re-ran the baseline regression using the inverse probability weighted sample and the results remain robust (

Table 5. IPW-DID.

	Dependent Variable: TobinQ
	(1)	(2)	(3)	(4)
Treat × Post	0.113 **	0.160 ***	0.159 ***	0.089 **
	(2.494)	(3.883)	(3.724)	(2.225)
Size		−0.225 ***	−0.223 ***	−0.224 ***
		(−14.104)	(−12.801)	(−13.058)
Lev		−0.009	0.081	0.000
		(−0.069)	(0.609)	(0.000)
ROA		1.502 ***	1.692 ***	1.328 ***
		(2.773)	(3.530)	(2.814)
Cashflow		0.561 **	0.733 ***	0.776 ***
		(2.510)	(2.975)	(3.296)
Growth		−0.009	−0.003	−0.023
		(−0.248)	(−0.067)	(−0.898)
ListAge		0.057 ***	0.053 **	0.046 **
		(2.854)	(2.540)	(2.298)
ATO		0.151 ***	0.030	0.039
		(4.160)	(0.557)	(0.711)
GDPct3		0.299 **	0.488 ***	0.413 ***
		(1.964)	(3.174)	(2.731)
POP		0.000	0.000	0.000
		(0.530)	(1.239)	(1.126)
YearFE	No	No	No	Yes
IdFE	No	No	Yes	Yes
R2	0.007	0.310	0.431	0.530
Observations	1282	1091	1091	1091

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. ** and *** indicate statistical significance at the 5% and 1% level, respectively.

).

4.2.5. Alternative Dependent Variables

To enhance robustness, we replace the dependent variable (Tobin’s Q) with annual stock returns (Ret). As shown in Column (1) of

Table 6. Robustness Tests.

	Stock Returns	Policy Shock Controls	High-Dim FE	Selection Bias Controls	Lagged Treatment
	(1)	(2)	(3)	(4)	(5)
Treat × Post	0.066 *	0.084 **	0.104 *	0.114 ***
	(1.920)	(2.281)	(1.921)	(2.603)
L.Treat × Post					0.077 *
					(1.755)
LowCarbon		0.026
		(0.833)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes
Pro*Id FE			Yes
Year*Id FE			Yes
Pro*Year FE			Yes
R2	0.407	0.529	0.826	0.542	0.516
Observations	1028	1091	873	894	857

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

, the policy increases Ret by 6.6% and it is significant at the 10% level, corroborating our baseline findings and confirming the policy’s value-enhancing effect across alternative metrics.

4.2.6. Controlling for Concurrent Policy Shocks

In this research, we consider the other related policy: Low-Carbon City Pilot Policy (denoted as LowCarbon), as a significant environmental regulation, it may indirectly affect new energy enterprise value through its energy-saving and emission-reduction initiatives. Thus, we should exclude the effect of this policy to ensure the credibility of our conclusion.

In Table 6 column (2), we control for the Low-Carbon City Pilot Policy (denoted as LowCarbon). The results show that the implementation of the New Energy Demonstration City policy increases enterprise value by 8.4%, statistically significant at the 5% level, remaining largely consistent with the baseline regression estimates, proving that our results are robust.

4.2.7. Other Robustness Tests

Additionally, we conduct multiple robustness checks, including high-dimensional fixed effects, sample selection controls, and a lagged policy term (L.Treat × Post). As shown in Table 6 column (3), the results remain consistent across all specifications.

Our high-dimensional fixed effects model which controlling for province, firm, and year, estimates a 10.4% enterprise value increase from the policy, confirming its robust positive effect. When restricting the sample to 2010–2020 to exclude COVID-19 impacts (Table 6 column (4)), the effect strengthens to 11.4%, statistically significant at the 1% level. The lagged term (L.Treat × Post) remains significant with a coefficient of 0.077 (Table 6 column (5)), demonstrating persistent policy effects.

Furthermore, we employ multiple clustering approaches (city, industry, and year levels), with all results remaining statistically significant as shown in

Table 7. Alternative Clustering Methods.

	Cluster (City)	Cluster (Ind)	Cluster (Year)
	(1)	(2)	(3)
Treat × Post	0.131 ***	0.131 *	0.131 **
	(2.723)	(1.756)	(2.215)
Controls	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Id FE	Yes	Yes	Yes
R2	0.439	0.439	0.439
Observations	1090	1090	1090

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

. Specifically, Table 7 demonstrates the robustness of the policy effect estimate under alternative clustering methods, with the Treat × Post coefficient remaining stable at 0.131 and statistically significant across all specifications—showing 1% significance with city-level clustering, 10% with industry-level clustering, and 5% with year-level clustering—while all models control for firm and year fixed effects, include full control variables, and maintain an R² of 0.439 based on 1090 observations, confirming the reliability of the results against varying dependency structures in the error term.

Additionally, we provide a table of Alternative Fixed Effects in the Appendix A 

Table A1. Alternative Fixed Effects.

	Cluster (Year)	Cluster (City)	Cluster (Ind)	Cluster (City Ind)
Variable	(1)	(2)	(3)	(4)
Treat × Post	0.082 **	0.138 *	0.138 *	0.138 *
	(2.273)	(1.989)	(1.876)	(1.801)
Size	−0.237 ***	−0.276 ***	−0.276 ***	−0.276 ***
	(−13.077)	(−12.679)	(−7.124)	(−10.179)
Lev	−0.179 **	0.049	0.049	0.049
	(−2.762)	(0.230)	(0.261)	(0.241)
ROA	0.839 *	1.292 **	1.292 **	1.292 **
	(1.854)	(2.657)	(2.338)	(2.390)
Cashflow	0.536 ***	0.433 *	0.433 *	0.433
	(3.330)	(1.892)	(1.992)	(1.523)
Growth	−0.012	0.039	0.039	0.039
	(−0.477)	(0.915)	(1.094)	(0.835)
ListAge	0.006	0.068 *	0.068 **	0.068 **
	(0.178)	(1.736)	(2.470)	(2.184)
ATO	0.157 ***	−0.014	−0.014	−0.014
	(5.140)	(−0.241)	(−0.210)	(−0.229)
GDPct3	0.232 **	1.085 **	1.085 **	1.085 **
	(2.523)	(2.384)	(2.298)	(2.207)
POP	0.000 ***	−0.000	−0.000	−0.000
	(3.107)	(−0.919)	(−0.862)	(−0.895)
YearFE	Yes	No	No	No
IdFE	No	Yes	Yes	Yes
City	No	Yes	Yes	Yes
R2	0.409	0.518	0.518	0.518
Observations	1091	1090	1090	1090

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

, which further confirms the robustness of our baseline results under different model specifications, it demonstrates robust policy effects across alternative clustering methods, with the Treat × Post coefficient remaining significant (ranging 0.082–0.138) under all specifications. Control variables maintain expected signs and significance, confirming the reliability of the baseline results. All robustness checks corroborate the baseline findings.

4.3. Heterogeneity Tests

4.3.1. Environmental Regulation Intensity

Firms’ competitive advantages are systematically shaped by regional environmental regulation intensity. Given significant cross-regional variation, we examine how the New Energy Demonstration City policy’s impact on firm value differs by regulatory stringency. In high-stringency regions, stringent standards compel firms to increase environmental investments and accelerate technological innovation. Conversely, low-stringency regions’ laxer standards reduce compliance pressures but may incur long-term environmental risks.

Following Shao et al. [112], we measure environmental regulation stringency by the ratio of environmental-related text to total word count in municipal government work reports. Using 2009 report data (preceding our 2010–2023 firm sample), we classify cities into weak, moderate, and strong regulation groups for stratified regression analysis.

As shown in Columns (1)–(3) of

Table 8. Heterogeneity Tests of Environmental Regulation Intensity and Ownership.

	Environmental Regulation Intensity			Ownership
	Weak	Modern	Strong	SOE	Non-SOE
	(1)	(2)	(3)	(4)	(5)
Treat × Post	0.088	0.061	0.135 **	0.086 **	0.068
	(1.158)	(0.873)	(2.188)	(2.094)	(0.719)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes
R2	0.534	0.576	0.650	0.558	0.608
Observations	365	368	356	831	260

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. ** indicate statistical significance at the 5%.

, they reveal significant heterogeneity in the policy’s impact across regions with different environmental regulation intensities. Most notably, the policy increases new energy enterprise value by 13.5% in regions with strong environmental regulations. This stronger effect likely stems from three key factors: first, these regions had already established proactive government support systems for green development prior to the policy; second, local enterprises possessed more mature environmental technology R&D capabilities through long-term accumulation; third, the implementation of the new policy created effective government–enterprise collaboration—with enterprises rapidly transforming their existing technological reserves into applications while local governments provided coordinated supporting services. This synergistic mechanism ultimately enhanced the value creation effect of the policy in high-regulation regions.

In contrast, the policy shows no significant effect on firm value in medium- and low-regulation regions. This may reflect weaker regions’ economic growth-oriented strategies, where some firms prioritize short-term profits over environmental protection and maintain pollution-intensive production modes. Such approaches face growing sustainability constraints from rising environmental costs and resource limitations. When the policy was implemented, these firms’ limited technical preparedness and inadequate policy response mechanisms hindered their ability to capitalize on policy benefits, resulting in weaker performance compared to high-regulation regions.

In summary, the impact of the New Energy Demonstration City policy on new energy enterprise value exhibits significant heterogeneity across regions with varying environmental regulation intensities, thereby confirming Hypothesis 4.

Environmental regulation intensity is measured by the ratio of environmental-related text word count to the total word count in municipal government work reports of the cities where firms are located. Firms are categorized into low, medium, and high intensity groups based on terciles of this ratio.

4.3.2. Ownership

Firms’ ownership types critically influence their responses to the New Energy Demonstration City policy. As shown in Columns (4)–(5) of Table 8, the policy significantly boosts SOEs’ value by 8.6% due to their superior access to government resources and established market positions. However, their complex decision-making processes may hinder rapid innovation. Conversely, while non-SOEs demonstrate greater operational flexibility for strategic adjustments, their limited scale and credit profiles constrain policy resource acquisition, resulting in statistically insignificant effects. These contrasting outcomes confirm Hypothesis 5 and reflect fundamental differences in how ownership structures mediate policy impacts through varying combinations of resource advantages and organizational constraints.

4.3.3. Firm Size

Firm size fundamentally shapes corporate capabilities and policy responsiveness. We classify firms into large and small groups based on total assets for separate regression analyses in

Table 9. Heterogeneity Tests of Firm Size and Capital Intensity.

	Firm Size		Capital Intensity
	Large	Small	Low	Medium	High
	(1)	(2)	(3)	(4)	(5)
Treat × Post	0.103 **	0.019	0.075	0.160 *	0.011
	(2.346)	(0.283)	(1.054)	(1.939)	(0.289)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes
R2	0.501	0.398	0.574	0.577	0.663
Observations	539	552	357	344	349

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, ** indicate statistical significance at the 10%, 5% level, respectively. Firm size is classified into large and small groups based on year-end total assets. Capital intensity is measured by the ratio of total assets to operating revenue, and firms are classified into low, medium, and high capital intensity groups based on terciles of this ratio.

with column (1)–(2). The results show that for large firms, significant 10.3% value increases due to superior resource access and economies of scale. For small firms, statistically insignificant effects, they may be constrained by limited R&D capacity and higher compliance costs. These findings confirm Hypothesis 5, demonstrating that the New Energy Demonstration City policy’s benefits concentrate disproportionately among larger firms with stronger implementation capabilities.

4.3.4. Capital Intensity

The analysis of capital intensity heterogeneity reveals nuanced policy effects across firms with different investment structures. Capital-intensive firms, characterized by substantial fixed asset investments and economies of scale, demonstrate distinct responses to the New Energy Demonstration City policy compared to their less capital-intensive counterparts. We measure capital intensity using the total-assets-to-revenue ratio and categorize firms into terciles (low, medium, high) for stratified regression analysis.

The results presented in Table 9 with column (3)–(5), which indicate that medium capital-intensity firms experience the most pronounced positive effects, with statistically significant value increases following policy implementation. These firms appear optimally positioned to leverage policy support, possessing sufficient physical capital to implement technological upgrades while maintaining flexibility in resource allocation. In contrast, low capital-intensity firms show no significant improvement, likely constrained by their limited fixed asset base and consequent difficulties in scaling production, even with policy assistance. At the other extreme, high capital-intensity firms also exhibit insignificant policy effects, suggesting that their existing substantial capital buffers and established market positions leave less room for additional policy-induced value creation. This pattern of results highlights the non-linear relationship between capital intensity and policy effectiveness, where intermediate levels of capital investment appear most conducive to translating policy support into measurable value gains. The findings underscore how firms’ pre-existing capital structures fundamentally shape their capacity to benefit from environmental regulations, with both insufficient and excessive capital intensity potentially limiting policy responsiveness. These insights carry important implications for policymakers seeking to tailor interventions to different segments of the new energy sector.

5. Mechanism

The New Energy Demonstration City policy effectively alleviates financing constraints for new energy enterprises through multiple channels. By providing subsidies, tax incentives, and industrial fund support, the policy increases corporate funding while reducing financing costs. The government’s credit endorsement further enhances confidence among financial institutions and investors, expanding financing channels and helping firms overcome financial bottlenecks to accelerate growth and enhance market value. As a key instrument for promoting new energy development, the policy stimulates green innovation and transformation through research subsidies and tax benefits. These measures encourage firms to increase R&D investment, boost independent innovation capabilities, and ultimately improve corporate social image, brand value, and profitability.

To systematically examine these mechanisms, we analyze three pathways—financing constraints (WW), green innovation (EnvrPat), and green transformation (Green)—using the same subgroup classifications from our heterogeneity analysis. This approach provides deeper insights into how the policy influences enterprise value across different firm characteristics.

5.1. Financing Constraints

The analysis of financing constraint mechanisms reveals nuanced policy effects across different firm characteristics. Recognizing that certain variables may influence financing constraints, we exclude Lev, Cashflow, and ATO from the regression.

Table 10. Mechanisms: Financing Constraints-Environmental Regulation Intensity and Ownership. (Excludes Lev, Cashflow, ATO variables).

	Environmental Regulation Intensity				Ownership
	All	Weak	Modern	Strong	All	SOEs	Non-SOEs
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	−0.005 **	−0.006	0.002	−0.011 ***	−0.005 **	−0.006 **	0.003
	(−2.482)	(−1.287)	(0.385)	(−2.672)	(−2.482)	(−2.274)	(0.584)
Size	−0.050 ***	−0.050 ***	−0.049 ***	−0.052 ***	−0.050 ***	−0.050 ***	−0.055 ***
	(−46.689)	(−20.925)	(−35.469)	(−31.473)	(−46.689)	(−40.884)	(−16.435)
ROA	−0.274 ***	−0.316 ***	−0.277 ***	−0.248 ***	−0.274 ***	−0.259 ***	−0.311 ***
	(−12.261)	(−9.112)	(−8.462)	(−5.676)	(−12.261)	(−9.590)	(−8.608)
Growth	−0.034 ***	−0.035 ***	−0.031 ***	−0.036 ***	−0.034 ***	−0.034 ***	−0.032 ***
	(−14.820)	(−10.661)	(−7.155)	(−9.495)	(−14.820)	(−12.143)	(−9.651)
ListAge	0.006 ***	0.005	0.005 **	0.005 **	0.006 ***	0.007 ***	0.002
	(4.081)	(1.531)	(2.576)	(2.522)	(4.081)	(3.902)	(0.834)
GDPct3	−0.009	−0.036 *	−0.015	−0.001	−0.009	−0.015	0.034
	(−1.090)	(−1.847)	(−0.805)	(−0.056)	(−1.090)	(−1.573)	(1.150)
POP	0.000	0.000	0.000	−0.000	0.000	−0.000	0.000
	(0.209)	(0.268)	(0.500)	(−0.925)	(0.209)	(−0.542)	(0.686)
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.893	0.880	0.927	0.897	0.893	0.904	0.849
Observations	1070	357	360	351	1070	815	255

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

reports results without these financing constraint-related variables. However, given these are conventional controls affecting Tobin’s Q, and considering that the WW index offers a more comprehensive measure of financing constraints, we present results including these variables in the Appendix A 

Table A2. Mechanisms: Financing Constraints-Environmental Regulation Intensity and Ownership.

	Environmental Regulation Intensity				Ownership
	All	Weak	Modern	Strong	All	SOEs	Non-SOEs
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	−0.004 **	−0.006	0.004	−0.009 **	−0.004 **	−0.005 *	0.004
	(−2.109)	(−1.346)	(1.108)	(−2.283)	(−2.109)	(−1.923)	(0.827)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.901	0.888	0.935	0.915	0.901	0.916	0.858
Observations	1070	357	360	351	1070	815	255

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, ** indicate statistical significance at the 10%, 5% level, respectively.

and

Table A3. Mechanisms: Financing Constraints-Firm Size and Capital Intensity.

	Firm Size			Capital Intensity
	All	large	Small	All	Low	Medium	High
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	−0.004 **	−0.005 *	−0.003	−0.004 **	0.001	−0.012 ***	0.002
	(−2.109)	(−1.888)	(−0.932)	(−2.109)	(0.275)	(−3.365)	(0.400)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.901	0.871	0.783	0.901	0.940	0.913	0.915
Observations	1070	535	535	1070	348	334	347

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

. As shown in Table 10, the New Energy Demonstration City policy significantly reduces financing constraints overall by 0.5%, with particularly strong effects in high environmental regulation regions. This regional variation suggests that stringent environmental standards enhance policy effectiveness, likely because compliant firms gain greater credibility with financial institutions and preferential access to green financing instruments. The results align with the baseline findings, indicating that financing constraint alleviation serves as an important channel through which the policy enhances enterprise value.

When examining ownership heterogeneity, the policy demonstrates distinct impacts—effectively alleviating financing constraints for SOEs while showing no significant effect for non-SOEs. This divergence reflects fundamental differences in their operating environments and characteristics: SOEs benefit from established relationships with government-backed banks, higher asset collateral values, and perceived implicit government guarantees, all of which amplify the policy’s financial channel effects. In contrast, non-SOEs face structural barriers in capital markets that appear to limit their ability to capitalize on the policy’s financing benefits.

In Appendix Table A2 and Table A3, which include Lev, Cashflow, ATO variables, reaffirm our core findings: the NEDC policy significantly alleviates financing constraints, as captured by the WW index, with coefficients remaining statistically significant and negative in the full sample. The heterogeneous effects also persist: the policy’s mitigating impact is particularly pronounced in regions with strong environmental regulations, among SOEs, and for firms with medium capital intensity. These consistent results across alternative model specifications underscore the reliability of financing constraints as a key mechanism through which the NEDCP enhances the value of new energy enterprises.

These findings collectively underscore how the financing constraint mechanism operates differently across regulatory and ownership contexts. The policy’s value-enhancing effects prove strong when implemented in environments with either stringent regulations that incentivize green compliance or ownership structures that facilitate access to policy-supported financing. The results help explain the heterogeneous treatment effects observed earlier while validating financing constraint reduction as a key transmission channel for the policy’s economic impacts. Importantly, the analysis demonstrates that the policy’s financial market effects depend critically on both external regulatory conditions and internal firm characteristics.

The analysis further examines how financing constraint mechanisms vary by firm size and capital intensity. Columns (1)–(3) of

Table 11. Mechanisms: Financing Constraints-Firm Size and Capital Intensity. (Excludes Lev, Cashflow, ATO variables).

	Firm Size			Capital Intensity
	All	Large	Small	All	Low	Medium	High
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	−0.005 **	−0.008 ***	−0.005	−0.005 **	−0.001	−0.014 ***	0.000
	(−2.482)	(−2.726)	(−1.321)	(−2.482)	(−0.162)	(−3.444)	(0.013)
Size	−0.050 ***	−0.047 ***	−0.049 ***	−0.050 ***	−0.050 ***	−0.047 ***	−0.053 ***
	(−46.689)	(−38.670)	(−21.569)	(−46.689)	(−38.762)	(−31.286)	(−32.591)
ROA	−0.274 ***	−0.338 ***	−0.248 ***	−0.274 ***	−0.250 ***	−0.253 ***	−0.284 ***
	(−12.261)	(−13.644)	(−8.507)	(−12.261)	(−9.027)	(−6.006)	(−7.305)
Growth	−0.034 ***	−0.033 ***	−0.032 ***	−0.034 ***	−0.033 ***	−0.036 ***	−0.033 ***
	(−14.820)	(−12.872)	(−8.919)	(−14.820)	(−12.790)	(−5.980)	(−8.161)
ListAge	0.006 ***	0.002 *	0.009 ***	0.006 ***	0.005 **	0.004 **	0.011 ***
	(4.081)	(1.786)	(5.291)	(4.081)	(2.341)	(2.114)	(4.715)
GDPct3	−0.009	−0.008	−0.029 **	−0.009	0.028 **	−0.019	−0.019
	(−1.090)	(−0.766)	(−2.056)	(−1.090)	(2.259)	(−1.055)	(−1.277)
POP	0.000	−0.000	0.000	0.000	0.000	0.000	−0.000
	(0.209)	(−0.398)	(0.199)	(0.209)	(1.176)	(0.328)	(−0.311)
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.893	0.852	0.759	0.893	0.934	0.900	0.901
Observations	1070	535	535	1070	348	334	347

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

reveals the policy significantly alleviates financing constraints for large firms but shows no meaningful effect for small firms, consistent with their divergent resource endowments and risk-bearing capacities. This size-based pattern aligns with earlier heterogeneity findings, suggesting large firms’ superior access to policy-supported financing channels enhances their responsiveness.

When considering capital intensity Table 11 with Columns (4)–(7), the policy’s constraint-reducing effects concentrate among medium-intensity firms, likely because their balanced capital structures optimally match policy support with financiers’ risk preferences. Neither low-intensity firms which have limited collateral nor high-intensity firms which own abundant internal funds that exhibit significant financing constraint improvements, again mirroring prior heterogeneity results. These findings collectively demonstrate that the policy’s financial channel operates most effectively for firms with intermediate resource profiles—sufficiently capitalized to meet lender requirements yet still dependent on external financing for green investments. The results underscore how firm characteristics systematically shape policy transmission mechanisms.

5.2. Green Innovation

This paper measures green innovation (EnvrPat) by taking the natural logarithm of the sum of green invention patent applications, green utility model patent applications, and 1. First, we examine the differential effects of the green innovation mechanism under varying environmental regulations. As shown in columns (1)–(3) of

Table 12. Mechanisms: Green Innovation Environmental Regulation Intensity and Ownership.

	Environmental Regulation Intensity				Ownership
	All	Weak	Modern	Strong	All	SOEs	Non-SOEs
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	0.198 **	−0.028	0.193	0.661 ***	0.198 **	0.257 **	0.182
	(2.035)	(−0.154)	(1.216)	(3.490)	0.198 **	0.257 **	0.182
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.466	0.586	0.545	0.538	0.466	0.526	0.427
Observations	1011	326	344	341	1011	767	244

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. ** and *** indicate statistical significance at the 5% and 1% level, respectively.

, the results indicate that the New Energy Demonstration City policy influences the value of new energy firms through green innovation, with a more pronounced effect in regions with stringent environmental regulations.

Specifically, column (1) reveals that the policy significantly enhances green innovation by 19.8%. Combined with the baseline regression results, this suggests that the policy indirectly boosts firm value by incentivizing greater green innovation. Comparing columns (2)–(4), the regression coefficient for regions with strict environmental regulations is 0.661, significant at the 1% level, indicating that the policy’s impact on green innovation is stronger in these areas. This may be because stringent regulations, coupled with policy guidance and market pressures, compel firms to engage more actively in green innovation. Increased investment in green R&D and eco-friendly process improvements not only aligns with policy objectives but also enhances firms’ sustainable development capabilities and social reputation. Consequently, the NEDCP exerts a more substantial positive effect on firm value through green innovation in high-regulation regions. In contrast, in regions with weak or moderate environmental regulations, the policy’s weaker constraints may fail to sufficiently motivate green innovation, limiting its indirect effect on firm value. These findings align with the earlier heterogeneity analysis on environmental regulations.

Then, we examine the differential effects of the green innovation mechanism across firms with different ownership types. As shown in columns (5)–(7) of Table 12, the results indicate that the NEDCP policy significantly enhances the value of state-owned enterprises (SOEs) by promoting green innovation, while its impact on non-SOEs is statistically insignificant.

Specifically, the regression coefficient for SOEs is 0.257, significant at the 5% level, suggesting that the policy effectively stimulates green innovation in SOEs. This divergence may stem from inherent differences in ownership structure, resource allocation, and institutional support. SOEs, being more responsive to policy directives, likely benefit from stronger government backing and preferential incentives in green innovation initiatives. In contrast, non-SOEs primarily innovate under market competition, where resource allocation is more fragmented and subject to competitive pressures, potentially leading to less stable and sustained investments in green innovation. These findings align with the earlier ownership-based heterogeneity analysis.

Next, we analyze the heterogeneous effects of the green innovation mechanism across firms of different sizes. As shown in columns (1)–(3) of

Table 13. Mechanisms: Green Innovation-Firm Size and Capital Intensity.

	Firm Size			Capital Intensity
	All	Large	Small	All	Low	Medium	High
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	0.198 **	0.701 ***	−0.192 *	0.198 **	1.122 ***	−0.319 **	0.091
	(2.035)	(4.089)	(−1.906)	(2.035)	(4.810)	(−2.027)	(0.650)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.466	0.560	0.392	0.466	0.510	0.561	0.392
Observations	1011	517	494	1011	326	319	328

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively.

, the NEDCP significantly affects firm value through green innovation, but with opposing effects for large and small firms. Specifically, the policy’s coefficient for large firms is 0.701, while for small firms it is −0.192. This divergence suggests that large firms, with their stronger financial and technological capabilities, are better positioned to align with the policy’s green innovation objectives. In contrast, small firms, constrained by limited resources and higher risk exposure, struggle to sustain substantial green R&D investments, leading to a negative valuation effect.

Finally, we examine the differential impacts across firms with varying capital intensity. Columns (4)–(7) of Table 13 show that the policy significantly influences low- and medium-capital intensity firms but not high-intensity ones. For low-capital-intensity firms, the coefficient is 1.122, indicating a strong positive effect, likely due to their operational flexibility in reallocating resources toward green innovation. Conversely, medium-capital-intensity firms exhibit a negative coefficient, possibly because their complex asset structures hinder agile transitions to green practices. These results further underscore the policy’s uneven effectiveness across firm characteristics.

This study further examines the green innovation mechanism by classifying it into two distinct types: green inventive innovation (EnvrInvPat), measured as the natural logarithm of green invention patent applications plus one, and green utility innovation (EnvrUtyPat), measured as the natural logarithm of green utility model patent applications plus one.

Table 14. Heterogeneous Effects of Green Innovation Mechanisms.

	Green Innovation	Green Inventive Innovation	Green Utility Innovation
	(1)	(2)	(3)
Treat × Post	0.198 **	0.165 *	0.061
	(2.035)	(1.805)	(0.879)
Controls	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Id FE	Yes	Yes	Yes
R2	0.466	0.446	0.422
Observations	1011	1011	1011

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, ** indicate statistical significance at the 10%, 5% level, respectively.

presents the differential impacts of these innovation types on the value of new energy firms.

The results reveal notable differences in policy effectiveness. As shown in columns (2) and (3), green inventive innovation exhibits a statistically significant coefficient of 0.165, while green utility innovation shows an insignificant coefficient of 0.061. This suggests that the NEDCP primarily enhances firm value through green inventive innovation rather than utility-oriented innovation.

The disparity likely stems from the fundamental nature of these innovation types. Green inventive innovation typically involves pioneering R&D efforts that demand substantial resource commitments and carry higher risks. However, successful breakthroughs in this domain can yield distinctive technological advantages, stronger market positioning, and enhanced brand reputation for firms. In contrast, utility-focused innovations, while valuable for incremental improvements, may lack the transformative potential to significantly elevate firm value under this policy framework. These findings align with the view that breakthrough innovations drive more substantial value creation in emerging technology sectors like new energy.

5.3. Green Transformation

This paper measures corporate green transformation (Green) using keyword frequency statistics from listed companies’ annual reports. This study draws on methodologies from multiple scholars [109,113,114,115,116,117,118,119], particularly following Wu and Li (2022) [120], to construct a green transition framework based on three dimensions: “institutional green transition,” “operational green transition,” and “supportive green transition.” Using machine learning, we identified additional keywords highly correlated with initial key terms to expand the green transition lexicon. We then employed Python3.13’s Jieba package to scan and match keywords in listed firms’ annual reports, counting the frequency of each term. To control for report length variation, we divided keyword frequency by the total word count to develop a Green Transition Strength (GTS) index, which was normalized for comparability. Higher GTS values indicate stronger corporate green transition performance. Our empirical analysis first investigates how this transformation mechanism operates differently across regions with varying environmental regulation intensities. The results presented in

Table 15. Mechanisms: Green Transformation Environmental Regulation Intensity and Ownership.

	Environmental Regulation Intensity				Ownership
	All	Weak	Modern	Strong	All	SOEs	Non-SOEs
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	0.085 *	0.102	0.010	0.194 **	0.085 *	0.109 *	0.054
	(1.732)	(1.122)	(0.104)	(2.157)	(1.732)	(1.848)	(0.468)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.721	0.774	0.744	0.713	0.721	0.735	0.701
Observations	1068	358	360	348	1068	818	250

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. *, ** indicate statistical significance at the 10%, 5% level, respectively.

with columns 1–4, which demonstrate that the NEDCP effectively enhances new energy firms’ value through promoting green transformation, particularly in regions with stringent environmental regulations. The estimation in column (1) indicates that the policy boosts green transformation by 8.5%, suggesting this channel significantly contributes to corporate value creation. More importantly, the coefficient reaches 0.194 and significant at 5% level in high-regulation regions from column (2), revealing that stringent environmental oversight strengthens the policy’s transformative effect by compelling firms to prioritize sustainable development and increase green technology investments. However, this positive impact becomes statistically insignificant in medium- and low-regulation areas, consistent with our previous findings on regulatory heterogeneity.

Further examination of ownership heterogeneity from columns (5)–(7) shows that the NEDCP’s effect on green transformation is concentrated in state-owned enterprises (SOEs), with no significant impact observed in non-SOEs. This differential effect likely stems from SOEs’ inherent advantages in policy implementation and resource acquisition due to their closer ties with government entities, whereas non-SOEs’ market-oriented operations may constrain their responsiveness to administrative policies. These findings not only validate our earlier ownership heterogeneity analysis but also highlight the importance of considering firm-specific characteristics when evaluating policy effectiveness. The study provides micro-level evidence on how environmental policies drive corporate sustainable transition, while emphasizing the need for differentiated policy designs that account for regional regulatory contexts and firm ownership structures to maximize their transformative impacts.

Finally, we examine the heterogeneous effects of the green transformation mechanism across firms of different sizes and capital intensity levels. As shown in

Table 16. Mechanisms: Green Transformation Firm Size and Capital Intensity.

	Firm Size			Capital Intensity
	All	large	Small	All	Low	Medium	High
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Treat × Post	0.085 *	0.091	0.074	0.085 *	0.112	−0.029	0.176
	(1.732)	(1.189)	(0.991)	(1.732)	(1.239)	(−0.294)	(1.603)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Id FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.721	0.693	0.750	0.721	0.766	0.738	0.717
Observations	1068	529	539	1068	351	338	338

Notes: Robust standard errors in parentheses. The controls and constant terms are omitted to save space. * indicate statistical significance at the 10% level.

, the results indicate that the NEDC policy’s impact through the green transformation channel is statistically insignificant for both firm size subgroups and capital intensity categories.

In summary, our analysis demonstrates that the NEDC policy affects new energy enterprise value through three distinct channels: alleviating financing constraints (WW), promoting green innovation (EnvrPat), and facilitating green transformation (Green). Importantly, the policy exhibits differential effects across various firm characteristics. These mechanism analysis results provide empirical support for both Hypothesis 2 and Hypothesis 3, confirming the multifaceted transmission channels through which the NEDC policy influences corporate value while highlighting the moderating role of firm heterogeneity in shaping policy outcomes. The findings suggest that while the policy effectively operates through financing and innovation channels, its transformative impact on corporate environmental practices may require stronger complementary measures, particularly for smaller firms and those with varying capital structures.

6. Conclusions and Policy Implications

This study examines the impact of China’s New Energy Demonstration City (NEDC) policy on the value of new energy enterprises within the context of the country’s dual-carbon goals and energy transition. Using panel data from listed new energy firms (2010–2023) and a difference-in-differences approach, we investigate both the policy’s effectiveness and its transmission channels. Our main findings are threefold. First, we document robust evidence that the NEDC policy significantly enhances the value of new energy enterprises. This positive effect withstands multiple rigorous tests, including propensity score matching, alternative outcome measures, exclusion of concurrent policy shocks, high-dimensional fixed effects, sample selection controls, and lagged specifications. Second, our heterogeneity analysis reveals important variations in policy effectiveness. The value-enhancing effect is particularly pronounced for: (1) firms in regions with stringent environmental regulations compared to those in moderate- or low-regulation areas; (2) state-owned enterprises relative to their non-state counterparts; (3) large firms versus small firms; and (4) medium capital-intensive enterprises compared to both low and high capital-intensive firms. Third, we identify three key transmission channels through which the policy operates: (1) alleviating financing constraints by reducing capital costs and expanding funding access—especially effective for firms in high-regulation regions, SOEs, large enterprises, and medium capital-intensive firms; (2) stimulating green innovation through increased R&D investment and technological upgrading—particularly evident in high-regulation areas, SOEs, large firms, and low capital-intensive enterprises; and (3) facilitating green transformation that enhances corporate reputation and competitiveness—most visible in strictly regulated regions and SOEs. These mechanisms collectively contribute to significant value creation in the new energy sector.

This study yields important policy implications for promoting the development of new energy enterprises. The findings suggest that policymakers should adopt differentiated approaches based on regional characteristics and firm heterogeneity to maximize policy effectiveness. For regions with stringent environmental regulations, enhanced policy support can further consolidate their advantages in the new energy sector, while other regions may require complementary measures in infrastructure and market development to improve policy implementation. This finding—that stricter environmental regulations amplify policy benefits—demonstrates the synergistic effect of incentive-based policies and stringent standards. It suggests policymakers should combine “carrots” (incentives) and “sticks” (penalties) to enhance policy design, integrating motivational and punitive environmental regulations for optimal impact. The government should tailor specific support mechanisms for different types of enterprises, including facilitating financing for non-SOEs, promoting industry chain integration among large firms, and providing technical assistance to small enterprises. Particularly for small private firms, a key implication of heterogeneous effects is that “carrot” policies (e.g., subsidies or demonstration zones) must be complemented by capacity-building programs—as these enterprises benefit less. Policymakers could tailor financial or technical support to such firms to extend policy impacts, enhance valuations, and boost multidimensional value (including profitability, influence, and external recognition), thereby creating broader spillover effects across the Small and medium-sized enterprises ecosystem. Strengthening financial support through dedicated industry funds and innovative green financial instruments is crucial to alleviate capital constraints. Meanwhile, increasing incentives for green innovation through R&D subsidies and industry-academia collaboration can enhance firms’ core competitiveness. Finally, accelerating green transformation by enforcing environmental standards and providing guidance on sustainable practices will help enterprises improve their market adaptability and long-term growth potential. These comprehensive measures, when properly implemented, can significantly enhance the effectiveness of new energy policies and contribute to China’s energy transition and dual-carbon goals.

This study also has several limitations. Due to the lack of accounting data on firms’ investment structure shifts toward green activities in all publicly available databases, we rely on textual analysis of annual reports to capture corporate commitment to green practices, even if this is widely used in the literature serves as a proxy for firms’ strategic emphasis on green transformation. However, whether these textual commitments translate into tangible investments or operational changes remains an open question due to data constraints. Future research would benefit from more granular firm-level investment data to directly track green capital expenditures, green asset ratios, or green revenue shares, thereby offering a more concrete understanding of how environmental incentive policies boost enterprises’ market value.