The Green Paradox in NEV Manufacturing: Regulatory Impacts on Innovation from a Stakeholder Perspective

Abstract

This paper explores the paradoxical impact of environmental regulations on green innovation in the manufacturing of new energy vehicles (NEVs) from a stakeholder perspective. We address the dual challenge of accelerating green innovation across various diffusion stages and refining environmental regulations for effective stakeholder engagement, including the central government, upstream suppliers, and internal operations teams. First, we utilize Stackelberg game theory to analyze the strategic interaction and behavioral rationale between local governments and NEV manufacturers at different stages of innovation diffusion, represented by specific parameter sets. Second, we examine the roles of four key stakeholders, exploring their unique impact mechanisms and potential to influence the game’s Nash equilibrium. Finally, the game models’ validity and primary conclusions are corroborated with real-world case studies, prominently including the ongoing shift of Chinese automakers towards NEVs. Results demonstrate that: (1) environmental regulations should be calibrated according to the diffusion stages of green innovation, and (2) misaligned stakeholder interests can lead to environmental regulations that inadvertently hinder, rather than promote, NEV manufacturers’ green innovation efforts.

1. Introduction

Environmental regulation aims to address the energy and environmental challenges that jeopardize human well-being and sustainable development. In China, a key target of environmental regulation is the manufacturing sector, which accounts for more than half of the country’s total energy consumption and hence emissions (see Figure 1). To mitigate its environmental impacts and enhance its resource efficiency, this sector is prompted by environmental regulation to invest more in green innovation, including the creation and adoption of new technologies, products, or processes [1]. Although the Porter hypothesis posits that compliance costs of environmental regulations are only transitory and that manufacturers can attain a competitive advantage through green innovation over time [2,3], this may not be fully applicable to China. In China, many manufacturers face challenges such as low profit margins, human resource shortages, and insufficient innovation capabilities. Furthermore, this situation can be exacerbated by the “pollution shelter effect”, whereby polluting industries may choose to relocate to regions with more lenient environmental regulation [4]. These challenges are not unique to China but are also prevalent in developing countries globally [5,6]. Therefore, to advance the manufacturing sector in developing countries, it is imperative to devise a more nuanced and effective environmental regulatory framework that balances environmental objectives with practical constraints. Additionally, it is essential to investigate the mechanisms and factors that influence the green innovation behavior of manufacturing firms.

Previous studies have investigated the impact of environmental regulation on green innovation in China and revealed that the overall effect may not always be favorable [7,8]. It implies that tighter environmental regulation, instead of incentivizing firms to adopt greener technologies, can deter them from doing so. However, this does not mean that environmental regulation is entirely detrimental; it depends on how it is implemented. For example, some literature suggests that market-based and voluntary environmental regulations are more likely to have intended effects, such as environmental taxes, subsidies, and certification. In contrast, command-and-control environmental regulation, such as emission standards and penalties, may not end favorably [9,10,11]. Furthermore, the relationship between environmental regulation and green innovation is more often than not found to be nonlinear. Specifically, it can display a U-shaped or inverted U-shaped pattern under different types of environmental regulation [12,13]. This suggests that there is an optimal level of environmental regulation for fostering green innovation, and it varies under different circumstances.

Therefore, scholars tend to agree with a weaker version of Porter’s hypothesis, which states that stakeholders, along with other factors, can also influence innovation besides environmental regulation itself [3]. For example, stakeholders such as customers and suppliers can affect firms’ decisions and behaviors regarding green innovation [14,15]. That is, the ultimate effect of environmental regulation on green innovation may also depend on stakeholders’ pressure. In this light, this study aims to further explore the dynamic process between the local government and manufacturers using a stakeholder approach. A Stackelberg game model with complete information is then developed and extended to incorporate stakeholders such as internal operation teams, upstream suppliers, and the central government. The local government may choose either strict or lax environmental regulation, and manufacturers may adopt either proactive or passive green innovation [16]. Additionally, through a sensitivity analysis of key model parameters, this paper also identifies three distinct Nash equilibrium scenarios, which exhibit characteristics reminiscent of the introduction, growth, and maturity stages of the innovation diffusion theory [17].

By employing a mixed-methods approach that combines a Stackelberg game modeling, stakeholder analysis, and an innovation diffusion perspective, this study delves into the complex dynamics of decision-making in environmental regulation and green innovation, and aims to answer the following key questions:

(1) How do manufacturers decide whether to adopt proactive or cautious green innovation strategies in response to environmental regulations? And how do operating costs, consumer preferences, and technological advancement affect these decisions?

(2) How do the adoption and diffusion of green innovation influence the effectiveness of environmental regulations? And under what conditions are these regulations unnecessary, necessary, or paradoxical for fostering green innovation?

(3) How do the interactions among key stakeholders, including internal operations teams, upstream suppliers, and the central government (or non-governmental organizations, as commonly seen in other countries), influence the decision-making dynamics within the game model? What are the implications of the model’s findings for real-world case studies or applications?

The rest of the paper is organized as follows: Section 2 reviews the relevant literature on environmental regulation and green innovation and identifies the research gap and contribution of this study. Section 3 presents the main assumptions and their rationale, key symbols and definitions, and the payoff matrix of the Stackelberg model. Section 4 analyzes the optimal strategy of manufacturers and the local government under different market conditions with the sensitivity analysis and extends the model by incorporating the influence of stakeholders, such as an internal operation team, upstream suppliers, and the central government. Section 5 verifies the extended model with a case study of China’s new energy vehicle industrial development. Section 6 concludes the paper by discussing the main findings, their implications for theory and policy, and the directions for future research.

2. Related Literature

Much academic research on environmental regulation and green innovation focuses on examining the validity and generalizability of the Porter hypothesis under diverse contexts and has found that the applicability of the Porter hypothesis (1) may vary across different industries, depending on their specific characteristics, such as the intensity of competition, the nature and pace of innovation, and the stringency of environmental regulations [4,18]; (2) may be influenced by the choice of data sources, such as patent applications, innovation inputs, productivity measures, environmental performance indicators, and firm-level surveys [19,20]; (3) may depend on the specific type of environmental regulation employed, such as command-and-control, market-based instruments, information disclosure, or voluntary approaches [21]. While revisiting the Porter hypothesis has emerged as a prominent research area, a limited number of studies have delved into the deeper theoretical underpinnings in cases where the hypothesis does not hold true, particularly with a game-theoretic approach.

Despite the scarcity of research utilizing game theory in this context, existing studies can be broadly categorized into two main groups: evolutionary game theory and dynamic game theory. The former group assumes bounded rationality for both local governments and manufacturers, implying that they adapt their strategies through repeated interactions [22,23]. The latter group, on the other hand, assumes perfect or complete rationality for both actors and focuses on analyzing their cooperative or non-cooperative behavior in different scenarios [24,25]. While both approaches offer valuable insights into strategic decision-making, they also have limitations: (1) evolutionary game theory, while capturing strategic adaptation through repeated interactions, might not fully capture the reality of China nowadays, where local governments face limitations in policy changes due to central government constraints. (2) Furthermore, dynamic game theory may overlook the influence of key stakeholders, such as internal operations teams, external suppliers, and central government inspection teams, who can significantly impact the dynamics of environmental regulation and green innovation.

To address the research gap in the literature, this study employs a mixed-method approach, combining the Stackelberg game model, stakeholder analysis, and an innovation diffusion perspective. It aims to analyze the strategic interaction between local governments and manufacturers at different stages of green innovation diffusion and explore how factors such as sales revenue, operating costs, consumer preferences, and technological advancement shape the effectiveness of environmental regulations. Additionally, a real-world case study of China’s NEV industry is utilized to explore the practical implications of the model’s findings.

This study contributes to the existing literature in the following ways: (1) it reveals that the effectiveness of environmental regulation in promoting green innovation varies across the stages of innovation diffusion. (2) It highlights the potential for diverse stakeholders, such as the central government (or non-governmental organizations), upstream suppliers, and the operation team, to impede the green innovation process, potentially acting as hindrances that may contradict the intended effects of environmental regulation. (3) Moreover, it applies the extended game model to a real-world case study, focusing on how stakeholder interactions impact the development of China’s NEV industry. This comprehensive analysis provides policy-relevant insights by unveiling the complex interplay between environmental regulations, stakeholders, and green innovation. This knowledge can equip policymakers with a nuanced understanding to develop more effective strategies for promoting innovation in the green technology sector.

3. Model Development: The Stackelberg Framework for Analyzing Environmental Regulation and Green Innovation

3.1. Local Government and Manufacturer Strategies in the Game Framework

Promoting green innovation through environmental regulation presents a complex challenge for both local governments and manufacturers. Local governments juggle competing objectives such as job creation, public funding, and aligning with the demands of the central government and non-governmental organizations (NGOs), all while balancing short-term gains with long-term benefits [26,27]. Manufacturers, on the other hand, often lack inherent motivation for green innovation due to their short-term profit focus and uncertainty surrounding policy expectations [28,29]. However, the landscape is evolving: market demands for environmentally friendly products are rising [30], and governments are implementing incentives like subsidies, tax cuts, and lower-interest loans to encourage green innovation. Therefore, a thorough understanding of this dynamic background is essential for informing the assumptions underlying the model developed hereafter.

To address these challenges, local governments and manufacturers can adopt various strategies that significantly influence the outcomes for both parties. In this paper’s Stackelberg model, the local government is assumed to choose between “lax environmental regulation” and “strict environmental regulation”. The former prioritizes limited government intervention, relying on market forces and typically not promoting green transition through additional compliance costs for manufacturers. Conversely, strict environmental regulation aims to achieve environmental sustainability through stricter regulations and enforcement mechanisms, often accompanied by environmental penalties [31]. Similarly, manufacturers can choose between “cautious green innovation” and “proactive green innovation”. Cautious green innovation resembles incremental innovation, utilizing and improving existing technologies or processes. This approach typically entails lower costs and risks but generates less environmental benefit. In contrast, proactive green innovation resembles radical innovation, encouraging investment in developing and adopting new technologies or processes. This strategy often involves high investment risks but also offers greater potential for environmental benefits [32]. Therefore, the specific payoffs for both governments and manufacturers will depend on the specific combinations of strategies they choose.

Table 1. Symbols and definitions used in the model.

Symbols	Definitions
${\psi }_1$	Manufacturers’ operating cost when implementing cautious green innovation (CI)
${\psi }_2$	Manufacturers’ operating cost when implementing proactive green innovation (PI)
${\rho }_1$	Manufacturers’ R&D investment when implementing cautious green innovation (CI)
${\rho }_2$	Manufacturers’ R&D investment when implementing proactive green innovation (PI)
${\delta }_1$	Manufacturers’ added value from cautious green innovation.
${\delta }_2$	Manufacturers’ added value from proactive green innovation.
${\eta }_c$	Local government’s social and political recognition under different strategy combinations
$\tau$	The local sales tax rate applied to the manufacturer’s sales revenue
P	The standard sales price charged by manufacturers for products without green innovation
$Q_c$	The sales quantity achieved by manufacturers under different strategy combinations
${\pi }_i$	Environmental penalty paid when manufacturers implementing cautious green innovation
$\widehat{{\pi }_i}$	Environmental penalty paid when manufacturers implementing proactive green innovation

All variables are expressed in per-unit terms, except for $\tau$, ${\eta }_c$, and $Q_c$.

provides the notation and definitions of the relevant parameters used in the model.

3.2. Modeling Technological Progress: The S-Curve and Its Influence on Key Parameters

Even though manufacturers can improve their environmental performance through green innovation, the relationship may not be linear. According to the technology S-curve theory, technological progress often follows an S-shaped curve. This perspective emphasizes that improvement induced by increasing R&D input is slow at first, then accelerates, and then slows down again [33]. Furthermore, ambidexterity theory offers complementary insights into this phenomenon by differentiating between two types of innovation: exploitative and exploratory. Specifically, exploitative innovation, characterized by incremental changes focused on improving existing technologies, aligns with the initial slow-growth phase of the S-curve. Conversely, exploratory innovation, emphasizing radical changes and new technologies, corresponds to the rapid acceleration phase [34]. Importantly, both types of innovation exhibit a strong degree of interdependence. Exploitative innovation can help build the resources and capabilities necessary for future exploratory endeavors, while successful exploratory innovation can create new avenues for further exploitative refinement [35]. This interconnection further reinforces the alignment with the technological progress described by the S-curve theory, as demonstrated in Figure 2.

Leveraging the insights from the technology S-curve theory, this paper analyzes the dynamic changes in key parameters of the model as a function of a manufacturer’s green innovation input. These key parameters include operating cost, penalty payment, and green-added value. Figure 2 illustrates that, as manufacturers incrementally increase their green innovation input, the green added-value exhibits an S-shaped trajectory, signifying growing environmental benefits. Conversely, the operating cost generated by green transition and the penalty payment incurred by environmental regulation decrease in an inverted S-curve fashion. This signifies diminishing costs associated with both operational efficiency and environmental compliance as innovations improve. It is noteworthy that the “take-off point”, also known as the tipping point on the S-curve, coincides with a significant technological breakthrough resulting from the accumulated green innovation efforts. This point marks the watershed where green innovation starts yielding substantial competitive advantages and environmental benefits. Additionally, the penalty payment scheme employed here is assumed to be market-based. This means the penalty amount is directly proportional to the volume of discharge produced by the manufacturer. Consequently, a sharp reduction in discharge translates to a sharp decline in penalty payment, while a moderate reduction translates to a moderate decline.

Equations (1)–(3) represent the mathematical expressions for operating cost, penalty payment, and green-added value, respectively. These functions are defined by two key parameters: (1) the y-intercept a, which influences the starting point of each curve. A higher value of a implies a larger room for future improvement through increased green innovation input; (2) the x-intercept b, which influences the location of the inflection point. A higher value of b means that a greater level of green innovation input is needed to achieve a significant change.

$$\psi =a_{\psi }-\sqrt[3]{\psi \left(\rho \right)-b_{\psi }},$$

(1)

$$\pi =a_{\pi }-\sqrt[3]{\pi \left(\rho \right)-b_{\pi }},$$

(2)

$$\delta =a_{\delta }+\sqrt[3]{\delta \left(\rho \right)-b_{\delta }}.$$

(3)

3.3. The Stackelberg Game Framework and Pay-Off Matrix

This study applies the Stackelberg leadership model to analyze the dynamic between local government and manufacturers in the context of green innovation in China [36]. The local government acts as the Stackelberg leader, setting the stage for environmental policies and incentives, thereby shaping the manufacturer’s response. This assumption aligns with the prevalent top-down approach in China, where governmental directives often precede corporate action in areas of significant public interest, particularly environmental sustainability [37]. Both the government and the manufacturer are assumed to be rational actors with pure strategies. This means their actions are consistent and predictable, aimed at maximizing their respective utilities: public welfare for the government and profit for the manufacturer. The rationale behind this assumption is two-fold: firstly, the government’s decisions are influenced by considerations of political legitimacy and public trust, and secondly, the producer’s strategies are driven by financial implications and the cost–benefit analysis of green innovation. Therefore, neither the local government nor the manufacturer will alter their chosen strategies without substantial consideration.

In the specified Stackelberg model, the strategy space for the local government is defined as (SR, LR), where (SR) denotes strict environmental regulation and (LR) represents lax regulation. Conversely, the manufacturer’s strategy space is denoted as (PI, CI), with (PI) indicating proactive green innovation and (CI) signifying cautious green innovation.

The local government, when selecting its strategy, must consider a balance among several critical factors: tax revenue from sales, potential income from environmental penalties, and the political and social recognition associated with environmental stewardship. The dilemma here is that while strict regulations may increase environmental penalty revenues and enhance the government’s environmental image, they could simultaneously deter manufacturers due to increased compliance costs, potentially leading to a reduction in tax revenue as a result of decreased sales or even relocation of manufacturing facilities.

On the other hand, manufacturers face their own set of trade-offs when determining their strategic approach. The decision to invest in proactive green innovation may lead to higher operating costs due to the adoption of advanced technologies and processes. However, this can be offset by the added value derived from green innovation and a reduction in environmental penalties. Conversely, a cautious approach may minimize immediate operating costs but could result in higher penalties and a loss of competitive advantage in the longer term.

To illustrate these strategic interactions and their associated pay-offs, we present a decision tree and a corresponding pay-off matrix in Figure 3 and

Table 2. Pay-off matrix for both players in the Stackelberg game.

Local Government	Manufacturer
	Cautious innovation	Proactive innovation
Strict	$(1-\tau )(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}-{\pi }_i{Q}_{c(SR,CI)}$	$(1-\tau )(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}-\widehat{{\pi }_i}{Q}_{c(SR,PI)}$
regulation	$\tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}+{\sum }_{i=1}^n{\pi }_i{Q}_{c(SR,CI)}+{\eta }_{c(SR,CI)}$	$\tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}+{\sum }_{i=1}^n\widehat{{\pi }_i}{Q}_{c(SR,PI)}+{\eta }_{c(SR,PI)}$
Lax	$(1-\tau )(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}$	$(1-\tau )(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}$
regulation	$\tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}+{\eta }_{c(LR,CI)}$	$\tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}+{\eta }_{c(LR,PI)}$

, respectively. These visual aids will provide a clearer understanding of the potential outcomes for each combination of strategies and help elucidate the strategic considerations that both the government and manufacturers must weigh.

4. Model Analysis

4.1. Rationale behind a Manufacturer’s Strategy Choice and Sensitivity Analysis

In the Stackelberg model’s pay-off matrix H, detailed in Equation (4), the manufacturer’s pay-offs under different regulatory and innovation strategies are delineated. The first row of H corresponds to the manufacturer’s pay-offs under SR, while the second row pertains to pay-offs under LR. Similarly, the first column represents the pay-offs when the manufacturer adopts CI, and the second column corresponds to PI.

$$\mathit{H}=\left(\begin{array}{cc}(1-\tau )(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}-{\pi }_i{Q}_{c(SR,CI)}& (1-\tau )(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}-\widehat{{\pi }_i}{Q}_{c(SR,PI)}\\ (1-\tau )(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}& (1-\tau )(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}\end{array}\right).$$

(4)

The manufacturer’s strategic choice is determined by comparing the horizontal pay-offs within matrix H. Specifically, if the condition outlined in Equation (5) is satisfied, the manufacturer will opt for proactive innovation under strict regulation. Conversely, if Equation (6) holds, proactive innovation will also be chosen under lax regulation.

$$(1-\tau )[(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}-(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}]+[{\pi }_i{Q}_{c(SR,CI)}-\widehat{{\pi }_i}{Q}_{c(SR,PI)}]>0,$$

(5)

$$(1-\tau )[(P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}-(P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}]>0.$$

(6)

For the sensitivity analysis, we categorize the model into three distinct scenarios based on the sales quantities resulting from different innovation strategies:

In Scenario 1, under both SR and LR, manufacturers adopting a proactive innovation strategy (denoted as $Q_{c(SR,PI)}$ and $Q_{c(LR,PI)}$ respectively) experience lower sales quantities compared to a cautious approach ($Q_{c(SR,CI)}$ and $Q_{c(LR,CI)}$). This aligns with the introduction stage of innovation diffusion theory, where initial adoption of new technologies is slow due to market uncertainties [38]. However, manufacturers’ decisions can still hinge on the cost–benefit trade-off despite lower sales. There following two key cases must be considered:

In the first case, when $(P-{\psi }_2-{\rho }_2+{\delta }_2)<(P-{\psi }_1-\rho 1+\delta 1)$, proactive innovation’s net benefit (profitability) may be lower than cautious innovation’s. This occurs when the per-unit profit (considering the base price, operating costs, research costs, and added value from innovation) under proactive innovation is less than that of cautious innovation. However, manufacturers might still choose proactive innovation if the potential savings from reduced environmental penalties outweigh the initial economic disadvantage.

Here, exceeding the “take-off point” of the technology’s S-curve becomes crucial. This point signifies a significant improvement in the technology, leading to a substantial decrease in waste discharge and, consequently, environmental penalty payments (as assumed in Figure 2). Therefore, to incentivize proactive innovation, it is vital to help manufacturers surpass this “take-off point”. This can be achieved not only through environmental regulations but also by providing public research and development infrastructure. Such infrastructure would make private research investments by manufacturers more worthwhile and efficient in achieving significant technological breakthroughs.

In the second case, when $(P-{\psi }_2-{\rho }_2+{\delta }_2)\ge (P-{\psi }_1-\rho 1+\delta 1)$, the net benefit from proactive innovation outweighs that of cautious innovation. This means the per-unit profit under proactive innovation (considering base price, operating costs, research costs, and added value from innovation) is greater than the net profit under cautious innovation. Consequently, environmental penalties become less relevant. The market itself rewards manufacturers for their innovative products, driving them to choose proactive innovation for the inherent economic advantage it offers. This case is particularly likely to occur when both operating cost reductions and the added value from innovation have surpassed the “take-off point” on the technology’s S-curve.

In Scenario 2, under both SR and LR, manufacturers adopting a proactive innovation strategy (denoted as $Q_{c(SR,PI)}$ and $Q_{c(LR,PI)}$ respectively) experience higher sales quantities compared to those adopting a cautious approach ($Q_{c(SR,CI)}$ and $Q_{c(LR,CI)}$). This trend is consistent with the maturity stage of the innovation diffusion theory, where the new technology becomes widely adopted by the mainstream market [38]. Consequently, customers become more receptive to the benefits of the innovation, leading to increased sales for manufacturers employing proactive strategies. Nevertheless, despite the higher sales quantity, manufacturers’ decisions still hinge on a cost–benefit analysis. The following two key cases must be considered:

In the first case, when $(P-{\psi }_2-{\rho }_2+{\delta }_2)<(P-{\psi }_1-\rho 1+\delta 1)$, proactive innovation may offer a lower net benefit (profitability) compared to cautious innovation. This can be attributed to the high operating costs and research investments associated with the innovation. The entire industry for the new innovative product may be in the throes of rapid technological advancements, and fierce competition could limit the added value manufacturers gain from their innovations.

Although the lower net benefit can be a deterrent to the choice of proactive innovation, it is not prohibitive as long as the sales advantage of proactive over cautious innovation is substantial. Moreover, the potential savings from reduced environmental penalties remain a significant incentive for manufacturers to consider proactive innovation. The pivotal factor for manufacturers is to reach a level of technological advancement that allows them to surpass the ‘take-off point’ on the technology’s S-curve, leading to significant operational efficiencies and cost savings, particularly in terms of environmental penalties.

In the second case, when $(E-{\psi }_2-{\rho }_2+{\delta }_2)\ge (E-{\psi }_1-\rho 1+\delta 1)$, the profitability of proactive innovation exceeds that of cautious innovation. This shift in net benefit diminishes the role of environmental penalties as a motivator for manufacturers to adopt proactive strategies. This situation is likely to occur when the cumulative benefits of operating cost reductions and the added value from innovation reach a critical threshold, known as the ‘take-off point’ on the technology’s S-curve. Surpassing this point typically indicates that the technology has matured enough to offer substantial cost advantages and market differentiation. This stage signifies a notable evolution in the industry, where proactive innovation becomes integral to business success and is propelled by market forces rather than regulatory pressures.

In Scenario 3, under both SR and LR, manufacturers adopting a proactive innovation strategy (denoted as $Q_{c(SR,PI)}$ and $Q_{c(LR,PI)}$, respectively) experience sales quantities equivalent to those adopting a cautious approach ($Q_{c(SR,CI)}$ and $Q_{c(LR,CI)}$). This situation aligns with the growth stage of the innovation diffusion theory and marks a critical juncture for proactive innovation, where the new innovative product has just reached a significant level of market penetration. This can be attributed to the adoption of the innovation by the early majority, meaning that acceptance is growing considerably, even though the late majority and laggards are still hesitant [38].

With this in mind, inequalities 7 and 8 can be derived from inequalities 5 and 6, which equate the sales quantities under the two innovation strategies (proactive and cautious) with $\overline{Q_c}$. In this context, a clear pattern emerges for the manufacturer among gains and losses in operating costs, research investment, added value, and environmental penalties under different innovation strategies. For instance, a substantial reduction in operating costs and environmental penalties, or a significant increase in added value due to proactive innovation, would incentivize manufacturers to choose a proactive strategy. Conversely, a substantial increase in investment costs would hinder proactive innovation. Thus, it is imperative for manufacturers to consider all factors and weigh the pros and cons before making the final innovation decision. In other words, even though a proactive innovation strategy can help reduce operational costs and environmental penalties while increasing added value, it is not justifiable for manufacturers to choose the proactive strategy if these advantages do not sufficiently offset the research investment. Therefore, in this scenario, strict environmental regulation with environmental penalties may still provide an additional incentive to promote proactive innovation.

$$\{(1-\tau )[({\psi }_1-{\psi }_2)+({\rho }_1-{\rho }_2)+({\delta }_2-{\delta }_1)]+({\pi }_i-\widehat{{\pi }_i})\}\overline{Q_c}>0,$$

(7)

$$\left\{(1-\tau )[({\psi }_1-{\psi }_2)+({\rho }_1-{\rho }_2)+({\delta }_2-{\delta }_1)]\right)\}\overline{Q_c}>0.$$

(8)

4.2. Rationale behind the Local Government’s Strategy Choice and Sensitivity Analysis

We can also construct matrix G, as detailed in Equation (9), to represent the local government’s pay-offs under different regulatory and innovation strategies. The first column of G corresponds to the local government’s pay-offs when the manufacturer adopts a cautious innovation strategy, while the second column reflects pay-offs associated with a proactive innovation strategy. Similarly, the first row represents the pay-offs when the local government imposes strict regulation, and the second row indicates the pay-off under lax regulation.

Given that the local government is the leader in the Stackelberg model, a simple vertical comparison of pay-offs within matrix G is insufficient. Instead, we must consider four possible outcomes, as the local government’s choice of strategy will prompt the manufacturer to respond with different strategic choices. This interplay results in four distinct scenarios:

$$\mathit{G}=\left(\begin{array}{cc}\tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}+\sum _{i=1}^n{\pi }_i{Q}_{c(SR,CI)}+{\eta }_{c(SR,CI)}& \tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}+\sum _{i=1}^n\widehat{{\pi }_i}{Q}_{c(SR,PI)}+{\eta }_{c(SR,PI)}\\ \tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}+{\eta }_{c(LR,CI)}& \tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}+{\eta }_{c(LR,PI)}\end{array}\right).$$

(9)

In Scenario 1, regardless of the local government’s regulatory strategy, whether SR or LR, the manufacturer is likely to opt for CI after evaluating the potential gains and losses. This situation aligns with the introduction stage of the innovation diffusion theory, where a new innovative product typically has limited market recognition and acceptance [38]. Furthermore, production inefficiencies at this stage, such as high operating costs, discourage manufacturers from adopting PI. Should this scenario reflect reality, environmental regulations in the form of penalties may not effectively promote green innovation. Instead, they might serve merely to increase the government’s non-tax revenue. Moreover, excessive environmental penalties could even lead manufacturers to relocate to areas with less stringent regulations to avoid such financial burdens.

In Scenario 2, conversely, the manufacturer would invariably choose PI, regardless of the local government’s regulatory strategy, be it SR or LR. This situation closely mirrors the maturity stage of the innovation diffusion theory, where the new innovative product has achieved considerable market recognition and production efficiency has improved sufficiently to reduce operating costs [38]. As a result, PI becomes the manufacturer’s dominant choice over CI, furthering the manufacturer’s competitive edge. Meanwhile, there is little justification for the local government to impose SR, especially under the pretext of promoting green innovation, as it would likely impose an unnecessary burden on manufacturers rather than incentivize them to engage in activities they would likely undertake anyway.

In Scenario 3, the manufacturer’s choice of innovation strategy is congruent with the local government’s environmental policy goals. Manufacturers typically adopt CI under LR. However, they shift to PI in response to SR enacted by the local government. This transition highlights the pivotal role of rigorous environmental regulations in stimulating green innovation within the manufacturing sector. The scenario mirrors the growth phase of the innovation diffusion theory, wherein an innovative product, having achieved initial market acceptance and momentum, necessitates further encouragement to surpass traditional products in preference [38]. Consequently, it is imperative for local governments to implement SR over LR to reinforce manufacturers’ dedication to a green transition.

Furthermore, for SR to become the dominant strategy for the local government in this scenario, inequality 10 must hold. Assuming a constant sales tax rate, the sales quantity under SR and PI should not significantly trail the quantity under LR and CI. Moreover, the net benefit (profit) from the SR–PI combination should exceed that from the LR–CI combination. Additionally, sufficient social and political recognition should incentivize the local government to implement SR, supplemented by revenue from environmental penalties.

$$\tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(SR,PI)}+\sum _{i=1}^n\widehat{{\pi }_i}{Q}_{c(SR,PI)}+{\eta }_{c(SR,PI)}>\tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(LR,CI)}+{\eta }_{c(LR,CI)}.$$

(10)

In Scenario 4, a regulatory paradox emerges where environmental policies produce counter-intuitive effects on manufacturers’ innovation strategies. Under LR, manufacturers voluntarily engage in PI. Conversely, introducing SR leads manufacturers to adopt CI. This development is antithetical to the policy’s objective, as it impedes, rather than promotes, green innovation. While this scenario shares similarities with Scenario 3’s alignment with the growth stage of the innovation diffusion theory [38], it also presents distinct differences. Here, the innovative product has overtaken traditional alternatives but retains a fragile competitive edge. A minor policy shift, such as the imposition of an additional environmental penalty, could jeopardize this position. Thus, in this scenario, environmental regulation not only does not facilitate green innovation but may also inadvertently weaken the manufacturer’s impetus for ecological transition.

In this stage, it is particularly concerning that local governments tend to maintain SR as in Scenario 3, under the assumption that it will continue to incentivize green innovation. This persistence often arises from an insufficient awareness of situational changes and a lack of agility in policy modification. Inequality 11 delineates the difficulties local governments encounter when considering policy shifts. In the inequality, the manufacturer’s net benefit (profit) and sales quantity associated with the SR–CI strategy combination typically exceed those of the LR–PI strategy combination. However, the local government’s pursuit of social and political recognition for appearing tough on environmental issues, along with the income from environmental penalties, encourages the perpetuation of SR, which may inadvertently impede green innovation.

$$\tau (P-{\psi }_1-{\rho }_1+{\delta }_1)Q_{c(SR,CI)}+\sum _{i=1}^n{\pi }_i{Q}_{c(SR,CI)}+{\eta }_{c(SR,CI)}>\tau (P-{\psi }_2-{\rho }_2+{\delta }_2)Q_{c(LR,PI)}+{\eta }_{c(LR,PI)}.$$

(11)

In summation, an analysis of the four scenarios reveals that environmental regulation does not uniformly enhance green innovation. For instance, in Scenarios 1 and 2, the manufacturer’s innovation strategy remains indifferent to the local government’s regulatory approach. Conversely, Scenario 4 demonstrates that environmental regulation can counteract its intended purpose. While Scenario 3 shows a positive correlation between strict regulation and proactive green innovation, it underscores the need for policy design to evolve alongside the manufacturer’s progression through the innovation diffusion stages. To effectively leverage environmental regulation as a catalyst for green innovation, local governments must consider the manufacturer’s current stage within the innovation diffusion spectrum. Additionally, the practical outcomes of environmental regulation policies become more complex when stakeholder responses are taken into account, which will be discussed in subsequent sections.

4.3. Incorporating Stakeholder Dynamics into the Stackelberg Model

Stakeholder theory offers a valuable lens for examining strategic interactions among entities within game theory. It elucidates how the attributes and actions of stakeholders can shape the pay-offs and strategies of players, thereby influencing their behavior. For instance, research employing a stakeholder perspective has demonstrated that a lead firm’s adoption of green, eco-friendly practices can compel suppliers in the supply chain to undertake a green transformation [14,15]. This suggests that the impetus for a manufacturer to engage in green innovation arises not solely from governmental pressure but also from other relevant stakeholders. Building on this premise, we can infer that various stakeholders may significantly influence both local government and manufacturers in their environmental decision-making. Similar to the distributed cooperative energy management approach used in multienergy systems [39], our Stackelberg model is extended to examine the effects of three principal stakeholders on the strategic behaviors of local governments and manufacturers. The interplay and dynamics among the key players are depicted in Figure 4.

Stakeholder 1: Central Government. In China, the central government plays a pivotal role in shaping and monitoring the environmental policies implemented by local governments through mechanisms such as the centralization of authority in controlling the appointment of officials and applying political pressure. Additionally, the Chinese central government is keener to set ambitious environmental goals, like reducing greenhouse gas emissions or increasing renewable energy usage, to enhance its public image on the global stage and fulfil its international obligations. As a result, local governments often find themselves compelled to adopt stricter environmental regulations, which can come at the expense of the local economy and may conflict with local interests. For instance, in 2020, the Chinese central government launched initiatives such as the ‘dual carbon targets’, aiming for a carbon peak before 2030 and carbon neutrality before 2060, and ‘carbon peak and carbon neutrality’. These initiatives exerted significant pressure on local governments to enforce stringent environmental regulations on manufacturers, leading to drastic measures like electricity outages that significantly disrupted the normal operations of manufacturers [40].

In our Stackelberg Model, the variable $\eta$ represents the social and political recognition accorded to the local government. Specifically, ceteris paribus, the value of $\eta$ is higher when the manufacturer opts for PI or when the local government enforces SR due to the associated positive appearance and environmental improvement. However, the central government, as a pivotal stakeholder in environmental governance, can wield even more influence over $\eta$ by rewarding local governments that adhere to its environmental mandates with SR and penalizing those that deviate from LR. For instance, irrespective of the manufacturer’s strategy, a pro-environmental central government can markedly raise the value of $\eta$ associated with any (*-SR) strategy combination and reduce it for any (*-LR) strategy combination, where * denotes the strategy (either PI or CI) chosen by the manufacturer. By increasing the disparity in $\eta$ values between (*-SR) and (*-LR) strategy combinations, the central government can effectively incentivize local governments to prioritize SR, thus eclipsing other economic factors detailed in Table 1.

Stakeholder 2: Upstream Suppliers. Unlike the environmental demands placed by downstream buyers [14,15], Chinese manufacturers often grapple with cost pressures originating from their suppliers. A case in point is the new energy vehicle sector in China, which has recently suffered from significant supply shortages. These shortages stem from bottlenecks in the delivery of essential components and raw materials, notably semiconductors and lithium-ion batteries. The crux of the issue lies in the green supply chain’s lack of holistic development, curtailing its potential. Specifically, the chain’s inadequate infrastructure and emergent technology foster persistent vulnerabilities, leading to bottlenecks that undermine both efficiency and sustainability. Moreover, these challenges are magnified by the swift expansion of downstream assembly and integration manufacturing, which escalates the demand for upstream components and materials. This supply–demand mismatch inevitably leads to a sharp cost escalation, thereby stalling the progress of the nascent green industry [41].

In our Stackelberg game model, the variable $\psi$ represents the manufacturer’s operating costs, which vary between PI and CI. These costs typically decrease following an S-curve pattern as investment in research intensifies. Moreover, the value of $\psi$ is considerably affected by stakeholders, especially upstream suppliers dealing with essential raw materials and components. The impact is particularly pronounced for manufacturers engaged in PI, as this type of innovation often requires the establishment of a new supply chain, which may be in its infancy and underdeveloped. Therefore, it is not uncommon for manufacturers introducing drastically innovative products to experience a surge in $\psi$, attributable to upstream supply strain. This surge can sometimes be substantial enough to erode profits and compel a shift from a PI to CI, thereby decelerating the innovation process.

Stakeholder 3: Operations Team. In addition to external ones, internal stakeholders significantly influence the efficacy of a Chinese manufacturer’s innovation strategy. For instance, when a manufacturer invests in new equipment and establishes a new production line for innovative products, the operations team, being directly impacted by these changes, may exhibit considerable resistance to the manufacturer’s aggressive innovation strategy, potentially leading to operational disruptions, such as production delays or quality issues. This resistance often stems from the team’s unfamiliarity with new workflows and the daunting prospect of adapting to rapid and radical changes. Such a transition may provoke concerns over job security and create a sense of being overwhelmed by the necessity to quickly acquire new skills, fostering uncertainty and distrust. If not properly managed, this environment will hinder the manufacturer’s ability to successfully implement its innovation strategy [42].

In our Stackelberg game model, the operations team’s impact is also captured by the variable $\psi$, as any operational changes would affect operating costs. In particular, $\psi$ is typically higher when a manufacturer opts for PI over CI. This is because PI often entails significant adjustments to operational procedures and practices, leading to operational inefficiencies and a consequent rise in $\psi$. Therefore, an increase in research investment does not guarantee a decrease in $\psi$. In fact, the value of $\psi$ may even surge if the operations team, a key stakeholder, is not effectively engaged and becomes resistant to the innovation process. In extreme cases, a significant profit downturn caused by a spike in $\psi$ could compel the manufacturer to abandon PI for CI, jeopardizing its green transition strategy.

In summary, the extension of the Stackelberg game model to include stakeholders adds layers of complexity to the relationship between local government’s environmental regulations and manufacturers’ green innovation efforts. Stakeholders such as the central government, upstream suppliers, and the operations team can exert significant influence on the game’s outcomes by adjusting the pay-off matrix of the players. This relationship sheds light on cases where environmental regulations can be driven by political rather than economic causes and may inadvertently hinder the manufacturer’s transition to innovative processes. Specifically, the model highlights how local governments, under political pressure from the central government, may aggressively implement environmental regulations without due consideration of economic repercussions or their own financial interests. Furthermore, manufacturers that are inadequately prepared for a green transition may encounter severe supply chain bottlenecks and internal operational resistance, leading to stalled innovation. The model thus serves as a cautionary tale, illustrating the potential pitfalls of misaligned incentives and the critical need for a holistic approach that aligns the interests of all stakeholders to ensure a smooth transition to green practices.

5. Case Study Application and Model Verification

To strengthen the empirical foundation of our research, we complement the theoretical framework with a case study application centered on the burgeoning Chinese NEV industry. This methodological approach directly engages with the stakeholder-inclusive extension of our Stackelberg game model and bridges the theoretical constructs with real-world scenarios. Specifically, the case studies elucidate the diverse influences exerted by stakeholders, ranging from central government mandates to upstream supplier collaboration and internal operations team dynamics, on the efficacy of environmental regulations in spurring green innovation. Through a meticulous dissection of these scenarios, we aim to achieve two key objectives: firstly, to validate the robustness of our model in a practical context, and secondly, to deepen our comprehension of the nuanced relationship between environmental regulation and green innovation.

5.1. The Central Government’s Role in NEV Transition

In 2009, the Chinese central government initiated a transformative journey in the automotive sector by issuing the “Automotive Industry Adjustment and Revitalization Plan”. This strategic initiative was aimed not only at combating global climate change but also at revolutionizing the domestic auto industry. It signified the beginning of a comprehensive subsidy program for the NEV industry, which garnered enthusiastic support from local governments. The implementation of this program catalyzed a significant surge in the NEV market, as depicted in Figure 5.

Nevertheless, the subsidy program had inadvertently precipitated certain unintended consequences, as illustrated in

Table 3. Fraudulent green innovation resulting from environmental regulation in 2015.

Fraudulent Behavior Type	Number of Vehicles	Fraudulent Subsidy Amount (10,000 RMB)	Percentage of Total Vehicles	Percentage of Total Subsidy
Registered Vehicles with No Car	3547	101,021	1.1%	3.0%
Vehicles with Car but No Battery	19,158	187,510	5.8%	5.6%
Idle Vehicles by Affiliates	30,414	361,857	9.2%	10.8%
Delivered but Unused Vehicles	21,362	254,159	6.5%	7.6%
Others	1893	22,522	0.6%	0.7%
Total	76,374	927,070	23.1%	27.7%

Sourced from the China Association of Automobile Manufacturers (http://www.caam.org.cn/ (accessed on 8 July 2024)).

. Specifically, a number of auto manufacturers, enticed by the prospect of government subsidies, commenced production of NEVs that were deficient in terms of battery energy density and driving range, with their efforts primarily directed at maximizing financial gains from the subsidies. This approach engendered an oversupply of NEVs that were misaligned with market demands. In an effort to rectify this, the central government incrementally heightened the eligibility criteria for subsidy acquisition and embarked on a gradual withdrawal of the subsidy program. Data from the Chinese Ministry of Industry and Information Technology indicate a discernible annual reduction in the proportion of subsidized NEVs, which declined from 78% in 2017 to 47% in 2021. Simultaneously, there was a notable contraction in the average subsidy allocated per NEV, diminishing from 67,300 RMB in 2017 to 19,800 RMB in 2021 [43,44]. For context, the average monthly income of Chinese urban citizens working in the private sector in 2017 was approximately RMB 4000.

$$\mathrm{CAFC}\mathrm{Credit}=(\mathrm{Average}\mathrm{Fuel}\mathrm{Consumption}\mathrm{Target}\mathrm{Value}-\mathrm{Average}\mathrm{Actual}\mathrm{Value})\times \mathrm{Annual}\mathrm{CV}\mathrm{Quantity},$$

(12)

$$\mathrm{NEV}\mathrm{Credit}=\mathrm{NEV}\mathrm{Model}\mathrm{Score}\times \mathrm{NEV}\mathrm{Quantity}-\mathrm{CV}\mathrm{Quantity}\times \mathrm{NEV}\mathrm{Credit}\mathrm{Ratio}\mathrm{Requirement}.$$

(13)

In 2018, the central government introduced the “Dual Credit Policy” (DCP), specifically designed to curb fuel consumption and promote qualitative growth within the NEV sector. The DCP stipulates two distinct types of credits for auto manufacturers: corporate average fuel consumption (CAFC) credit and NEV credit. CAFC credit, as delineated by Equation (12), quantifies the average fuel consumption level of a manufacturer’s fleet and compares it against a benchmark established by the government. NEV credit, explicated in Equation (13), translates a manufacturer’s actual NEV production and sales figures into credits, taking into account factors such as electric range, energy efficiency, and vehicle type, including hybrid vehicles, pure electric vehicles, and fuel cell vehicles, relative to another government-mandated target.

Under the stipulated framework of the Dual Credit Policy, manufacturers failing to meet the mandated CAFC or NEV credit requirements are subject to a range of punitive measures. These include restrictions on the production and sale of non-compliant vehicle models, fines proportional to the credit deficit, and the mandatory purchase of NEV credits from entities with a surplus. Additionally, the government retains the discretion to modify credit requirements, thereby promoting ongoing technological advancement. Data published annually by the Chinese Ministry of Industry and Information Technology (MIIT) reveal that the DCP has significantly influenced the automotive industry, leading to a marked reduction in the average fuel consumption of passenger cars produced in China. Consumption rates declined from 5.74 L per 100 km in 2018 to 3.99 L per 100 km in 2022, representing an average annual decrease of 8.69%. Concurrently, the allocation of positive credits for new energy passenger cars by domestic manufacturers increased from 3.9374 million in 2018 to 15.2278 million in 2022, indicating an average annual growth rate of 40.24% [45,46], as illustrated in

Table 4. Impact of stakeholders on green innovation performance.

Stakeholder	Action	Results
Central Government	Dual Credit Policy	Average fuel consumption rate decreased by 8.69% annually: – 5.74 L/100 km in 2018 – 3.99 L/100 km in 2022 Positive credits for NEV manufacturers increased by 40.24% annually: – 3.9374 million in 2018 – 15.2278 million in 2022
Upstream Suppliers	Semiconductor and lithium-ion battery supply strain	Semiconductor shortage: – Reduced automotive production by 5–8% in early 2021 – Delivery time extended from 3–4 months to >6 months Lithium-ion battery shortage: – Battery raw material price increased from 37,000 RMB/ton to 99,000 RMB/ton in 2021 Impact on sales and profitability of assembled vehicles: – BYD vehicle retail price increased by 4000–13,000 RMB in Q1 2022 – BYD net profit reduced by 57.53% in 2021
Internal Operations Team	Human resources inadequacy	Higher salary offers in NEV manufacturers: – 30% to 50% higher than CFV manufacturers in 2021–2022 Higher turnover rate in NEV manufacturers: – 20.2% compared to CFV manufacturers’ 15.9% in 2022

Source from MIIT (http://www.miit.gov.cn (accessed on 16 April 2024)); the China Association of Automobile Manufacturers (http://www.caam.org.cn (accessed on 16 April 2024)); the Institute for Energy Research (https://www.instituteforenergyresearch.org (accessed on 16 April 2024)).

.

In this case study, rather than imposing an environmental penalty payment, denoted as $\psi$ in our model, subsidies are provided to auto manufacturers. These subsidies act as a “negative” penalty payment, effectively incentivizing manufacturers to engage in PI. This dynamic corroborates our model’s predictions, illustrating that local governments can indeed motivate manufacturers to escalate their green innovation efforts through environmental regulation. Nonetheless, due to information asymmetry, this incentive mechanism may inadvertently foster green-washing, where manufacturers exploit the subsidy system rather than commit to authentic innovation. To counteract this, the central government has implemented the DCP, which addresses the issue of information asymmetry by requiring direct and quantifiable reporting to the central government, thus enhancing the transparency and efficacy of local governments’ environmental efforts. In alignment with the game model, the recognition variable, influenced partially by the central government and denoted as $\eta$, becomes a more precise measure of local government environmental efforts, encouraging the adoption of SR as their strategic preference. This case study underscores the central government’s role as a critical stakeholder in exerting influence over local governments, ensuring their stringent adherence to environmental regulations, and promoting manufacturers’ green innovation.

5.2. The Upstream Supplier’s Impact on NEV Development

Policy incentives implemented by both central and local governments have catalyzed many manufacturers to expedite their green innovation initiatives. Consequently, there has been a notable shift in production from CFVs to an increasing number of NEVs. In 2023, China’s NEV production and sales soared remarkably to 9.587 million and 9.495 million units, respectively. This surge had established China as the global frontrunner in the NEV market for nine consecutive years, commanding a domestic market share of 31.6% [47]. Considering that 77.8% of China’s NEVs in 2023 were electric vehicles [48], this rapid ascension has also exerted a considerable strain on upstream supplies, particularly in critical areas such as semiconductors and lithium-ion batteries.

Semiconductors are indispensable for electric vehicle (EV) systems ranging from powertrains to advanced driver-assistance systems. The demand for these components is substantial, with each EV requiring nearly 2000 chips, almost three times more than its fuel-powered counterpart [49]. However, owing to the low rate of domestic self-sufficiency in semiconductor production in China, this heightened demand has precipitated a significant increase in global automotive chip shipments, which have nearly tripled over the past decade. Particularly during the years of the COVID-19 pandemic, a global shortfall in automotive chips has further exacerbated supply chain disruptions, leading to pronounced production and logistical challenges [50]. For instance, in early 2021, the semiconductor shortage reduced Chinese automotive production by 5% to 8% while extending MCU delivery times from 3–4 months to over 6 months, with some delays reaching 9 months or more [51].

China, which possesses only 8 percent of the world’s lithium reserves and relies on imports for two-thirds of lithium raw materials [52], finds its lithium-ion battery supply chain under significant pressure due to international market dynamics, geopolitical tensions, and environmental concerns. As indicated in Figure 6, China’s lithium battery shipments experienced a substantial surge, escalating from 131.6 GWh in 2019 to 887.4 GWh in 2023, with an average annual compound growth rate of 61.14%. This increased demand for lithium-ion batteries, driven by the rapid expansion of China’s NEV market, had, in turn, precipitated a spike in the cost of lithium battery raw materials. For instance, the price of lithium–iron phosphate, a key material used in lithium-ion batteries, dramatically increased by 167.57% from CNY 37,000/ton in January to CNY 99,000/ton in December 2021 [53]. The substantial cost spike compelled major NEV manufacturers such as BYD to implement two price hikes in the first quarter of 2022: the initial increase ranged from RMB 1000–7000, followed by a second increase of RMB 3000–6000, culminating in a net profit reduction of 57.53% [54], as illustrated in Table 4.

Mitigating the adverse effects of critical upstream supply disruptions, leading firms in China’s NEV market have engaged in pre-emptive measures, such as supply chain collaboration and vertical integration. The former entails establishing a stable supply chain for critical raw materials through strategic alliances with upstream suppliers. In the context of China’s lithium carbonate and lithium hydroxide refining market, long-term contractual agreements are prevalent, constituting roughly 80% of total transactions. Notably, this percentage escalates to approximately 90% among battery manufacturing facilities that acquire these critical lithium-based raw materials [55]. The adoption of long-term contracts has demonstrated efficacy in stabilizing the supply chain and securing price consistency for raw materials in the lithium-ion battery sector, even amid market volatility.

The latter entails integrating upstream operations to secure control over indispensable components, thereby mitigating vulnerabilities within the supply chain. For example, BYD’s foray into semiconductor manufacturing began in 2004, with significant advancements made in 2018 and 2019 through the release of 8-bit and 32-bit automotive-grade microcontrollers (MCUs). This initiative towards vertical integration has not only consolidated BYD’s market presence amid the semiconductor shortage but also culminated in significant achievements: by the end of 2020, more than 5 million units of BYD’s automotive-grade MCUs were operational within vehicles produced by BYD and other automotive firms, including XPeng Motors [56,57].

In this case study, we examine the impact of upstream suppliers on manufacturers’ innovation strategies. The manufacturer’s operating cost, denoted by $\psi$, is represented by a decreasing S-curve in relation to research investment $\rho$, as illustrated in Equation (1). Additionally, the cost can be further influenced by upstream suppliers through the baseline parameter $a_{\psi }$. For example, supply constraints or shortages can lead to an increase in an $a_{\psi }$, thereby raising operating costs and negating the benefits of research investment. Such scenarios, commonly observed during supply chain disruptions, underscore the vulnerability of manufacturers’ cost structures to supply shocks and the significant influence of upstream suppliers on manufacturers’ strategic decision-making. As predicted by our model, increased operating costs are likely to diminish manufacturers’ profits, thereby prompting the manufacturer to strategically pivot from PI to CI. This strategic shift is particularly evident when CFVs become more appealing to manufacturers than NEVs, as the former benefit from a more established and reliable supply chain. Furthermore, this case study also sheds light on how concerns regarding upstream stakeholders can act as a driving force for supply chain collaboration and vertical integration, thereby securing the advancement of green innovations.

5.3. The Operations Team’s Influence on NEV Production

Beyond external factors such as environmental regulations and supply chain dynamics, internal elements within an organization can also exert significant influences on a manufacturer’s innovation strategy. Notably, organizational inertia, characterized by the tendency to adhere to established trajectories and resist change, plays a pivotal role in this context [58]. This inertia often originates from the operations team, a crucial internal stakeholder tasked with managing supply chains, production processes, and logistics. Their resistance to change can entrench established routines and hinder the innovation process [59], which elucidates the reluctance of traditional manufacturers, like Volkswagen and Toyota, to fully embrace NEVs. For instance, in 2023, electric cars were estimated to constitute 18% of global car sales [60], yet Volkswagen and Toyota’s electric vehicle sales comprised merely 8.3% and 1% of their respective total sales [61,62]. These figures partially reflect the significant impact that the operations team, as an internal stakeholder, has on strategic decision-making.

Specifically, operations teams are encountering significant challenges during the manufacturer’s transition to NEVs, which require not only comprehensive operational re-engineering but also the acquisition of new skills and knowledge. NEVs, particularly EVs, are distinguished by their electric motors, batteries, and electronic control systems, and they place a greater emphasis on advanced technologies such as autonomous driving and vehicle-to-everything (V2X) connectivity. These demanding requirements and complexities necessitate that NEV manufacturers offer salaries typically 30% higher than those in the CFV sector to attract skilled operational professionals. Despite these financial incentives, NEV manufacturers in 2022 still experienced a turnover rate as high as 20.2%, in contrast to 15.9% in the CFV sector [63], as illustrated in Table 4. This disparity once again highlights the challenges faced by operations team members in adapting to the rapidly changing workflows and evolving job content associated with NEVs.

The alignment of internal stakeholders, particularly the operations team, is crucial for CFV manufacturers transitioning to green technologies. Without proper integration of the operations team’s expertise and commitment, manufacturers may face significant challenges in pursuing green innovation, especially in the face of stringent environmental regulations. The rationale behind this is that the manufacturer’s capacity for innovation and development is intrinsically linked to the operations team’s adaptability, resilience, and overall disposition. For example, Volkswagen’s inaugural pure electric SUV, the ID.4, has encountered significant quality issues in the Chinese market, characterized by recurring problems such as black screens and connectivity disruptions. Despite attempts at resolution through after-sales service, these issues persist [64], highlighting systemic operational difficulties within Volkswagen China in its NEV transition. Similarly, Chevrolet and Peugeot have markedly delayed the launch of their electric vehicles in China and have experienced a precipitous drop in annual sales, over 80% from their peak in recent years. Ironically, however, Stellantis, the parent company of Chevrolet and Peugeot, has instead invested EUR 1.5 billion to secure a 20% stake in the Chinese NEV startup Leapmotor [65], indicating a shift in confidence towards the NEV startup’s operations team over its own established CFV brands.

In this case study, we examined the critical role of the operations team as an internal stakeholder in shaping a manufacturer’s innovation strategy, particularly focusing on its influence on the operating cost, denoted by $\psi$. The case confirms an inverse relationship between operating costs and the resilience and adaptability of the operations team. For example, a lack of operational capability or engagement in innovation correlates with increased operating costs due to process inefficiencies and post-sale management. Additionally, the operations team’s impact on operating costs is captured by the baseline parameter $a_{\psi }$, as illustrated in Equation (1). Therefore, an underperforming operations team would lead to increased operating costs, which may, in turn, erode the manufacturer’s resolve to pursue PI due to reduced profits. This economic repercussion may further incline the manufacturer towards CI, potentially hindering the progress of green innovation. In sum, by integrating the operations team into the game model, the analysis once again underscores the imperative for a more collaborative stakeholder approach.

6. Conclusions and Policy Implications

6.1. Conclusions

This study employs a Stackelberg game theory model from an innovation–diffusion perspective to investigate the effectiveness of local government environmental regulations in incentivizing NEV manufacturers’ innovation behavior under different scenarios. The model is then extended to incorporate key stakeholders, such as the central government, upstream suppliers, and internal operations teams. Furthermore, the findings of the model are corroborated through case study validation of China’s NEV manufacturers. In summary, while environmental regulations can provide some incentives for NEV manufacturers to engage in proactive innovation, the key drivers include robust technology infrastructure and effective stakeholder engagement. The specific conclusions are as follows and are illustrated in

Table 5. The paradoxical relationship between environmental regulation and green innovation in NEV manufacturing.

Conclusions	Evidences	Questions Answered
During the introduction and growth stage of innovation diffusion, – environmental regulations can be ineffective in promoting green innovation among NEV manufacturers; – the effectiveness of environmental regulations is also influenced by technological maturity and prevailing market conditions.	In 2015, fraudulent practices were identified in China’s emerging NEV industry, particularly concerning environmental regulations in the form of subsidies intended to promote NEVs: – among 330,000 NEVs, 23.1% were found to be fraudulent. – In a total subsidy of RMB 33.435 billion, 27.7% was found to be fraudulent.	Research Question 1 Research Question 2
Stakeholders such as the Central Government can influence environmental regulation through the Dual Credit Policy, – to ensure transparency in NEV manufacturers’ green innovation progress. – to exert precise pressure on NEV manufacturers to advance green innovation.	From 2018 to 2022, significant progress in green innovation were made by NEV manufacturers: – The average fuel consumption rate decreased by 8.69% annually. – Positive credits for NEV manufacturers increased by 40.24% annually.	Research Question 1 Research Question 3
Stakeholders such as upstream suppliers can influence environmental regulation through: – increasing the prices of raw materials and components.	In 2021 and early 2022, significant supply chain issues and challenges were encountered by NEV manufacturers: – NEV production reduced by 5–8% due to semiconductor shortages. – NEV semiconductor delivery times nearly doubled. – NEV battery raw material prices nearly tripled. – NEV retail prices increased by 2.3–7.6%.	Research Question 1 Research Question 3
Stakeholders such as internal operations teams can influence environmental regulation by: – creating human resource inadequacies.	From 2021 to 2022, significant human resource costs and instability were encountered by NEV manufacturers: – NEV manufacturers paid 30–50% higher salaries than their CFV counterparts. – NEV manufacturers experienced a 4.3% higher employee turnover rate than their CFV counterparts.	Research Question 1 Research Question 3

Source from MIIT (http://www.miit.gov.cn (accessed on 16 April 2024)); the China Association of Automobile Manufacturers (http://www.caam.org.cn (accessed on 16 April 2024)); the Institute for Energy Research (https://www.instituteforenergyresearch.org (accessed on 16 April 2024)).

:

• At the introduction stage of innovation diffusion, environmental penalties can somewhat boost proactive innovation. However, helping NEV manufacturers surpass the “take-off point” on the technology’s S-curve is more crucial. Excessive environmental penalties can backfire and even lead to unintended negative consequences.
• During the growth stage of innovation diffusion, environmental penalties are initially imperative to provide NEV manufacturers with sufficient encouragement to maintain their proactive innovation momentum. However, these penalties should gradually fade away once proactive innovation becomes an inherent drive within NEV manufacturers.
• In the maturity stage of innovation diffusion, a proactive innovation strategy can be considered the dominant choice for manufacturers. At this stage, environmental penalties can become unnecessary and may impose additional burdens on manufacturers.

6.2. Policy Implications

Based on the model findings and the insights gained from the case studies, the policy implications for promoting NEV innovation are as follows:

• Enhancing Public R&D and Innovation Platforms: A suite of endeavors, including bolstering public R&D investments and constructing synergistic innovation platforms, is essential to accelerate the maturation and adoption of green technologies. These initiatives will support manufacturers in broader application and further refinement of green technologies.
• Customized Regulatory Frameworks: The development of tailored regulatory frameworks is recommended to incentivize manufacturers to proactively engage in green innovation. These frameworks should be designed to address the specific needs and capabilities of different manufacturers.
• Precise and Adaptive Environmental Policies: The formulation and execution of environmental policies should be timely and contextually aware, incorporating a framework that is both flexible and adaptive. Environmental policies should not serve merely as fiscal instruments for accruing additional government revenue but should consistently focus on promoting NEV innovation.
• Holistic Support for Green Innovation: While environmental regulation is indispensable, it cannot single-handedly drive green innovation. Without thorough preparation and robust support systems, green innovation initiatives may encounter obstacles from stakeholders they rely upon. Therefore, establishing a comprehensive, supportive ecosystem that actively engages stakeholders in a unified, collaborative effort is crucial for fostering green innovation.

6.3. Future Work

Considering the evolutionary nature of the gaming process, future research could benefit from utilizing an evolutionary game theory approach. This would provide insights into how to foster a supportive ecosystem for proactive NEV innovation. Additionally, it is valuable to explore the factors influencing NEV manufacturers’ selection of technological routes, such as hybrid, electric, and fuel-cell electric vehicles. Understanding these factors will help identify the conditions and strategies that encourage manufacturers to adopt and innovate in green technologies.