Promoting the Development of China’s New-Energy Vehicle Industry in the Post-Subsidy Era: A Study Based on the Evolutionary Game Theory Method

Abstract

Government policy constraints and the green credit support of banks have played an indispensable role in promoting the development of the new energy vehicle (NEV) industry. To study the relationship between the government, the banks, and automobile manufacturers in the post-subsidy era and to promote the development of the NEV industry in China, we constructed a tripartite evolutionary game model for the government, the banks, and automobile manufacturers during the subsidy decline, analyzed the evolutionary process of the system, and used MATLAB to simulate the evolutionary stable strategies (ESSs) and the sensitivity of related parameters. The results demonstrate the following: (1) There are five possible evolutionary equilibrium points in the early, middle, and late stages of the NEV industry; (2) with the increase in the phase-out rates and the transaction prices of NEV points, the government is more inclined to low subsidies, the banks are more inclined to implement green credit, and enterprises are more inclined to produce NEVs; (3) there is a threshold for the impact of government incentives on the evolutionary results of the government and the banks, beyond which the evolutionary process of the government and the banks will be unstable; (4) with the increase in financing costs saved by green credit, the government is more inclined to low subsidies and enterprises are more inclined to produce NEVs, while changes in financing costs have less impact on the strategies of banks. According to these findings, the government, the banks, and automobile manufacturers can be relied upon to promote the development of the NEV industry in China.

1. Introduction

The transportation industry has always been a significant carbon emitter, accounting for about 15% of the total carbon emissions in China. In the context of carbon peak and carbon neutrality, it is essential to realize the low-carbon transformation of the transportation industry. In considering the advantages of new-energy vehicles (NEVs) in improving urban climate, reducing carbon emissions, and protecting the environment [1,2], the Chinese government has been subsidizing the promotion and use of NEVs since 2009. In 2015, the production and sales of NEVs in China surpassed those of the United States, with China ranking first in the world. However, government subsidies have also led to problems, including an excessive financial burden for the government, fraudulent corporate subsidies, and weak innovation by NEV manufacturers [3]. Therefore, the government decided to gradually reduce the subsidy for NEVs to ensure the overall stability of the subsidy policy. The decline in subsides not only reduces the government’s financial burden, but also reduces the dependence of NEV manufacturers on government subsidies in a manageable way.

In addition to reducing subsidies, in September 2017, the Chinese government issued a new policy called the New-Energy Vehicle Credit Program and Corporate Average Fuel Consumption Regulation (hereafter referred to as the dual-credit policy) [4]. Unlike the subsidy policy, the dual-credit policy realizes fuel savings and the development of NEVs by stipulating the fuel consumption of fuel vehicles (FVs) and the positive/negative points of the vehicles. In addition, the dual-credit policy shifts the incentive mechanism for producing of NEVs from subsidy-driven incentives to market-driven incentives, thereby shifting the financial burden from the government to enterprises [5]. The dual-credit policy plays a crucial role in developing China’s automobile industry by improving fuel savings and developing NEVs.

The decline of government subsidies increases the debt risk of NEV manufacturers [6]. Therefore, in the post-subsidy era, bank credit support will play an increasingly important role in the NEV industry. The Green Credit Policy is a corporate lending guideline proposed by the International Finance Corporation in 2002, requiring financial institutions to comprehensively assess the environmental impact caused by the financing of projects. This policy prioritizes financing green ecological protection projects and uses economic leverage to promote environmental protection and development [7]. As an effective way to develop green finance, green credit plays an essential role in encouraging cooperation between financial institutions and the government to promote the low-carbon transformation of the automobile manufacturing industry. Therefore, in the post-subsidy era of the NEV industry, how the strategic choices of government departments, banks, and automobile manufacturers influence each other and what measures the government should take to encourage banks to implement green credit and reduce the impact of the decline in subsidies on the NEV industry have become theoretical and practical problems that need to be studied urgently.

Most previous studies of these problems have focused on the impact of the evolution of government policies on the NEV industry. However, the direction of government departments, banks, and automobile manufacturers in the post-subsidy era is still unclear. To fill the research gap and address the above problems, we constructed a tripartite evolutionary game model for government departments, banks, and automobile manufacturers during the subsidy decline, analyzed the evolutionary process of the system, and used MATLAB to simulate the evolutionary stable strategies (ESSs) and the sensitivity of related parameters. This paper aims to answer two questions: (1) What are the evolutionarily stable states and corresponding preconditions of the system composed of government departments, banks, and automobile manufacturers? (2) How do the strategic choices of system agents, and related parameters affect the evolution of the system equilibrium?

The contributions of this paper are summarized as follows: (1) Most existing studies focused on the impact of government subsidies on the NEV industry while ignoring the role of bank loans. This paper constructs a tripartite evolutionary game based on government departments, banks, and automobile manufacturers and studies the evolutionary process of the system and the factors affecting the evolutionary behavior of the tripartite participants. (2) When studying China’s NEV industry in the post-subsidy era, most existing studies considered only the impact of the decline in subsidies on the NEV industry while ignoring the role of the dual-credit policy. The decline in subsidies and the dual-credit policy are parallel policies in the post-subsidy era. Therefore, this paper studies the combined effects of the decline in subsidies and the dual-credit policy on the NEV industry. (3) The conclusions of our research can provide relevant decision-making suggestions for government departments, banks, and automobile manufacturers, help to promote the development of the NEV industry in China and contribute to the realization of carbon neutrality goals.

2. Literature Review

Research for this paper focused mainly on three areas: (1) research on the impact of government policies on the NEV industry; (2) research on the impact of green credit on enterprises; and (3) research on evolutionary game theory.

2.1. Research on the Impact of Government Policies on the NEV Industry

Research on the impact of government policies on the NEV industry focused mainly on three areas: research on government subsidies, research on the dual-credit policy, and research on the decline in subsidies. First, regarding research on government subsidies, some scholars compared the impact of government subsidy methods, namely government research and development subsidies, production subsidies, and mixed subsidies, on the NEV industry [8,9]. Some scholars have compared the impact of different government subsidy objects, that is, government subsidies to NEV manufacturers or subsidies to consumers on the NEV industry [10,11,12]. Secondly, regarding studies on dual-credit policy, some scholars have studied the impact of the dual-credit policy on the production decisions of automobile manufacturers [13] and the development of the NEV industry [14]. Other scholars have further studied the formulation of the NEV credit ratio requirements [15] and the coordination of the automobile supply chain under the dual-credit policy and proposed a cost-sharing contract, a revenue-sharing contract and two pricing contracts to coordinate the supply chain [16]. Finally, regarding studies on the subsidy decline policy, some scholars have studied the impact of the subsidy decline on the NEV industry [17], and they found that the subsidy decline policy is conducive to the development of the NEV industry [18], appropriate decline maximized the use of resources [19], and attention should also be paid to the synergy of the subsidy decline policy and other polices [20]. Some scholars have compared the impact of subsidy policy and dual-credit policy on the production decisions of automobile manufacturers considering battery recycling, and the study found that the dual-credit policy is more efficient than the subsidy policy [21]. Other scholars have studied the impact of the subsidy decline and dual-credit mixed policy on the NEV industry, the study found that the dual-credit policy can effectively promote the development of the NEV industry [22]. Under the mixed policy, the retail price of NEVs will be reduced, and the implementation of the dual-credit policy can reduce the dependence of enterprises on government subsidies for R&D investment, and improve the enthusiasm of enterprises for R&D [23], and the mixed policy has a more substantial guiding effect on the NEV industry [24].

2.2. Researches on the Impact of Green Credit on Enterprises

The research on the impact of green credit on enterprises is mainly divided into three aspects: research on the impact of green credit on the green technology innovation of enterprises, research on the impact of green credit on green debt financing costs of enterprises, and research on the impact of government on banks’ implementation of green credit. Firstly, regarding studies on the impact of green credit on enterprises’ green technology innovation, they found that the green credit can promote the green technology innovation of enterprises [25], and the green credit has an undeniable role in improving the green innovation of high-polluting and high-energy-consuming enterprises [26]. Secondly, regarding studies on the impact of the green credit on enterprises’ debt financing costs, they found that the green credit reduced the debt financing costs of green enterprises but suppressed the debt financing of heavily polluting enterprises [27,28]. Green credit concessions under certain conditions are more beneficial to both banks and enterprises and can increase consumer surplus and total social welfare [29]. Finally, in terms of research on government’s influence on the implementation of green credit, Zhou et al. [30,31] established an evolutionary game model between the government and banks and analyzed the ESS of the two parties involved.

2.3. Researches on the Evolutionary Game Theory

However, the above research mainly focused on government policies and enterprise bank loans. Few studies considered the mechanism of promoting the development of the NEV industry from the perspective of behavioral strategy interaction among participants. However, it is necessary to study the evolution path of the government, banks and NEV enterprises. Thus, to fill this gap, this paper will adopt evolutionary game method, which has been proved a promising method for analyzing the behavioral strategy of multi-agent [32,33]. Regarding clean energy, Wang et al. [34] built an evolutionary game model composed of investment companies, hydrogen-powered vehicle users and solar photovoltaic power plants, which provided a reference for promoting hydrogen production by hydrogen-powered vehicles and solar photovoltaic in China. Liu et al. [35] constructed an evolutionary game model composed of clean power generation enterprises and thermal power generation enterprises, and discussed the necessity and role of orderly coordination in the sustainable development of China’s power generation industry during the transition period. Regarding energy regulation, Yang et al. [36] built an evolutionary game model composed of regulators, energy enterprises and whistleblowers, which provided suggestions for optimizing China’s energy regulatory system. Regarding green building, Chen et al. [33] constructed an evolutionary game model composed of government and construction stakeholders, and studied the role and influence of government policies on the adoption of green building technology. Su et al. [37] constructed an evolutionary game model composed of government, construction contractors and waste recycling plants, which provided management enlightenment for policy makers to promote sustainable construction and demolition waste recycling.

3. Construction of the Three-Party Evolutionary Game Model

The evolutionary game is suitable for the limited rationality of the research object, and a little equilibrium state can be achieved through learning and imitation. The theory is currently widely used in the research of energy-saving and environmental protection fields such as clean energy, energy regulation, green building and NEVs [32,33,34,35,36,37,38]. Therefore, this paper uses the evolutionary game to analyze the evolutionary behaviors between the government departments, banks and auto manufacturers in the post-subsidy era of the NEV industry.

3.1. Model Assumptions

The game model has three participants: government departments, banks, and automobile manufacturers. As for government departments, the alternative strategies include low-subsidy (LS) and high-subsidy (NS). As for banks, the alternative strategies include implement green credit (I) and non-implement green credit (NI). As for automobile manufacturers, the alternative strategies include produce NEVs (N) and non-produce NEVs (NN). Combined with China’s current policy practice in the NEV market, the following basic assumptions are made:

(1)

Strategy selection and probability: The probability that government departments select LS is $x(0\le x\le 1)$, thus the probability that they select NS is $1-x$. The probability that banks adopt I is $y(0\le y\le 1)$, thus the probability that they adopt NI is $1-y$. The probability that automobile manufacturers choose N is $z(0\le z\le 1)$, thus the probability that they choose NN is $1-z$.

(2)

Government departments: When government departments choose high-subsidy, the subsidy amount for each NEV is $S_1$. In this case, the reputation of government departments is improved, and the corresponding income of reputation improvement is $F_1$. When government departments choose low-subsidy, the phase-out rate of per NEV is $k$, therefore, the subsidy amount for per NEV is $(1-k)S_1$. In this case, in order to alleviate the impact of subsidy decline on the production of NEVs by automobile enterprises, government departments will adopt the dual-credit policy and give incentives to banks that implement green credit, and the incentives to banks is $S_2$. Apart from spending on government subsidies and incentives, government departments will also obtain environmental benefits $R$ from NEVs and the environmental governance costs $C_1$ from FVs.

(3)

Dual-credit policy: Automobile manufacturers will receive NEV positive points when producing NEVs. Otherwise, they will receive NEV negative points and corporate average fuel consumption (CAFC) positive/negative points. Because there are only a few enterprises that produce FVs whose fuel consumption meets the standard, we assume that CAFC points of automobile manufacturers are negative. Besides, NEV positive points can be sold on the point trading market, and CAFC negative points can be compensated by NEV positive points [4,39]. We suppose that the CAFC negative points borne by each FV are $b$, the NEV positive points borne by each NEV are $a$, the NEV points ratio requirement is $\lambda$. The calculation formula for points is: CAFC negative points $=b\times Q_2$, NEV points $=a\times Q_1-\lambda \times Q_2$, and the transaction price of per NEV point is $P^{NEV}$.

(4)

Banks: When banks decide to implement green credit, the benefits of banks is $P_1$, and the reputation improvement due to the implemented green credit is $F_2$. When banks do not implement green credit, the benefits of banks is $P_2$.

(5)

Automobile manufacturers: When automobile manufacturers choose to produce NEVs, their net profits is $M_1{Q}_1$. Meanwhile, if banks decide to implement green credit, it will save financing costs for automobile manufacturers, which is denoted as $C_2$. When automobile manufacturers do not produce NEVs, their net profits are $M_2{Q}_2$. $M_1$ and $M_2$ represent the unit profit of NEV and FV, respectively, $Q_1$ and $Q_2$ represent the sales volume of NEV and FV, respectively.

3.2. Formulas for Modeling

Let $N_1=M_1+(1-k)S_1+a{P}^{NEV}$, $N_2=M_2-(b+\lambda )P^{NEV}$. Based on the above assumptions, the benefits of the tripartite evolutionary game between government departments, banks and automobile manufacturers can be obtained, as shown in

Table 1. The game profit value of the government, banks and automobile manufacturers.

		Auto Manufacturers (N)	Auto Manufacturers (NN)
Government (LS)	Banks (I)	$a_1:R-(1-k)S_1{Q}_1-S_2$	$a_2:-C_1$
		$b_1:P_1+F_2+S_2$	$b_2:P_2+F_2$
		$c_1:N_1{Q}_1+C_2$	$c_2:N_2{Q}_2$
	Banks (NI)	$a_3:R-(1-k)S_1{Q}_1$	$a_4:-C_1$
		$b_3:P_2$	$b_4:P_2$
		$c_3:N_1{Q}_1$	$c_4:N_2{Q}_2$
Government (HS)	Banks (I)	$a_5:R-S_1{Q}_1+F_1$	$a_6:-C_1+F_1$
		$b_5:P_1+F_2$	$b_6:P_2+F_2$
		$c_5:(M_1+S_1)Q_1+C_2$	$c_6:M_2{Q}_2$
	Banks (NI)	$a_7:R-S_1{Q}_1+F_1$	$a_8:-C_1+F_1$
		$b_7:P_2$	$b_8:P_2$
		$c_7:(M_1+S_1)Q_1$	$c_8:M_2{Q}_2$

.

Evolutionary game theory discards the assumption that participants are entirely rational, which the traditional game theory requires and is based on participant-bounded rationality. According to this theory, we conduct a balanced analysis of each subject’s strategic behavior and conclude through dynamic replicator analysis [31]. The dynamic replicator equation is a emotional differential equation that essentially gives the frequency at which a particular strategy is adopted or accepted within a population [40]. From Table 1, the expected return and average return of government departments can be calculated, as shown in Equation (1):

$$\left\{\begin{array}{l}{u}_{G1}=yz{a}_1+y(1-z)a_2+(1-y)z{a}_3+(1-y)(1-z)a_4\\ u_{G2}=yz{a}_5+y(1-z)a_6+(1-y)z{a}_7+(1-y)(1-z)a_8\\ u_G=x{u}_{G1}+(1-x)u_{G2}\end{array}\right.$$

(1)

From Equation (1), the replicator dynamic equation of the government can be determined, as shown in Equation (2):

$$F(x)=x(u_{G1}-u_G)=x(1-x)[-F_1+k{S}_1{Q}_{1}z-S_{2}yz]$$

(2)

Similarly, the expected return, average return and the replicator dynamic equation of banks can be determined, as shown in Equations (3) and (4), respectively:

$$\left\{\begin{array}{l}{u}_{B1}=xz{b}_1+x(1-z)b_2+(1-x)z{b}_5+(1-x)(1-z)b_6\\ u_{B2}=xz{b}_3+x(1-z)b_4+(1-x)z{b}_7+(1-x)(1-z)b_8\\ u_B=y{u}_{B1}+(1-y)u_{B2}\end{array}\right.$$

(3)

$$F(y)=y(u_{B1}-u_B)=y(1-y)[F_2+(P_1-P_2)z+S_{2}xz]$$

(4)

In the same way, the expected return, average return and the replicator dynamic equation of automobile manufacturers’ strategy adjustment are shown in Equations (5) and (6), respectively:

$$\left\{\begin{array}{l}{u}_{M1}=xy{c}_1+x(1-y)c_3+(1-x)y{c}_5+(1-x)(1-y)c_7\\ u_{M2}=xy{c}_2+x(1-y)c_4+(1-x)y{c}_6+(1-x)(1-y)c_8\\ u_M=z{u}_{M1}+(1-z)u_{M2}\end{array}\right.$$

(5)

$$F(z)=z(u_{C1}-u_C)=z(1-z)[(M_1+S_1)Q_1-M_2{Q}_2+(-k{S}_1{Q}_1+\overline{CV}{P}^{NEV}{Q}_1+(\Delta +\lambda )P^{NEV}{Q}_2)x+C_{2}y]$$

(6)

According to replicator dynamic equations, the probability $x$, $y$ and $z$ fluctuate with time $t$, which means that when all the replicator dynamic equations are equal to 0, that is $x$, $y$ and $z$ no longer change, the replicator dynamic system tends to reach a stable state. From Equations (2), (4) and (6), the replicator dynamic system can be written as

$$\left\{\begin{array}{l}F(x)=x(1-x)[-F_1+k{S}_1{Q}_{1}z-S_{2}yz]\\ F(y)=y(1-y)[F_2+(P_1-P_2)z+S_{2}xz]\\ F(z)=z(1-z)[(M_1+S_1)Q_1-M_2{Q}_2+(-k{S}_1{Q}_1+a{P}^{NEV}{Q}_1+(b+\lambda )P^{NEV}{Q}_2)x+C_{2}y]\end{array}\right.$$

(7)

3.3. Stability Analysis of the Equilibrium Points in The Tripartite Game

Let $F(x)=0$, $F(y)=0$, $F(z)=0$, we can obtain eight evolutionary equilibrium points for pure strategy solutions of the three parties: $E_1=(0,0,0)$, $E_2=(1,0,0)$, $E_3=(0,1,0)$, $E_4=(0,0,1)$, $E_5=(1,0,1)$, $E_6=(1,1,0)$, $E_7=(0,1,1)$, $E_8=(1,1,1)$. Friedman, Ritzberger and Weibull confirmed that a replicator dynamic system’s hybrid strategy Nash equilibrium can never become a stable point [41]. Therefore, only the stability of these eight pure equilibrium points must be determined. Additionally, the Lyapunov system stability theory is usually used to judge the asymptotic stability of the equilibrium point. The Jacobian matrix based on the Lyapunov theory is as follows:

$$J=\left[\begin{array}{ccc}\frac{\partial F(x)}{\partial x}& \frac{\partial F(x)}{\partial y}& \frac{\partial F(x)}{\partial z}\\ \frac{\partial F(y)}{\partial x}& \frac{\partial F(y)}{\partial y}& \frac{\partial F(y)}{\partial z}\\ \frac{\partial F(z)}{\partial x}& \frac{\partial F(z)}{\partial y}& \frac{\partial F(z)}{\partial z}\end{array}\right]$$

(8)

Due to space limitations, the value of each partial derivative in the matrix is not listed, and the eigenvalues and eigenvalue symbols of each equilibrium point are shown in

Table 2. Eigenvalue and eigenvalue symbol of equilibrium points.

Equilibrium Points	Eigenvalue			Eigenvalue Symbol
	${\mathit{\lambda }}_1$	${\mathit{\lambda }}_2$	${\mathit{\lambda }}_3$
$E_1=(0,0,0)$	$-F_1$	$F_2$	$(M_1+S_1)Q_1-M_2{Q}_2$	$(-,+,\times )$
$E_2=(1,0,0)$	$F_1$	$F_2$	$N_1{Q}_1-N_2{Q}_2$	$(+,+,\times )$
$E_3=(0,1,0)$	$-F_1$	$-F_2$	$(M_1+S_1)Q_1-M_2{Q}_2+C_2$	$(-,-,\times )$
$E_4=(0,0,1)$	$-F_1+k{S}_1{Q}_1$	$F_2+P_1-P_2$	$-[(M_1+S_1)Q_1-M_2{Q}_2]$	$(\times ,\times ,\times )$
$E_5=(1,0,1)$	$-(-F_1+k{S}_1{Q}_1)$	$F_2+P_1-P_2+S_2$	$-(N_1{Q}_1-N_2{Q}_2)$	$(\times ,\times ,\times )$
$E_6=(1,1,0)$	$F_1$	$-F_2$	$N_1{Q}_1-N_2{Q}_2+C_2$	$(+,-,\times )$
$E_7=(0,1,1)$	$-F_1+k{S}_1{Q}_1-S_2$	$-(F_2+P_1-P_2)$	$-[(M_1+S_1)Q_1-M_2{Q}_2+C_2]$	$(\times ,\times ,\times )$
$E_8=(1,1,1)$	$-(-F_1+k{S}_1{Q}_1-S_2)$	$-(F_2+P_1-P_2+S_2)$	$-(N_1{Q}_1-N_2{Q}_2+C_2)$	$(\times ,\times ,\times )$

.

According to Lyapunov theory, the asymptotic stability of the replicator dynamic system needs to satisfy that all eigenvalues of the Jacobian matrix are negative [42]. From the results in Table 2, $E_1=(0,0,0)$, $E_2=(1,0,0)$ and $E_6=(1,1,0)$ can be excluded, and we only need to analyze five other equilibrium points of $E_3=(0,1,0)$, $E_4=(0,0,1)$, $E_5=(1,0,1)$, $E_7=(0,1,1)$ and $E_8=(1,1,1)$.

Scenario 1:

$(M_1+S_1)Q_1-M_2{Q}_2+C_2<0$.

In this situation, for automobile manufacturers, under high subsidy and banks implementing green credit, the income of producing NEVs is less than that of not producing NEVs. Therefore, automobile manufacturers prefer not to produce NEVs. At this time, $E_3=(0,1,0)$ is ESS.

Scenario 2:

$-F_1+k{S}_1{Q}_1<0$, $F_2+P_1-P_2<0$, $(M_1+S_1)Q_1-M_2{Q}_2>0$.

In this situation, reputation improvement is more significant than saved fiscal expenditures for government departments. For banks, the sum of benefits of implementing green credit and reputation improvement is less than that of not implementing green credit. For automobile manufacturers, the income of producing NEVs under high-subsidy and banks not implementing green credit is higher than that of not producing NEVs. Then, government departments prefer high subsidy, banks prefer not to implement green credit, and automobile manufacturers prefer producing NEVs. At this time, $E_4=(0,0,1)$ is ESS.

Scenario 3:

$-F_1+k{S}_1{Q}_1>0$, $F_2+P_1-P_2+S_2<0$, $N_1{Q}_1-N_2{Q}_2>0$.

In this situation, for government departments, saved financial expenditures are greater than the reputation improvement. For banks, the benefits of not implementing green credit are greater than the sum of the benefits of implementing green credit, reputation improvement and government incentives. For automobile manufacturers, the net income of producing NEVs under low subsidy and banks not implementing green credit is greater than that of not producing NEVs. Therefore, government departments prefer low subsidy, banks prefer not implementing green credit, and automobile manufacturers prefer producing NEVs. At this time, $E_5=(1,0,1)$ is ESS.

Scenario 4:

$-F_1+k{S}_1{Q}_1-S_2<0$, $F_2+P_1-P_2>0$, $(M_1+S_1)Q_1-M_2{Q}_2+C_2>0$.

In this situation, reputation improvement is higher for government departments than the difference between saved financial expenditures and incentives to banks. For banks, the sum of the benefits of implementing green credit and reputation improvement is more significant than benefits of not implementing green credit. For automobile manufacturers, the sum of the income from producing NEVs under high subsidy and financing costs saved by the green credit is greater than that of not producing NEVs. Therefore, government departments prefer high subsidy, banks prefer implementing green credit, and automobile manufacturers prefer producing NEVs. At this time, $E_7=(0,1,1)$ is ESS.

Scenario 5:

$-F_1+k{S}_1{Q}_1-S_2>0$, $F_2+P_1-P_2+S_2>0$, $N_1{Q}_1-N_2{Q}_2+C_2>0$

In this situation, the difference between saved fiscal expenditures and incentives to banks is higher than the reputation improvement for government departments. For banks, the sum of benefits of implementing green credit, reputation improvement and government incentives are more significant than the benefits of not implementing green credit. For automobile manufacturers, the sum of the income from producing NEVs saved financing costs under low subsidy, and banks implementing green credit is greater than that of not producing NEVs. Therefore, government departments prefer low subsidy, banks prefer implementing green credit, and automobile manufacturers prefer producing NEVs. At this time, $E_8=(1,1,1)$ is ESS.

4. Numerical Simulation

The parameter settings in the numerical simulation are shown in

Table 3. Simulation parameter setting.

	${\mathit{E}}_3 = (0,1,0)$	${\mathit{E}}_4 = (0,0,1)$	${\mathit{E}}_5 = (1,0,1)$	${\mathit{E}}_7 = (0,1,1)$	${\mathit{E}}_8 = (1,1,1)$
$S_1$	1	2	1	1	1
$F_1$	15	65	15	25	15
$k$	0.2	0.2	0.2	0.2	0.2
$a$	2	2	2	2	2
$P^{NEV}$	0.3	0.3	0.3	0.3	0.3
$b$	3	3	3	3	3
$\lambda$	0.1	0.1	0.1	0.1	0.1
$S_2$	10	10	10	10	10
$P_1$	100	100	100	110	100
$P_2$	120	120	130	120	120
$F_2$	15	15	15	15	15
$M_1$	2	2	2	2	2
$M_2$	4	3.2	3.2	3.2	3.2
$Q_1$	150	150	150	150	150
$Q_2$	180	170	170	170	170
$C_2$	50	50	100	100	100

. Firstly, all parameter settings must meet the stable conditions of the corresponding equilibrium point. Secondly, some parameters are set according to reality. For example, the subsidy per NEV is 0 to 60,000 yuan. The NEV subsidy standard will be reduced by 10% or 20% based on 2020 in 2021 [17]. The NEV points calculation method is determined by New Energy Passenger Vehicle Model Points Calculation Method, the transaction price of NEV points reached 3000 yuan per point in 2021, and the Ministry of Industry and Information Technology’s requirements for the ratio of NEVs in 2019, 2020 and 2021 are 10%, 12% and 14%, respectively. Therefore, we let $S_1=1$, $k=0.2$, $a=2$, $P^{NEV}=0.3$, $\lambda =0.1$.

4.1. Evolutionary Equilibrium Points Simulation Analysis

Firstly, we simulate and analyze the five equilibrium points derived from the above-mentioned theory. Without loss of generality, the initial state of the evolutionary game system is set as $(0.4,0.4,0.4)$. That is, the government chooses a low-subsidy strategy with a probability of 0.4, banks choose to implement a green credit strategy with a probability of 0.4, and automobile manufacturers choose a strategy for producing NEVs with a probability of 0.4.

There are five possible equilibrium points in the system. At the same time, according to the financial expenditure of subsidies and the willingness of enterprises to produce NEVs, the development of the NEV industry in China can be divided into three stages, namely early, middle and late stages and five equilibrium points correspond to these three stages, respectively.

In the early stage of the NEV industry, the government promoted the development of the NEV industry with high subsidies and bank support with green credit. However, the income of automobile manufacturers producing NEVs is still lower than those not producing NEVs, enterprises tend not to produce NEVs. In the early stage, the development of the NEV industry was slow, corresponding to $E_3=(0,1,0)$ at this time, as shown in Figure 1a. In the middle stage, the government still supports enterprises to produce NEVs with high subsidies. With technological upgrading and the expansion of the NEV consumer market, the income of automobile manufacturers producing NEVs is increasing gradually. Enterprises tend to produce NEVs under high subsidies, and banks choose whether to simultaneously implement green credit according to the income obtained from the implementation of green credit. In the middle stage, the NEV industry is developing rapidly, corresponding to $E_4=(0,0,1)$ and $E_7=(0,1,1)$ at this time, as shown in Figure 1b,c. In the late stage, excessive subsidies will hinder the technological upgrading of NEV manufacturers. Therefore, the subsidy will gradually decline at this stage, and enterprises will produce NEVs individually or under the support of the green credit. In the late stage, the NEV industry has matured, corresponding to $E_5=(1,0,1)$ and $E_8=(1,1,1)$ at this time, as shown in Figure 1a,e.

4.2. Parameter Sensitivity Analysis

Currently, the NEV industry in China is going through a transitional stage from middle to late. The government is helping automobile manufacturers to transform in this process, which not only implements the subsidy policy phase-outs and dual-credit policy, but also strengthens policy requirements for banks to implement green credit. Therefore, the influences of several typical parameters on the evolutionary behaviors of three participants are estimated, including the phase-out rate, the transaction price of NEV points, the government incentives to banks and financing costs saved by the green credit for enterprises. When analyzing the influence of one of the parameters, the value of other parameters is the same as the simulation value of $E_8=(1,1,1)$ in Section 4.1.

4.2.1. Analysis on the Phase-Out Rate

$k$ is set to 0.1, 0.2, and 0.3 to investigate its effects on the strategies of stakeholders. Figure 2a shows that if the phase-out rate is low, i.e., $k=0.1$, government departments tend to high-subsidy because the financial expenditure that the subsidy decline saves is less than the benefits of reputation improvement. However, with the increase in the phase-out rate, the financial expenditure that the subsidy decline saves is higher than the benefits of reputation improvement. Thus, government departments tend to be low-subsidy. Figure 2b shows that when $k=0.1$, the strategy of banks is unstable because government departments tend to offer high subsidies. At the same time, there are no incentives for banks under high subsidies, and banks lack the motivation to implement green credit. With the increase in the phase-out rate, government departments tend to be low-subsidy, and government departments will give incentives to banks at this time. Thus, banks tend to implement green credit. Figure 2c shows that when the phase-out rate is low, i.e., $k=0.1$ the strategy of automobile manufacturers fluctuates. With the increase in the phase-out rate, automobile manufacturers tend to produce NEVs, which indicates that the rise in the phase-out rate will encourage enterprises to produce NEVs.

4.2.2. Analysis on the Transaction Price of NEV Points

Let $P^{NEV}$ equals 0.08, 0.3, and 1 to identify the effects of the transaction price of NEV points on the strategies of stakeholders. Figure 3 shows that when the transaction price is low, the strategies of government departments, banks and automobile manufacturers fluctuate slightly. Still, in the end, the three parties tend to implement active strategies. With the increase in the transaction price of NEV points, the three parties are more inclined to implement active strategies. That is, government departments are more prone to low-subsidy, banks are more willing to implement green credit, and enterprises are more inclined to produce NEVs. When the transaction price is 3000/min and 10,000/min, the respective evolution paths of the three parties almost overlap, indicating that when the transaction price of NEV points rises to a certain level, the continued increase will not have much impact on the strategies of the three parties.

4.2.3. Analysis on Government Incentives to Banks

$S_2$ is set to $S_2=10$, $S_2=15$, and $S_2=20$ to investigate its effects on stakeholders. Figure 4a,b show a threshold between 15 and 20 for the impact on system evolution results. When this threshold is exceeded, excessive incentives will bring a financial burden to government departments, and strategies of government departments and banks will be unstable. When the government incentives are lower than this threshold, with the improvement of incentives, government departments tend to be low-subsidy, and banks tend to implement green credit. Figure 4c shows that three curves of automobile manufacturers’ evolutionary behaviors almost overlap, which indicates that the government incentives to banks have little influence on the strategies of enterprises.

4.2.4. Analysis on the Financing Costs Saved by the Green Credit

$C_2$ is assigned with values $C_2=10$, $C_2=45$, and $C_2=100$ to investigate the effects of the financing costs saved by the green credit on the strategies of stakeholders. Figure 5 shows that with the increase in the saved financing costs, government departments tend to be low-subsidy, and automobile manufacturers tend to produce NEVs. However, the changes in saved financing costs have little impact on the strategies of banks.

5. Discussion

This subsection illustrates the differences between this study and existing research on China’s NEV industry post-subsidy era. Sun et al. [17] and Ji et al. [18] studied the evolutionary game model between government departments and enterprises in the post-subsidy era and found that the gradual withdrawal of subsidy policy benefits to the development of the NEV industry. This study also draws a similar conclusion, but this paper extends the situation of low subsidy to the situation where the subsidy policy and the dual-credit policy are implemented at the same time. Many scholars have studied the impact of the dual-credit policy on China’s NEV industry. Lu et al. [16] studied automobile manufacturers’ pricing and emission-reduction strategies under the dual-credit policy. They found that the dual-credit policy can actively promote fuel automobile manufacturers to reduce emissions, and the government should set higher transaction prices for NEV points. Li et al. [24] studied the impact of the dual-credit policy on the cooperative innovation of upstream and downstream enterprises in the NEV industry. They found that only when the transaction price of NEV points is high enough can the innovation of upstream and downstream enterprises be effectively promoted. Zhang et al. [39] studied the impact of dual-credit policy on automobile manufacturers’ production strategy and social welfare. They found that a dual-credit policy can effectively promote the development of China’s NEV industry. Still, these scholars did not consider the impact of the subsidy decline and dual-credit policy on China’s NVE industry under the synergistic effect. The reasons for considering these two policies at the same time are mainly based on the reality of China’s NEV industry, that is, the government adopted high-subsidy in the industrial development, gradually reduced the subsidy rate in recent years, and implemented a series of measures such as a dual-credit policy to promote the steady growth of China’s NEV industry. Li et al. [3] considered the impact of subsidy decline and dual-credit policy on the production strategy of automobile manufacturers. They found that subsidy decline can partially offset the impact of dual-credit policy on the NEV industry, and there is an optimal transaction price of NEV points to maximize the system’s profits. This is similar to the conclusion of this study. But based on studying the influence of government subsidies on the production strategy of automobile manufacturers, this paper further studies the influence of bank green credit. It constructs an evolutionary game model based on government departments, banks and automobile manufacturers to study the evolutionary behavior of the three parties and the influence of related factors on their evolutionary behavior.

6. Conclusions, Policy Recommendations and Limitations

6.1. Conclusions

Regarding the development of the NEV industry under the subsidy decline, most scholars have only studied the two-party evolutionary game between government departments and enterprises, and in considering the impact of the subsidy decline policy on automobile manufacturers, most studies only consider the impact of subsidy decline on enterprises. Therefore, in the context of the post-subsidy era of NEVs, this paper considers both the subsidy policy and the dual-credit policy, constructs a tripartite evolutionary game model among government departments, banks and automobile manufacturers, analyzes the evolution process of the system, and uses MATLAB to simulate the ESS and the sensitivity of related parameters. The results show that: (1) five evolutionary equilibrium points of the system correspond to three stages of the NEV industry in China. In the early stage, $E_3=(0,1,0)$ is the only equilibrium point. In the middle stage, there are two possible equilibrium points, namely $E_4=(0,0,1)$ and $E_7=(0,1,1)$. In the late stage, $E_5=(1,0,1)$ and $E_8=(1,1,1)$ are two possible equilibrium points. (2) With the increase in the phase-out rate and the transaction price of NEV points, government departments tend to be low-subsidy, banks implement green credit, and automobile manufacturers produce NEVs. However, when the transaction price of NEV points rises to a certain level, its continued increase will not have much impact on the strategies of the three parties. (3) There is a threshold for the impact of government incentives on the evolutionary behavior of government departments and banks. When this threshold is exceeded, the strategies of government departments and banks will be unstable. When it is below this threshold, with the increase in government incentives, government departments tend to be low-subsidy, and banks tend to implement green credit. However, the government incentives to banks have little impact on the strategies of automobile manufacturers. (4) With the increased financing costs saved by the green credit, government departments are more inclined to be low-subsidy, and automobile manufacturers are more prone to produce NEVs. However, the changes in saved financing costs have less impact on banks’ strategies.

6.2. Policy Recommendations

Given the reality that the NEV industry in China is in the post-subsidy era, this paper puts forward relevant policy recommendations for government departments, banks and automobile manufacturers based on the above conclusions. The research results have the following policy implications for government departments: (1) government departments should gradually increase the phase-out rate and supervise the implementation of the dual-credit policy of the NEV industry; (2) government departments should set the minimum transaction price of NEV points to guide the formation of the points trading market; (3) government departments have insufficient incentives for banks to implement green credit at present, so government departments should increase incentives for banks in the future.

The research results have the following policy implications for banks: banks and auto manufacturers only have connections in terms of loan funds. In the future, the relationship between banks and enterprises can be deepened. Banks can also consider developing products other than green credit to support the development of the NEV industry.

The research results have the following policy implications for automobile manufacturers: the dual-credit point amendment weakens the impact of pure NEV driving range in the calculation of model points and strengthens the requirements for indicators such as energy consumption that reflect the progressive nature of the vehicle. Therefore, the future development direction of automobile manufacturers should be to produce NEVs with higher vehicle performance, quality, and safety levels rather than based on production volume.

6.3. Limitations

Although this study obtained many exciting findings about China’s NEV industry in the post-subsidy era, it has limitations that need further investigation. It is appreciated that the NEV industry is a complex systematic project that involves many stakeholders and factors. Still, this study only considered 18 key parameters after determining the three stakeholders and their relationships. Further studies can consider additional factors to conduct comprehensive system and policy simulations. For example, a four-way evolutionary game model can be built based on the government, banks, NEV enterprises and consumers, and more factors can be considered in the parameter setting.