The Impact of Firm Heterogeneity and External Factor Change on Innovation: Evidence from the Vehicle Industry Sector

Abstract

We explore the connection between firms’ technological innovation capabilities and their internal and external factors. To empirically test this relationship, we use panel data for new energy vehicle (NEV) firms and traditional fuel vehicle firms in China from 2010 to 2020. Our findings show that public subsidies do have a positive impact on firms’ technology innovation capability, and there are consistent findings for both NEV and traditional fuel vehicle firms. Firms have a supportive effect on their innovative ability when they satisfy conditions of high profitability, low leverage, high equity concentration, and highly educated employees. The inability to maximize the effectiveness of public subsidies is due to an imbalance in the internal and external factors of firms. Therefore, we innovatively analyze the internal and external factors of NEV firms as an integrated system, taking into account the high correlation between them, rather than discussing them separately. The paper is not only of academic significance to the development of NEV firms to improve their technological innovation capability and the transformation of traditional fuel vehicle firms, but also of practical significance to the reduction of greenhouse gas emissions and the achievement of the “double carbon” goal.

1. Introduction

The increasing global concern for a low-carbon economy and sustainable energy development has been triggered by global warming [1]. New energy vehicles (NEVs), which have the advantage of using non-polluting and renewable power fuels, are highly valued by governments aiming to address energy transition development. NEVs play a vital role in the implementation of sustainable development strategies. They reduce fossil energy consumption and mitigate the greenhouse gas effect. Accompanied by the realization of the double carbon target, NEVs have undergone rapid expansion under encouraging policies from the Chinese government and are now entering a phase of annually decreasing public subsidies. Research on the impact of heterogeneity and of the external environmental changes of automobile firms on their innovation capacity enhancement is not only of theoretical significance to the transformation and upgrade of the automobile industry, but also of practical significance to the alleviation of environmental pressures brought about by climate change.

With the gradual removal of public subsidies, the pressure on the NEV industry has increased dramatically, making it difficult for many emerging NEV firms to maintain growth in production and sales. The further development of the industry relies on NEV technology research and transformation development of traditional vehicle manufacturers. Transformation is a process change that includes technological innovation [2], the rapid transformation of traditional fuel vehicle firms, usually accompanied by intensive R&D investment and the coordinated development of specific social factors, is still in the beginning stages [3]. However, the sociotechnical system of NEVs is not yet well developed and cannot satisfy consumer needs [2]. In the process of the technological breakthrough of NEV firms and the business transformation of traditional fuel vehicle firms, both types of businesses need to pursue long-term competitive advantages through innovation [2]. Endogenous innovation in the automotive industry is triggered by the exogenous environment, making the firm “balanced inside and outside” by considering the internal and external factors systematically, is crucial to keeping innovation alive.

The R&D process is complex and volatile, which is reflected in the irregularity of the activity cycle and the process non-separability; it requires high investments and high risks [4]. The struggle to obtain external financing for R&D activities in NEV firms stems from the problematic perceived barriers to their development potential and legitimacy by individuals with external stakeholder relationships. It has been widely discussed that public subsidies contribute to solving market failures caused by the positive externalities of innovation, improving the innovation performance of NEV firms and alleviating resource scarcity [5]. However, the Chinese government has, as of 2017, decided to eliminate these subsidies to avoid subsidy dependence in the NEV industry, and so reducing the burden on government finances [4]. The promotion of NEV firms has changed from an initial promotion-oriented universal industry-driven policy to a market-driven policy for new energy technology innovation. Thus, does public subsidy actually contribute to the technological innovation capability of NEV enterprises? Compared with traditional fuel vehicles, what are the specific differences in the amount and effect of subsidies for the two types of enterprises? Is there any negative impact of the subsidy withdrawal on the development of NEV enterprises?

Market-driven industries require stronger endogenous dynamics, which depend on their endowments and characteristics, as well as the interaction between various factors within the enterprise. Currently, the literature discusses the impact of government policies on firm innovation developed at the industry and national levels [1], few papers have studied firm heterogeneity in depth based on the micro-entity dimensions [6]. Considering the heterogeneity between industries and enterprises, we selected variables representing the business capability and firm characteristics from the micro-entity, comprehensively studying the balanced relationship between the internal and external factors of the firms. Firm heterogeneity includes financial performance (profitability, cash flow, asset-liability ratio, etc.) and firm characteristics determined by their stage of development (firm size, governance level, employee education level, etc.). The firm’s reserve financial flexibility can effectively and quickly mobilize funds, grasp the investment opportunities affected by uncertainty, release financing constraints, and promote the improvement of enterprise innovation ability [7]. Firm characteristic variables are also determinants of innovation behavior. For example, the influence of public subsidies on firm performance varies with the size of firms [8], while ownership concentration reflects the restraining ability of shareholders to supervise management and so helps to strengthen supervision, restrain rent-seeking behavior, improve corporate governance level, and make a better capital investment [9]. NEV firms and traditional fuel vehicle firms have different characteristics due to their different stages of development, how will this contribute to their STI capabilities? Does the financial performance of different firms reflect the support of endogenous dynamics among them, and does it have a different impact on their innovation capability?

We explore the factors the internal and external factors affecting the STI capability of automotive firms by using the data of listed firms from 2010 to 2020 to examine the contribution of public subsidies, the exogenous currency entries into the automotive industry, to firms’ STI capabilities. Continuous investment capability in R&D technology is directly determined by the current financial situation and the stage of development within the firms, constructing financial performance and a firm’s characteristics index based on heterogeneity. We find that positive motivation of public policy support for enterprises’ STI ability and that improving the balance between enterprises’ endogenous development motivation and their external environment can increase the efficiency of public subsidies. High profitability and low leverage within the firm are conducive to innovation, and high equity concentration and highly educated employees contribute positively to innovation at an NEV firm. We analyzed internal and external factors of NEV firms as an integrated system, taking into account the high correlation between them, rather than discussing them separately. The success of systemic innovation requires the industry as a whole to engage in a transition [10]. This research is conducive to the coordination of the synergistic development of the internal and external factors of NEV firms and traditional fuel vehicle firms, and to the maximization of their innovation potential, helping to achieve breakthroughs in NEV technology and the transformation and upgrading of traditional fuel vehicle firms. It has practical significance in the achievement of the goal of curbing global warming.

The three contributions of this paper are as follows: (1) based on the NEV industry, and in contrast to the development of the traditional fuel vehicle industry, we consider the different mechanisms of public subsidies on improving STI capability, due to the enterprises’ heterogeneity. (2) We research the different characteristics of NEV and traditional fuel vehicle firms at the micro-entity layer and their ability to improve STI capability at the different stages of development, as well as the mechanism behind the internal and external factors’ impacts on STI capability. (3) We select financial performance and characteristics variables describing the sustainability of the firm, studying the variables within firms in different dimensions, as these variables play a decisive role in STI capacity enhancement. This paper has theoretical contributions to the internal and external factors of firm innovation and related research in the field of the transformation and upgrade of the automotive industry.

In the following section, we present a literature review on the interrelations between firms’ internal and external factors and innovation. Then, in Section 3 and Section 4, we provide a model to investigate the relationship and describe the data used to test the model. In Section 5, we outline the empirical methodology for measuring the relationship, firms’ internal and external factors and innovation, perform panel framework tests to establish and estimate the empirical model, and present the findings. Finally, we summarize our main findings in Section 6, contributions, limitations, and future research in Section 7.

2. Literature Review

According to the literature on enterprise innovation [11], innovation drive factors are classified into external and internal factors [12]. External factors include R&D subsidies, tax relief, etc., while internal factors include financial performance and firm characteristics (e.g., financial performance, firm size). Shao (2021) has concluded that there have been obvious incentive effects of public subsidies on the R&D behavior of NEV firms [13], while Yu (2020) has argued that the motivating effect gradually diminished with the increase in public subsidies [9]. Based on intra-firm characteristics, studies in the renewable energy industry have shown that firms in a overall healthy financial position have a positive impact on R&D performance [14]. However, the relevant empirical findings have not been confirmed in the NEV industry, which is not beneficial to the sustainable development of the NEV industry in China. This paper examines the role of the internal and external factors on STI capability improvement for NEV and traditional fuel vehicle firms. Further, internal and external factors are analyzed as an integrated system, taking into account the high correlation between internal and external factors of an automotive firm, rather than discussing them separately.

2.1. Public Subsidies and STI Capability

That public subsidies decrease a firm’s barriers to joining the NEV industry, reducing investment risks and costs, reflects the role of public subsidies in promoting enterprises [15]. The rapid growth of production and consumption in the early development stage of the NEV industry is the result of large-scale financial support and various policies provided by the government [16]. It has been argued that the scale of subsidies determines how they affect firms’ R&D behavior; that the innovation efficiency decreases when the subsidies are small-scale and increases when public subsidies reach a certain size [17].

Some scholars form the opposite conclusion; they point out that public subsidies do not have a positive impact in promoting the improvement of firms’ STI capability as we expect [18]. Long-term public subsidies allow NEV firms to become dependent, even inducing some companies to “seek compensation”. We have made great progress in the innovation of NEVs, but there is still a core technology bottleneck at the moderate stage of overall R&D [19]. The background of “deglobalization” caused by COVID-19 has disrupted the production of automotive chips, directly reducing the production of NEVs. “The subsidized” industrial policy has enhanced the pioneering and scale advantages of the NEV industry for a short period of time but has failed to form core technological breakthroughs and brand advantages. The growth of a firm’s competencies and performance need to be supported simultaneously by the assurance of R&D funding and the endogenous drivers themselves.

As mentioned above, public subsidies do have a positive effect on the R&D performance of the NEV firms at an early stage, but the drawbacks of industry-driven policy in a universal benefit appear as the industry expands, which leads to the following hypothesis in this paper.

Hypothesis (H1).

Public subsidies have a positive impact on improving the STI capability of NEV firms.

2.2. Financial Performance and STI Capability

Previously, research on policies related to enhancing firms’ innovation capabilities has focused on the macro level [20], or on the crowding-out effect of public subsidies on firms’ innovation activities [21]. The heterogeneity reflected how different automotive firms have different resources and endowments [22], which has not only led public support to be at different levels, but also varied the policy effect over time. Therefore, it is necessary to consider the firm’s heterogeneity and explore the innovation driving forces within the firm, so that it can formulate and implement a reasonable development strategy for its own resources and capabilities.

A firm’s endogenous driving forces are represented by the current state of its finances and characteristics. Enterprise characteristics are indicated by the technical knowledge base and the human resources accumulated. The positive relationship between subsidies and innovation capacity also applies to the financial status of the firm’s operations. There are abundant resources to support the expansion of its research areas and support higher risk R&D projects when the financial status of a NEV firm is at a relatively favorable level. Those projects tend to be of higher technical demand and offer higher economic value. Profitability is the core of a sustainable and long-term business viability [23]. The level of paid equity determines whether it has sufficient available resources to innovate. High leverage would limit its ability to innovate, as bondholders tend to prefer to reduce investment risk [24]. With a high proportion of internal funding reducing the dependence on bondholders, a firm avoids the information loss of new inventions and promotes firm innovation. We put forward the following hypothesis in this paper:

Hypothesis (H2).

The financial performance of NEV firms has a positive effect on enhancing STI capability.

2.3. Firm Characteristics and STI Capability

Firm size is supportive of innovation activities to a certain extent, and small-scale firms can more easily face the complexities of innovation and address the shortage of corporate R&D funds, compared with larger firms. It was found that the direct impact of R&D subsidies on a firms’ R&D expenditures was positive, and that degree depended on the firm size [25]. In general, the technological resources available to the firm were also related to the number of employees [26]; personnel with high education levels play a crucial role in the successful implementation of new energy technologies in firms. The degree of equity concentration is also known as ownership concentration, which determines the level of corporate governance and decision-making style. Enterprise ownership concentration or fragmentation may positively or negatively affect the innovation capability of high-tech enterprises; when equity is concentrated, the shareholder can positively or negatively influence the R&D performance by increasing or decreasing its investment in innovation. On the contrary, equity diversification balances the interests of shareholders, thus moderating their individual influence on corporate decisions [27]. Based on the above analysis, the following hypothesis is put forward in this study:

Hypothesis (H3).

Firm characteristics have a positive effect on the STI capability.

2.4. Internal and External Factor of Traditional Fuel Vehicle Firm

From the perspective of path dependence, traditional fuel vehicle firms tend to maintain their existing business models and technology levels. The green environmental development and NEV breakthrough innovations put pressure on the traditional vehicle industry, as NEV firms receives more support from the exogenous context, traditional fuel vehicle firms tend to break the existing path dependence. Innovation essentially entails a process of coevolution within the features of systematic characteristics [10]; the auto industry is a complex system comprising various interrelated elements. it is also crucial for traditional fuel vehicle firms to transform and upgrade in the process of decarbonization of the auto industry. The replacement of internal combustion vehicles by NEVs will require technological advancement and various coordinated external factors. Linking the external factors, supported by the government, with the firm’s internal factors, and emphasizing the importance of industrial coordination is an effective way to break this path dependency [10]. The sociotechnical system of traditional fuel vehicles is stable in the current stage, but there have been certain fluctuations within the system, such as policy support toward NEVs and consumer preference changes. As a result, creating a mutually supportive external factor and a positive internal operating mechanism between the government, NEV firms, and traditional fuel vehicle firms helps to form a sustainable and low-carbon transition path for China’s automobile industry [28].

This text, thus, proposes three hypotheses corresponding to NEV firms and explores the path of innovation and the state of traditional fuel vehicle firms.

Hypothesis (H4).

Public subsidies, financial performance, and firm characteristics all have a positive impact on the STI capability in traditional vehicle firms.

3. Sample and Data Collection

This paper used panel data for the NEV firms and traditional fuel vehicle firms in China from 2010 to 2020. We collected data on 439 listed firms in the automotive industry, there are 1918 samples of NEV firms and 2882 samples of traditional fuel vehicle firms. The data were obtained from WIND and CSMAR databases and calculated after collation.

Table 1. Description of the variables.

Type of Variables	Variable	Definition
Dependent Variables
	STI capability (STI)	Scientific, technological and innovative capability
Input	R&D expenditures	Research and development expenses
	R&D personnel	Number of Research and Development personnel
Output	Patent	Number of patents granted
	Operating Income	Total operating income
Independent Variables
External factor	Public subsidy	Total Public Subsidy
Internal factors	Financial Performance
	Return on Total Assets (ROA)	The ratio of net income to the total assets
	Capital Leverage (LEV)	Assets and Liabilities
	Dividend Payout Ratio (Payouts)	Share of dividends in net income
	Equity Attributable to Shareholders of Parent Company (Equity Attributable)	Equity of the parent company in the consolidated statements
	Firm Characteristics
	Herfindahl Index (Herf)	Ownership concentration
	Number of employees (No. of Employee)	Firm size
	Employee education level (Education)	Graduate degree or above
	Gender	Gender of executives
Control variables	Regional Economic Development level (REDI)	Annual regional GDP value

shows all variables and their definitions in this paper.

3.1. Scientific, Technological and Innovative (STI) Capacity

In most studies, the patents granted were selected as a proxy variable for firms’ in-novation capability [29]. Factors affecting a firm’s innovative ability are compound and multiple as considering only one side is inadequate to express innovation capability. The integrated STI capability index was derived through the scientific calculation of enterprise R&D and innovation-related indicators.

In a typical DEA model, the objective is to compare the efficiency of similar elements based on predetermined inputs and outputs. The indicators related to the R&D of firms were divided into inputs and outputs. R&D input was measured in two aspects, human capital investment and financial investment. The quality of R&D personnel is related to the economic level, which is crucial to innovation [30]. The R&D output was also illustrated in two aspects, the firm patent applications and its product sales revenue. Since DEA cannot have null values in the calculation process, we obtained the complete dataset by multiple interpolation of data through the Miceforest algorithm in python, and the DEA results were also calculated by python.

3.2. Financial Performance

Corporate financial performance was characterized by the following four variables: return on total assets, dividend payout ratio, capital leverage, and shareholders’ equity attributable to the parent company, all of which examine the reflection of financial performance due to changes in the current operating status in different dimensions. Return on assets (ROA) was selected as the profitability indicator, as profitability may affect innovation decisions. Enterprises with strong profitability may carry out more exploratory innovation to maintain long-term competitive advantages. We used the cash dividend payment rate (Payouts) to measure the financing constraints of enterprises. The capital leverage (LEV) was used as an instrumental variable to measure financial leverage. A high leverage will leave companies with insufficient capital to invest [31]. Equity attributable to shareholders of the parent company divided by all invested capital indicates the stability of firm management, which varies according to the growth stage the firm is in.

3.3. Firm Characteristics

The firm’s characteristics, determined by its operation stage, were represented by the following variables: total number of employees, level of employee education, ownership concentration, and gender. Firm size is an important factor affecting the firm’s survival and performance [13]. Studies have found small firms are often more susceptible to the macroeconomic environment than large ones. The influence of public subsidies on performance varies with the size of the enterprise. Ownership concentration is associated with high shareholder involvement and closer monitoring of firm management. Dominant shareholders may exert greater power, limiting managerial discretion and actively influencing corporate management. The level of employee education indicates to a certain extent the attractiveness and importance of the firm to highly qualified personnel, which can affect the improvement of R&D performance.

4. Model Specification

4.1. Basic Regression Model

The current study considered two factors in its analysis of the relationship between public policies and innovation. First, our model includes firms’ heterogeneous characteristics and financial performance as control variables, which may affect the relationship between public subsidies and STI capability, thereby identifying the effects of public subsidies on STI capability. Second, our study examined the contribution of enterprise’ heterogeneity to STI capability, taking financial performance and firm characteristics as independent variables, controlling for the level of economic development in different regions. Using financial performance and firm characteristics as independent variables, we investigated the contribution of firm heterogeneity to STI capability.

We used the following panel regression model to investigate the relationship between public subsidies and firms’ innovation:

$$ST{I}_{i,t}={\phi }_0+{\phi }_{1}subsid{y}_{i,t}+{\phi }_2{Z}_{i,t}+{{\displaystyle \sum }}^{ }year+{\eta }_i+{\epsilon }_{i,t}$$

(1)

where i is the firm and t is the year. η_i is the unobservable firm-specific effect and ε_(i,t) is the error term. STI is a proxy variable for a firm’s innovation while ∑year is the year fixed effect.

4.2. GMM Regression Model

Considering that the STI capability may be characterized by continuity and dynamism, the lagged terms of the explanatory variables were included in the econometric model to avoid possible errors caused by omitted variables and to address the estimation bias caused by endogeneity issues.

$$ST{I}_{i,t}={\psi }_0+{\psi }_{1}ST{I}_{i-1,t}+{\psi }_{2}SI{T}_{i-2,t}+{\psi }_{3}subsid{y}_{i,t}+{\phi }_4{Z}_{i,t}+{{\displaystyle \sum }}^{ }year+{\eta }_i+{\epsilon }_{i,t}$$

(2)

The regression model used both differential GMM and systematic GMM, where the lagged term of the explanatory ψ₁STI_(i−1,_t) is set as the endogenous variable, the exogenous variable is subsidy, and the rest of the control variables are predetermined. The estimation results need to be tested after use, and usually the differences of the random error term are tested for the existence of first-order and second-order autocorrelation to ensure the consistent estimation of GMM. In general, the random error term differences will have first-order autocorrelation, because the lagged terms of the explanatory variables are included in the model, but there is no second-order autocorrelation, and the estimation results are judged by the significance of AR(1) and AR(2). In addition, a Hansen test is required to determine whether the instrumental variables used are valid.

5. Empirical Analysis

5.1. Panel Framework Tests

Panel A and panel B of

Table 2. Descriptive statistics (variables in natural logarithm).

Panel A: Sample Characteristics—NEV Firms
	STI	Subsidy	ROA	LEV	Payouts	Equity Attributable	Herf	No. of Employees	Gender	Education
Mean	0.5155	12.9601	2.7509	3.6727	3.4055	15.0119	4.0732	7.4752	0.0408	3.7992
Std. Dev	0.1685	7.8159	0.6962	0.6999	0.6872	0.3263	0.2883	1.2994	0.1978	1.3516
Minimum	0.2350	0.0000	−1.1802	0.8855	0.8890	1.6958	2.1724	3.7612	0.0000	0.0000
Maximum	1.0000	21.4164	4.9451	15.3630	10.4839	16.0639	4.6051	12.3207	1.0000	8.0983
Sample Characteristics—Traditional Fuel Vehicle Firms
Mean	0.3909	9.0254	2.8110	3.7062	3.5574	11.9789	4.0968	7.6587	0.0409	3.3254
Std. Dev	0.0983	7.7565	0.5918	0.5388	0.6352	1.1235	.2572	1.2157	0.1981	1.5582
Minimum	0.2038	0.0000	−1.8127	0.9303	0.8902	5.3188	2.6012	1.6094	0.0000	0.0000
Maximum	1.0000	22.1095	7.6442	7.9880	7.2373	17.0736	4.6167	12.2900	1.0000	10.1821
Panel B: Correlation Between Independent Variables
		Subsidy	ROA	LEV	Payouts	Equity Attributable	Herf	No. of Employees	Gender	Education
Subsidy		1.0000
ROA		0.0362 **	1.0000
LEV		0.0033	−0.0515 ***	1.0000
Payouts		0.0122	−0.0149	−0.1200 ***	1.0000
Equity Attributable		−0.3746 ***	−0.0436 ***	−0.0088	−0.1083 ***	1.0000
Herf		0.0396 **	0.1371 ***	−0.1449 ***	−0.0049	−0.0078	1.0000
No. of Employee		0.1620 ***	−0.0586 ***	0.3178 ***	−0.0424 ***	0.2253 ***	−0.0680 ***	1.0000
Gender		0.0182	−0.0296 **	−0.0488***	0.0144	−0.0216	−0.0090	−0.0507	1.0000
Education		−0.0132	−0.0796 ***	0.2211 ***	−0.0760 ***	0.4429 ***	−0.1116 ***	0.6573	−0.0696 ***	1.0000
Panel C: Multicollinearity
		Subsidy	ROA	LEV	Payouts	Equity Attributable	Herfindahl Index	No. of Employees	Gender	Education
Variance inflation factor		1.02	1.03	1.13	1.02	1.07	1.04	1.95	1.03	1.84
Tolerance		0.9834	0.9662	0.8847	0.9334	0.9141	0.9636	0.5123	0.9664	0.5438

Note: ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively, same below Tables.

provide descriptive statistics and correlation matrix of NEV firms and traditional fuel vehicle firms, respectively. Comparing the mean values of STI capability (0.5155 and 0.3909), we initially concluded that the STI capability of NEV firms is greater than traditional fuel vehicle firms; that there are higher public subsidies for the NEV firms, the average value is 12.9601 and 9.0254; that the mean value of ROA is slightly higher for traditional fuel vehicle firms, as it has a significant advantage at the maximum value; and that there is a higher education level of employees in NEV firms. A simple statistical description of the data can confirm the status of the NEV and traditional fuel vehicle firms above. Since NEV firms’ are in an early stage of industry, and since policies are more favorable to them, they have more high-tech personnel with advanced educations but have limited profitability. Traditional fuel vehicle firms are in a more stable state and enjoy stronger profitability.

Panel B of Table 2 corresponds to the bivariate correlations between dependent, independent, and control variables. The coefficients suggest the absence of multicollinearity problems. To check the characteristics of the data, we conducted various panel framework tests. Performing a multicollinearity test to check whether or not the independent variables are correlated. The results of panel C of Table 2 demonstrate that the data do not display multicollinearity, all independent variables meet the criteria that the variance inflation factor is less than 10, and that the tolerance value is over 0.1.

5.2. Panel Estimator and Empirical Results

5.2.1. Firms’ External Factor

The fixed-effect estimates deal with unobserved heterogeneity and potential endogeneity in a certain extent. The result of the Hausman test (p = 0.001) indicated the fixed effects regression model is better than the random effect estimation in our study. Therefore, we used the fixed effects regression model and the clustering robust standard error to reduce the interference of heteroscedasticity and improve the robustness of the regression results. The results of our multi regression analysis are presented in

Table 3. Fixed-effect regressions for external factors.

Independent Variables	Dependent Variable STIit
	Panel A: Fixed-Effect Regressions			Panel B: Group Test for Regional Factors
	(1) NEV	(2) Traditional Fuel Vehicle	(3) Automobile Industry	(4) First-Tier Cities	(5) Non-First-Tier Cities	(6) First-Tier Cities	(7) Non-First-Tier Cities
Subsidy	0.0024 *** (0.0005)	0.0009 *** (0.0002)	0.0023 *** (0.0005)	0.0038 *** (0.0009)	0.0016 ** (0.0007)	0.0017 *** (0.0006)	0.0006 *** (0.0002)
ROA	0.0051 (0.0042)	0.0054 *** (0.0018)	0.0047 * (0.0032)	−0.0028 (0.0083)	0.0008 (0.0058)	0.0102 ** (0.0039)	0.0038 * (0.0020)
LEV	0.0059 (0.0052)	−0.0038 (0.0048)	0.0051 (0.0046)	−0.0131 * (0.0224)	0.0073 (0.0049)	−0.0043 (0.0090)	−0.0025 (0.0053)
Payouts	0.0049 * (0.0054)	0.0066 *** (0.0019)	0.0034 ** (0.0035)	−0.0056 (0.0126)	0.0083 (0.0061)	0.0088 (0.0059)	0.0054 *** (0.0019)
Equity Attributable	−0.0099 *** (0.0027)	−0.0129 *** (0.0030)	−0.0279 *** (0.0061)	0.0293 ** (0.0721)	−0.0086 *** (0.0032)	−0.0135 ** (0.0056)	−0.0116 *** (0.0032)
Herf	0.0513 ** (0.0206)	0.0063 (0.0091)	0.0491 *** (0.0150)	0.0414 (0.0308)	0.0540 ** (0.0252)	0.0002 (0.0302)	−0.0074 (0.0080)
No. of Employees	−0.0132 *** (0.0038)	−0.0307 *** (0.0037)	−0.0157 *** (0.0037)	−0.0103 * (0.0094)	−0.0207 *** (0.0070)	−0.0172 *** (0.0042)	−0.0388 *** (0.0043)
Education	0.0031 (0.00489)	−0.0089 *** (0.0020)	−0.0041 (0.0030)	0.0101 (0.0106)	0.0104 * (0.0058)	−0.0156 *** (0.0047)	−0.0060 *** (0.0020)
Gender	0.01241 (0.0338)	0.0018 (0.0109)	0.0035 (0.0166)	0.0012 (0.0511)	0.0322 (0.0495)	0.0196 (0.0235)	0.0019 (0.0135)
REDL	0.1107 *** (0.0127)	0.0081 * (0.0043)	0.0817 *** (0.0094)	−0.0156 (1.1177)	0.4685 (0.1200)	0.7045 (0.1808)	0.8440 (0.0607)
_cons	−0.4572 ** (0.2081)	0.7183 *** (0.0814)	0.2686 *** (0.0947)
Individual effect	controlled	controlled	controlled	controlled	controlled	controlled	controlled
Time effect	controlled	controlled	controlled	controlled	controlled	controlled	controlled
F (N (0,1))	13.86	27.46	47.08	3.43	4.15	24.35	57.56
Observations	1918	2882	4818	660	1257	638	2244

Note: t statistics in parentheses, same below.

, panel A.

The three groups of panel A showed the effect of variables on SIT capability in a quantitative manner using public subsidy as the dependent variable in the three groups at the same time. According to the empirical results, public subsidies had a positive significant effect on STI capability, and the coefficient of NEV firms was comparatively higher than the traditional fuel vehicle firms; hence, hypothesis H1 is proved. In analyzing the possible reasons for this, NEV firms are able to utilize public subsidies in a more efficient way at their early stage of industrial development, with public financial support external to the firms and with sufficient highly technical talent inside, there is a competitive advantage for NEV firms that equips them with the ability to convert their invested physical capital and human capital into innovative physical outputs. Because the main profit source of the automobile industry is still the traditional enterprises, coupled with the investment risk brought by the technological change, enterprises are inclined to be opposed to new energy technology system reform. Although public subsidies have performed an irreplaceable role in improving R&D output for NEV firms, it cannot be denied that there still have been inevitable drawbacks during the subsidy policy [32]. These drawbacks include how public subsidies can crowd out an enterprise’s original investment [33], and how some enterprises will send false signals, leading to the adverse choice of public subsidies [34]. Subsidy optimization mechanisms have also been much discussed and include the regulation of subsidies and establish a carbon emissions trading market that is beneficial to NEV firms. This can become an effective way to promote the innovation of NEV firms by establishing a more reasonable institutional system in the practice.

5.2.2. Firms’ External Factor: Financial Performance

For NEVs, we found limited but substantial support for Hypothesis 2 in

Table 4. Fixed-effect regressions for internal factors.

Dependent Variable	Dependent Variable STIit
	Panel A New Energy Vehicle Firms			Panel B Traditional Fuel Vehicle Firms
	(1)	(2)	(3)	(4)	(5)	(6)
ROA	0.0092 ** (0.0043)		0.0074 * (0.0043)	0.0097 *** (0.0019)		0.0054 *** (0.0018)
LEV	−0.0025 (0.0074)		0.0055 (0.0056)	−0.0251 *** (0.0042)		−0.0054 (0.0048)
Payouts	0.0052 (0.0052)		0.0042 (0.0052)	0.0036 * (0.0020)		0.0055 *** (0.0020)
Equity Attributable	−0.0187 ** (−0.0187)		−0.0103 *** (0.0029)	−0.0414 *** (0.0025)		−0.0134 *** (0.0030)
Herf		0.0655 *** (0.0215)	0.0680 *** (0.0213)		−0.0042 (0.0090)	0.0005 (0.0087)
No. of Employees		−0.0141 *** (0.0047)	−0.0137 *** (0.0048)		−0.0360 *** (0.0036)	−0.0300 *** (0.0037)
Education		0.0060 * (0.0049)	0.0057 (0.0048)		−0.0135 *** (0.0017)	−0.0099 *** (0.0020)
Gender		0.0175 (0.0355)	0.0108 (0.0335)		0.0010 (0.0105)	0.0005 (0.0106)
REDL	0.0984 *** (0.0984)	0.1090 *** (0.0133)	0.1106 *** (0.0133)	0.0050 (0.0045)	0.0040 (0.0042)	0.0070 (0.0043)
_cons	−0.2780 (0.1939)	−0.7764 *** (0.1549)	−0.7074 *** (0.1676)	0.8880 *** (0.0596)	0.6872 *** (0.0721)	0.7246 *** (0.0803)
Individual effect	controlled	controlled	controlled	controlled	controlled	controlled
Time effect	controlled	controlled	controlled	controlled	controlled	controlled
F (N (0,1))	13.98	18.21	28.36	69.11	105.10	48.28
Observations	1918	1918	1918	2882	2882	2882

panel A, as the independent variables ROA and Payouts both had a positive sign for STI capability, but only the ROA was at a significant level. This indicates that companies with stronger profitability are also capable of investing and creating more wealth under the same conditions. The empirical results for traditional fuel vehicle firms are presented in Table 4 panel B, ROA and Payouts had a significant positive impact, and the leverage ratio had a significant negative impact (Hypothesis 4). Firms with a high leverage ratio are not always a good thing, because the creditors prefer a low leverage ratio, as firms with a low leverage ratio are less risky. Furthermore, firms with higher leverage need to pay more interest, and their investment capacity would be limited, it may also mean their need to reining overinvestment. From the above conclusion, we found that firms in a healthier financial situation leads to a higher STI capability. Moreover, different forms of public subsidies have different effects on financial performance; relevant government agencies should further refine their subsidy policies for the NEV industry and provide the appropriate form of subsidies based on an NEV firm’s conditions, something which benefits corporate financial performance in a healthy state.

5.2.3. Firms’ External Factor: Firm Characteristics

From the results in Table 4, we found that the Herf had a positive significant effect on NEV firms (φ_1 = 0.0655, p < 0.001) (Hypothesis 3) and a negative effect on traditional fuel vehicle firms at a non-significant level (Hypothesis 4); the empirical findings of Model 3 suggest that ownership concentration is beneficial to emerging firms’ innovative behavior. Dominant shareholders may exert greater power, limiting managerial discretion, and actively influencing management [27], which can bring about serious agency problems. However, from the other side, dominant shareholders can impose their view and use their influence positively by promoting value-enhancing projects. Due to the high risk of innovative activities, high ownership concentration can address situations that shareholders may give up innovation because they value short-term gains. Instead, low ownership concentration can balances the interests of dispersed shareholders, thereby moderating their personal influence on corporate decisions. Education has a significant positive effect on the STI capability of NEV firms. The number of employees can be used as a proxy variable for firm size under certain circumstances, and its impact on STI capability has shown that large firms with public subsidies are more motivated to engage in R&D than smaller ones; though it has also been argued that large firms lead to R&D innovation activities being inefficient. The conclusions obtained in this paper are more supportive of the latter.

5.3. Heterogeneity Test

The results for the control variables provide additional insights. The support policies for NEV firms varied according to economic, social, and environmental factors. For each regression, we controlled the economic development of the firm’s geographic location; the empirical results are show that it significantly affected three regression models of NEV firms. However, the innovation effect due to the geographical location of the enterprise is not obvious in the traditional fuel vehicle model. Thus, in industries already at a mature stage, the geographic location is not important enough to affect its technological innovation capabilities. On the contrary, the new industries such as the NEV industry which need the local government policy support, it is more conducive to the sustainable development when it is located in a areas with higher levels of economic development.

Further, we examined the influence of public subsidies on STI capabilities for firms in different geographic locations by distinguishing whether the firm was in a first-tier city or not, reports the test results in Table 3 panel B. In this paper, the first-tier cities included Beijing, Shanghai, and Shenzhen, and the listed firms were geographically located in the above three cities as one category, with the remaining cities as another category. This article categorized NEV and traditional fuel vehicle firms by whether they were located in first-tier cities. The empirical results show that all four subgroups of firms had a significant positive effect of public subsidies on their STI capabilities. However, we obtain that public subsidies had the greatest impact on the STI capability of NEV firms in first-tier cities from the coefficients of the model, which was higher than the NEV firms in non-first-tier cities and also higher than the traditional fuel vehicle firms in first-tier cities. The conclusion above is proved again.

5.4. Robustness Check

To exclude the interference of endogeneity in the regression results, this study adopted the generalized method of moments model to test the empirical results. The results in

Table 5. GMM regressions.

Dependent Variable	Dependent Variable STIit
	Panel A New Energy Vehicle Firms		Panel B Traditional Automobile Firms
	(1) Difference_GMM	(2) System_GMM	(3) Difference_GMM	(4) System_GMM
STIi-1,t	0.3657 *** (0.0680)	0.4942 *** (0.0353)	0.2109 *** (0.0555)	0.5043 *** (0.0729)
STIi-2,t	0.1705 ** (0.0624)	0.2582 *** (0.0381)	0.0084 (0.0369)	0.0929 *** (0.0280)
Subsidyi-1,t	0.0025 *** (0.0015)	0.0013 * (0.0010)	0.0094 *** (0.0022)	0.0075 *** (0.0022)
Subsidyi-2,t	0.0033 * (0.0016)	0.0017 (0.0012)	−0.0041 * (0.0023)	−0.0054 * (0.0031)
ROA	0.0011 (0.0096)	0.0105 (0.0087)	0.0041 (0.0055)	0.0052 (0.0061)
LEV	0.0215 (0.0168)	−0.0082 (0.0126)	0.0044 * (0.0044)	−0.0003 (0.0083)
Payouts	−0.0222 (0.0172)	−0.0156 (0.0105)	0.0087 (0.0087)	0.0095 * (0.0050)
Equity Attributable	−0.0390 (0.0304)	−0.0026 (0.0150)	0.0008 (0.0068)	−0.0056 (0.0061)
Herf	0.0776 * (0.0448)	0.0761 ** (0.0307)	0.0346 * (0.0197)	0.0116 (0.0178)
No. of Employees	0.0042 *** (0.0189)	0.0142 (0.0091)	−0.0303 *** (0.0109)	−0.0023 (0.0096)
Education	0.0029 (0.0109)	0.0110 * (0.0065)	−0.0036 (0.0031)	−0.0083 ** (0.0040)
Gender	−0.0007 (0.0478)	−0.0396 ** (0.0174)	0.0012 (0.0273)	−0.0020 (0.0093)
REDL	0.0038 (0.0265)	0.0010 (0.0062)	0.0147 (0.0114)	0.0168 *** (0.0051)
AR(1)	0.000	0.000	0.000	0.000
AR(2)	0.588	0.993	0.619	0.630
Sagan Test	109.15	166.43	284.33	323.65
Hansen Test	110.94	157.54	162.56	161.19

are consistent with the empirical results in Table 3 and Table 4, indicating that there was no serious endogeneity problem.

6. Conclusions

We explored the connection between firms’ technological innovation capabilities and their internal and external factors. To empirically test this relationship, we used panel data for the NEV firms and traditional fuel vehicle firms in China from 2010 to 2020 and collected data on 439 listed firms in the automotive industry. Based on the empirical analyses of this study, the following main results and implications can be drawn. Public subsidies do have a positive impact on firms’ STI capability; there are consistent findings for both types of firms. NEV firms receive higher public subsidies than traditional fuel vehicles and have a higher capacity for technological innovation. The NEV firms are able to convert public subsidies into R&D output in a more efficient way. Further, regarding financial performance, having high profitability and low leverage favors the ability to innovate capably for both types of firms. Regarding a firm’s characteristics, and in contrast to traditional fuel vehicle firms, high shareholder concentration and highly educated employees make positive contributions to an NEV firm’s innovation. This paper is not only of academic significance to the development of NEV firms to improve their technological innovation capability and the transformation of traditional fuel vehicle firms, but also of practical significance to the reduction of greenhouse gas emissions and the achievement of the “double carbon” goal.

First, the study found convincing evidence of a positive bidirectional relationship between R&D public subsidy and firms’ innovation. However, different from previous research, we also found that public subsidies have a more positive effect on promoting the STI capability of NEV firms than traditional fuel vehicle firms in the automotive industry. Thus, traditional fuel vehicle firms need to maintain their innovation dynamics through endogenous development. Most of the literature and reports have advocated for the direct elimination of subsidies for NEVs. We posit that this course of action is unadvisable [28]. The conclusion suggests the importance of persistent commitment to dynamically and efficiently implementing R&D policy, the implementation of a slow decline of the incentive policy is in line with the law of industrial development and an effective solution to the subsidy extrusion R&D. The relevant government agencies should further refine their subsidy policies for the NEV industry and provide the appropriate form of subsidies based on the conditions of NEV firms. Meanwhile, the enterprises should broaden their financing channels. The stock market, as a key source for listed firms to raise money [35], is particularly important for ensuring NEV firms have access to adequate capital to keep their business running and manage the consequences of shocks due to the reducing subsidies [36].

Next, we found consistent effects of corporate profitability and of equipping for the promotion of innovation capabilities of NEV firms versus traditional fuel vehicle firms, with high equity concentration and employee education level having a positive impact on the STI capability of NEV firms. However, the innovation process is not independent and comprises multiple divisional and relational activities; the success of systemic innovation requires the industry as a whole to engage in a transition [37]. Unlike previous studies, we analyzed the development of two types of firms in the automotive industry in comparison. The conclusions confirm that NEV and traditional fuel vehicle firms are at different stages, which have different characteristics. NEV firms are in an early stage of industrial development with limited profitability; support from public subsidies provides the advantages of science and technology innovation. Managers concentrate more on the long-term sustainability of the firm, stemming from the high concentration of equity, thus facilitating the firms to carry out STI activities, as corporate economic growth relies on highly technical and highly educated staff. The traditional fuel vehicle firms are in a mature stage of industry with high profitability, lower STI capability, and have balanced and distributed equity interests. Also, for both types of firms, the results imply that they need to maximize their profitability by implementing appropriate strategies, as well as by maintaining an adequate leverage ratio suited for their own conditions. The NEV firms should be aware of reserve financial flexibility [7] while holding excess cash and maintaining a low leverage to obtain financial flexibility. Traditional fuel vehicle firms should make full use of their own capital advantages to improve their innovation capabilities in R&D output by engaging in R&D and innovation activities, introducing and purchasing more advanced production technologies, and achieving low-carbon and environmentally friendly development [38].

7. Contributions, Limitation, Future Research

Innovation essentially entails a process of coevolution that features systematic characteristics. The automobile industry is a complex system comprising various interrelated elements; it is of practical significance for the automobile industry to integrate resources and collaborate in innovation development. Synergistic development in the automotive industry can contribute to an improvement in the STI capacity of NEV firms and to the sustainable transition from traditional fuel vehicles. Theoretically, by integrating research on the transformation and development of the automotive industry, can deepen our understanding of how state and market actors can attempt to drive sociotechnical change.

There were limitations in the data collection and measurement methods used. In the study of synergistic development and interconnectedness among things, it may be worthwhile to further optimize the measurement methods to adapt the study content, and further methodological innovations are still being explored. Based on the current study data, we hope to explore in with alternative statistical learning methods in future steps. One possibility is to modify the combined methods, where we analyze for latent clusters and relationships with stability and reproducibility [39]. Alternatively, we may employ grey models for short-term predictive modeling of firms’ technological innovation capabilities with limited samples [40]. Furthermore, Future research could be in a more specific direction, studying the business model of the innovation of NEV firms and the business model of the innovation of traditional fuel vehicle firms and how the two types of firms may cooperate. We will use scenario simulation to investigate what results may be generated under various modes of cooperation and whether there is any potential for firms’ technological innovation capabilities in the process.