The Effect of Outward FDI on Capabilities of Sustained Innovation: Evidence from China

Abstract

Based on theoretical analysis, this paper examines the impact of China’s outward foreign direct investment on regional capabilities of sustainable innovation. Empirical tests are used to analyze the mechanism of OFDI in regional capabilities of sustainable innovation. OFDI’s impact on sustainable innovation is tested through the fixed effect, mediation effect, and threshold effect models by using provincial panel data from 30 provinces in China from 2008 to 2019. The results show that OFDI promotes regional capabilities of sustained innovation. Moreover, the effect of OFDI on sustainable innovation is found to be mediated by regional human capital accumulation. Furthermore, this study also finds that regional absorptive capacity and financial market development are two significant threshold variables in the process by which OFDI affects sustainable innovation. Before the threshold variables exceed the threshold level, OFDI encourages sustainable innovation, but this positive association becomes stronger after threshold variables reach the threshold level. The results have strong policy implications, which will be useful for local governments to make policies with the aim to maximize the benefits of OFDI and encourage sustainable innovation.

1. Introduction

Since China joined the World Trade Organization (WTO), China has gradually emphasized encouraging domestic enterprises to “go global”. The scale of outward foreign direct investment (OFDI) has also expanded to many fields and regions. In 2019, the scale of China’s non-financial foreign investment reached USD 139.11 billion. Compared with 2003, the scale of foreign investment has increased by nearly 50 times. In 2020, China’s foreign investment flow was USD 153.71 billion, ranking first in the world for the first time, with a year-on-year increase of 12.3%.

At the same time, China constantly emphasizes the important role of innovation and regards innovation as a development strategy to enhance the country’s comprehensive strength. As a part of China’s overall innovation capability, regional innovation is an important component of international overall innovation, and it has also become an important factor to promote regional economic development [1]. In such a context, innovation capabilities will be continuously optimized and improved in the region [2] and are needed to achieve sustainable economic development [3]. Against the background of global economic integration, regional innovation is also influenced by international capital flows [4], such as OFDI [5]. Thus, the regional capability of sustained innovation is one of the essential factors of regional sustainable and stable development. Therefore, this paper will explore the influence of OFDI on China’s regional capability of sustainable innovation.

After Hymer put forward the monopolistic advantage theory to explain the motivations of foreign investment in 1960, the impact of OFDI on the economic performance of home countries has been discussed by scholars theoretically since the 1960s. Multinational companies transfer technologies through foreign investment, and this process affect the external economy in the home and host countries. The theory of the technology spillover effect describes the positive externality caused by foreign investment. In the 1990s, researchers found that developing countries introduce advanced technologies through outward foreign investment activities, thus accumulating knowledge, and then acquiring higher-level technologies. Based on the model of R&D spillover by Coe and Helpman [6], Lichtenberg and Potterie expanded the theoretical model to describe the technology spillover effect of OFDI [7]. The existence of reverse technology spillover of OFDI was examined in empirical studies [8,9]. Based on the reverse spillover effect of OFDI, researchers began to empirically study how OFDI affects economic performance in home countries. The related empirical research on the OFDI reverse spillover effect began with the empirical study of Japan’s direct investment in the United States, and it was found that investment was mainly concentrated in industries with high R&D density in both countries [10]. With the reverse spillover effect, OFDI can enhance domestic employment [11], total factor productivity [12], and economic growth [13] in home countries. Researchers also found that OFDI generates more technological opportunities to promote innovation in emerging countries [14,15].

With the rapid growth of China’s foreign direct investment, a large number of studies have empirically tested the impact of OFDI on regional innovation by using panel data at the provincial level in China. Based on the results of empirical research in China, the existing research is divided into (i) promotion theory, (ii) non-significant or negative effect theory, and (iii) nonlinear “threshold effect” theory. (i) Promotion theory: OFDI significantly improved domestic innovation efficiency [16], capacities [17], and performance [18]. (ii) Non-significant or negative effect theory: Different from the above conclusions, the conclusions of some existing empirical studies show that OFDI has no significant impact on domestic innovation in the home country [19,20]. (iii) Nonlinear “threshold effect” theory: With the deepening of research, some scholars have realized that OFDI’s reverse spillover effect is affected by the absorptive capacity of the home country [21,22,23], which may show a “threshold effect”. Whether the parent company can effectively absorb and transform the foreign technical knowledge transferred from overseas subsidiaries is the key. Scholars set a variety of threshold variables, including workers’ education level, opening level, R&D intensity, market openness, and so on [24]. These studies provide theoretical support for the reverse spillover effect of OFDI on regional innovation in China.

The effect of OFDI on regional innovation has been discussed by scholars theoretically. However, few studies have empirically examined the impact of OFDI on regional sustainable innovation [25]. The effect mechanism of OFDI on sustainable innovation is rarely discussed in existing studies. In this paper, we attempt to fill this gap by empirically examining the effect of OFDI on regional sustainable innovation. Furthermore, the mediating effect of human capital accumulation is analyzed. In addition, this paper discusses impact factors on the relationship between OFDI and regional sustainable innovation by examining the threshold effect. The threshold variables include regional absorptive capacity and the development of the financial market. Through the exploration of these problems, this paper expands the theoretical framework of the OFDI reverse spillover effect and helps to reveal the complex relationship between OFDI and domestic sustainable innovation capability.

Based on the existing research, this paper contributes to the existing literature in several ways. First, this paper makes an in-depth analysis on the mechanism of the reverse spillover effect of OFDI on regional sustainable innovation. Previous studies theoretically analyze the effect of OFDI on regional innovation; this paper extends it to regional sustainable innovation. It supplements to extend the existing literature on the reverse spillover effect of OFDI. Second, it provides a mechanism analysis on the impact of OFDI on regional sustainable innovation with the test on mediating effect. Previous studies provide a theoretical analysis of the effect mechanism of OFDI on regional sustainable innovation. This study empirically analyzes the mechanism of the effect of OFDI on regional sustainable innovation from the perspective of human capital. The findings of mediating effect enrich the research on the mechanism of OFDI affecting sustainable innovation. Third, it adds to a better understanding of the relationship between OFDI and innovation by introducing the impact factors of the relationship. While previous studies obtained inconsistent results on the effects of OFDI on regional innovation, these may be caused by the neglect of the impact factors on their results. Our study discusses whether OFDI’s impact on regional sustainable innovation has a threshold effect on regional absorptive capacity and the development of the financial market. It helps to provide a better understanding of the relationship between OFDI and regional sustainable innovation. This paper expands the theoretical framework of the OFDI spillover effect and helps to reveal the complex relationship between OFDI and regional sustainable innovation. It also provides a reference for the government’s macro-level policymaking. Therefore, it is of theoretical and practical significance to explore the influence of OFDI on China’s sustainable innovation ability and analyze its mechanism.

The remainder of the paper is organized as follows. The second part is the theoretical analysis; the third part is the model construction and variable description; the fourth part is the empirical analysis; the fifth part discusses the empirical results; the last part concludes the results, contributions, policy recommendations, and suggestions on future work.

2. Theoretical Analysis

2.1. OFDI’s Effect on Regional Sustainable Innovation Capability

The motives of the home country’s direct investment in the host country mainly include expanding the international market, introducing advanced technology, and obtaining local resources. The mechanism of OFDI reverse technology spillover to promote regional technological innovation can be summarized into four main mechanisms: R&D resource sharing mechanism, knowledge transfer mechanism, market feedback mechanism, and peripheral stripping mechanism.

According to the resource-based theory, the sharing mechanism of R&D resources is mainly subdivided according to multinational enterprises’ foreign investment, which can increase innovation output by increasing joint technology, R&D cost sharing, and worker mobility mechanism. One of the most common channels of R&D resource sharing is that multinational enterprises enter the host country through the establishment of branches, or through joint ventures, strategic alliances, and mergers and acquisitions, etc. In the host country, they engage in joint technology R&D integration or joint R&D centers with local enterprises, cooperate to develop new products and other activities, and share the R&D resources of both parties in the process. Among them, China’s resource-seeking investment in developing countries is conducive to optimizing the efficiency of resource allocation and reducing production and operation costs. Moreover, the savings in R&D expenses reduced by resource sharing can be used for other innovative activities such as the technology development of the parent company. Second, international investment is beneficial for the cooperative enterprises in the host country to share a certain degree of R&D expenses, so as to ease the R&D financing constraints within the home country enterprises, and thus have sufficient funds to invest in R&D activities. Third, multinational corporations and technology enterprises in developed host countries jointly train R&D personnel and employ highly skilled talents to make use of the working skills of talents in host countries to innovate, so as to improve the innovation ability of subsidiaries and overseas institutions. Technical personnel working in overseas R&D centers and branches of multinational corporations can make use of the superior R&D conditions of the host country to continuously improve their technical level and communicate with other employees working in China to realize the reverse spillover of knowledge. By sending local employees to overseas branches for training, knowledge transfer is realized through personnel exchanges, and knowledge is introduced to domestic parent companies.

Knowledge transfer mechanisms mainly include technology trading and imitation and transmission of advanced technology. Among them, technology trading refers to the fact that domestic enterprises set up overseas branches or subsidiaries through foreign direct investment, which can help the home country enterprises to blend with developed countries more conveniently in terms of technology and management knowledge. For example, after an enterprise enters the international market through cross-border M&A, it may directly obtain the R&D capital of the acquired enterprise, thus increasing the R&D elements of multinational enterprises. The advanced technology transfer mechanism stems from the fact that enterprises embed into the value chain of the host country through strategic alliances and other ways, thus absorbing and imitating the technology and management experience of the host country’s enterprises and spreading and imitating these technologies. Especially, most of the direct investments in developed countries have strong technical intentions and promote the technological progress of the home country through technology acquisition.

Market feedback mechanisms mainly include international market competition, income feedback, and consumer feedback. First of all, the foreign direct investment of multinational corporations in the host country’s market will be affected by the competitive pressure of the international market. Especially, when multinational companies invest abroad in the form of greenfield investment, they are affected by different factors such as the foreign market environment and institutional environment, which leads to competitive pressure. The competitive pressure in the host market may prompt multinational companies to increase R&D investment to better cope with the production and operation environment of the host country. Second, direct investment in developing countries mainly strengthens the independent innovation capability of the parent company through market income feedback. Through OFDI, enterprises can expand overseas markets, increase their sales revenue, and thus reduce the cost of production and sales. The increase in enterprises’ benefits can prompt more R&D investment, so as to further enhance the technological capability and absorption level of enterprises and enhance their independent innovation capability. Specifically, China’s market-seeking OFDI can increase the market share and sales income of enterprises in the host country by exporting part of the surplus production capacity in the domestic market, thus producing economies of scale and reducing the R&D cost per unit product, and then feeding the increased profits back to the home country to make up for the expenses of the home country’s technology research and development. Third, overseas subsidiaries and R&D institutions carry out targeted innovative R&D in order to adapt to local consumers’ preferences. Feedback from consumers in the host country can promote investment enterprises to continuously improve the performance of consumer goods, and the innovation level of enterprises will be enhanced through continuous improvement.

Peripheral technology divestiture means that enterprises transfer peripheral R&D technologies to other countries through OFDI, including products of non-core technologies, etc., so as to improve the allocation efficiency of domestic innovation resources and enable home enterprises to concentrate their core resources, such as manpower and financial resources, on research and development of core technologies, thus increasing domestic technological innovation.

In the host country, enterprises have promoted technological innovation through channels such as R&D resource sharing mechanisms, knowledge transfer mechanisms, market feedback mechanisms, and peripheral stripping mechanisms, so as to reverse “spillover” to the parent company to improve the R&D activities and R&D level of multinational companies and enhance the technological innovation capability of enterprises. The advanced technology acquired by the parent company of multinational companies will be further transferred to other companies in the same industry in the home country through the demonstration effect and competition effect. At the same time, reverse technology spillovers will spread to other industries by transferring to other enterprises in related industries such as upstream and downstream enterprises. By virtue of the advanced technology and innovative elements acquired by foreign direct investment activities, enterprises in the industry and related industries are encouraged to make full use of the factor resources in the region through the “demonstration effect” and “competition effect”. With the increasing liquidity and agglomeration capacity of innovation R&D capital in the region, regional innovation will be increased through the diffusion effect. Then, it will stimulate the effect mechanism of “winners are more successful”; that is, the previous innovation gains further promote the later-stage innovation, and further reach the regional sustainable innovation. Figure 1 explains the effect mechanism of OFDI on sustainable innovation.

2.2. The Mediating Effect of Human Capital Accumulation

In order to further understand the effect of OFDI in China on regional sustainable innovation capability, it is necessary to theoretically analyze its mechanism. The regional human capital is the comprehensive embodiment of local ability, knowledge, and technology. No matter whether the absorption of foreign advanced knowledge and technology or domestic independent research and development, it is inseparable from the investment of high-tech human capital. By constructing the technology spillover model of foreign direct investment, multinational corporations use advanced technology in the host country after training local workers, and the workers will have a spillover effect when they are employed by the host domestic company [26]. This paper expounds on the mechanism of technology spillover caused by transnational investment through personnel flow. By learning the knowledge and experience of the host country, the outward foreign direct investment provides opportunities for domestic employees to receive further training, thus spilling knowledge and technology back to their home countries. In neoclassical economics and new growth economics, human capital is regarded as an important factor of economic growth, and human capital is not completely exclusive, so it has a “spillover effect”. Human capital is regarded as the key factor to promote technological innovation because of its increasing marginal return. For example, human capital plays an important role in regional patent applications in the research of manufacturing enterprises in China [27]. Therefore, in theory, foreign direct investment affects domestic economic activities by increasing regional human capital accumulation. Based on the abovementioned analysis of foreign direct investment, human capital accumulation, and regional innovation, this paper holds that OFDI can enhance its sustainable innovation capability by enhancing the accumulation of regional human capital.

2.3. The Threshold Effect of OFDI on Sustainable Innovation Ability

2.3.1. Regional Absorptive Capacity

The concept of absorptive capacity was first put forward in the 1960s, which refers to the ability of an economy to absorb and utilize external information and resources. Absorptive capacity is a dynamic process of organization and practice, and its connotation includes knowledge acquisition, absorption, digestion, conversion, and utilization. Regional absorptive capacity refers to the ability of a region to identify external information and use it to transform it into its own development. Due to the technological gap between the home country and the host country, absorptive capacity is the decisive factor that affects the digestion and utilization of reverse technology spillovers in the home country. The theory of absorptive capacity originates from Cohen and Levinthal’s firm-level analysis, which points out that absorptive capacity is endogenous and is an additional learning effect produced in R&D activities. On the basis of the absorptive capacity theory, enterprises need to recognize and understand new external knowledge technologies, transform them, and create commercial value for the acquired knowledge. Based on these concepts of enterprise-level absorptive capacity, regional absorptive capacity is understood as the ability of regional acquisition, transformation, and utilization of new knowledge. Regional innovation also depends on digesting and absorbing foreign superior resources. It is necessary to have a certain absorptive capacity to absorb knowledge and technology from outside the region and transform it into its own innovative output. The absorptive capacity in the existing research includes regional human capital, infrastructure, etc. The comprehensive absorption capacity of a region depends on the level of economic development, education, and government support. When the absorption capacity of the home country is weak, it is difficult for the imported advanced technology to exert its positive externality. When the absorptive capacity reaches a certain threshold, the home country can re-innovate the acquired knowledge and technology and realize the regional innovation effect of OFDI reverse technology spillover. The stronger the absorptive capacity of a region, the stronger the absorptive capacity of technology and knowledge brought by foreign capital, and the greater the spillover effect of foreign investment [28,29]. The absorptive capacity of domestic enterprises and regions is the prerequisite to absorb the externalities of foreign direct investment when exploring the overall spillover effect of foreign direct investment on domestic enterprises [30]. The stronger the regional absorptive capacity, the easier it is to absorb and acquire external knowledge and transform the external knowledge and technology absorbed into the capability of independent innovation. If the regional absorptive capacity is weak, it will limit the ability to acquire external knowledge and technology, so that external inflow factors such as knowledge and technology cannot be transformed into innovation, hindering the regional innovation output. Therefore, OFDI’s impact on regional sustainable innovation capability may have a threshold effect on regional absorptive capacity.

2.3.2. Development of Regional Financial Market

The level of financial development will also affect the technology spillover effect of OFDI. The spillover effect of OFDI technology may depend on a sound financial system and a high degree of financial market development. Domestic enterprises need a lot of financial support when absorbing the knowledge spillover of international direct investment and carrying out technological innovation, including the cost of learning, purchasing technology and equipment, etc. Therefore, external financing can solve the problem of insufficient domestic savings. At this time, we need to rely on a sound financial system and the level of financial market development. Financial development is mainly through financing, venture capital, and innovative risk transfer mechanism to ensure that local enterprises absorb technology spillovers and carry out secondary development. The higher the financial development level of the host country, the more support it can provide for enterprise financing. This financing facility can attract more multinational companies to set up new enterprises in the host country, which will have a positive impact on the economic growth of the host country. Developed financial markets help multinational companies to obtain the required funds, and then increase R&D investment to promote technological innovation. The developed financial market helps OFDI to influence the host country enterprises to increase their R&D investment through the reverse technology spillover effect by alleviating the financing constraints of enterprises. Multinational companies need to invest in relevant equipment and hire R&D personnel to carry out technological innovation activities in the host country. A sound financial market system helps multinational companies to obtain the funds needed for enterprise R&D activities, thus promoting the spillover effect of international direct investment. The level of financial development affects the relationship between OFDI and technological innovation. OFDI’s influence on continuous innovation will also change with the level of financial development. When the development level of the regional financial market is low, enterprises cannot obtain enough funds from it to solve the problem of financing constraints, so it is difficult to exert OFDI’s promoting effect on sustainable innovation ability. Only when the enterprises in the region can source enough initial fixed investment funds from the financial market can they fully absorb and utilize OFDI reverse technology spillovers to improve their technical level. Therefore, OFDI has the threshold characteristics of financial market development level for regional sustainable innovation. Therefore, OFDI’s impact on regional sustainable innovation capability may have a threshold effect on the financial market development level.

3. Model Construction and Variable Description

3.1. Research Model

The empirical research on technology spillover of international direct investment mainly adopts the technology regression method represented by Coe and Helpman. First of all, it is necessary to find the proxy variables of technology by using the technology regression method. It is common to measure the technology spillover based on the input and output of technology activities. Scholars Lichtenberg and Potterie put forward the L-P model on the basis of revising the C-H model and combining the actual development of international direct investment, and for the first time, foreign direct investment was included in the model as an international technology spillover path. This study draws on the idea of the L-P model to construct the model of this study. The foreign direct investment amount is taken as the explanatory variable of the model, the regional sustainable innovation capability is taken as the explained variable, and the control variable is added to the model to carry out the research. To test the influence of OFDI on China’s regional innovation, this paper sets up Equation (1):

$$inno{v}_{it}={\beta }_0+{\beta }_{1}ofd{i}_{it}+{{\displaystyle \sum }}_k{\beta }_{k}contro{l}_{it}+{\epsilon }_{it}$$

(1)

In the equation, $i$ is the province and $t$ is the year. The explained variable $inno{v}_{it}$ is the sustainable innovation capacity of the provinces, and $ofd{i}_{it}$ is the explanatory variable of foreign direct investment. In addition, $contro{l}_{it}$ are the control variables and ${\epsilon }_{it}$ is the random disturbance term.

In order to further explore the mechanism of OFDI’s impact on regional sustainable innovation, the mediating effect model will be applied to the mechanism test. Based on the theoretical analysis of the effect mechanism of OFDI on regional sustainable innovation, this paper introduces the mediating variable of human capital accumulation. By constructing a mediating effect model, this paper empirically explores the transmission mechanism of the outward FDI affecting regional innovation. The total effect of OFDI on sustainable innovation can be divided into two parts to describe the transmission mechanism. One is the effect of OFDI on human capital accumulation. The other one is the effect of human capital accumulation on sustainable innovation. Based on the stepwise method, the empirical test of the mediating effect model is divided into three steps [31]. First, the explanatory variables are regressed to the explained variables, that is, refer to the Equation (1). The coefficient ${\beta }_1$ in Equation (1) is the direct effect of OFDI on sustainable innovation. Second, the explanatory variables are regressed to the human capital accumulation of mediating variables, as shown in Equation (2). The coefficient ${\theta }_1$ in Equation (2) measures the impact of OFDI on human capital accumulation. Third, the explanatory variables and the mediating variable $huma{n}_{it}$ are added to the regression analysis of the explained variables, as shown in Equation (3). The coefficient ${\gamma }_1$ is used to describe the impact of OFDI on sustainable innovation with controlling human capital accumulation. The coefficient ${\gamma }_2$ is used to describe the impact of human capital accumulation on sustainable innovation with controlling OFDI. If the coefficient ${\theta }_1$, ${\gamma }_2$ and ${\beta }_1$ are all significant, it implies that the mediating effect of human capital accumulation exists. If ${\gamma }_1$ is significant, the mechanism of human capital accumulation on the relationship between OFDI and sustainable innovation is the partial mediating effect. Otherwise, the mechanism of human capital accumulation is the full mediating effect if ${\gamma }_1$ is insignificant.

$$huma{n}_{it}={\theta }_0+{\theta }_{1}ofd{i}_{it}+{{\displaystyle \sum }}_k{\theta }_{k}contro{l}_{it}+{\epsilon }_{it}$$

(2)

$$inno{v}_{it}={\gamma }_0+{\gamma }_{1}ofd{i}_{it}+{\gamma }_{2}huma{n}_{it}+{{\displaystyle \sum }}_k{\gamma }_{k}contro{l}_{it}+{\epsilon }_{it}$$

(3)

As for the impact factors of the effect of OFDI on regional sustainable innovation, this paper mainly discusses the threshold effect. To test the threshold effect of the absorptive capacity and financial market development level on the impact of OFDI on the regional innovation level, a threshold model is constructed. As there may be multiple threshold effects in practical application, according to the threshold regression model proposed by Hansen [32], based on Equation (1), this paper establishes the threshold regression model of China’s OFDI on innovation, as shown in Equation (4).

$$\begin{gathered} inno{v}_{it}=a_0+a_{1}ofd{i}_{it}·I\left(q_{it}\le {\gamma }_1\right)+a_{2}ofd{i}_{it}·I\left({\gamma }_1<q_{it}\le {\gamma }_2\right)+\dots +a_{n}ofd{i}_{it}· \\ I\left({\gamma }_{n-1}<q_{it}\le {\gamma }_n\right)+a_{n+1}ofd{i}_{it}·I\left(q_{it}>{\gamma }_n\right)+{{\displaystyle \sum }}_k{a}_{k}contro{l}_{it}+{\epsilon }_{it} \end{gathered}$$

(4)

where $I(·)$ is the indicator function $q_{it}$ is the threshold variable. In the empirical study of this paper, the threshold variables are absorptive capacity and the level of financial market development. The threshold values are set to n, including ${\gamma }_1, {\gamma }_2,\dots , {\gamma }_n$.

3.2. Description of Variables

This paper selects the provincial panel data of China from 2008 to 2019 for the empirical test. Due to the data accessibility, our study includes 30 provinces in mainland China (except Tibet, Hong Kong, Macau, and Taiwan).

The data on outward foreign direct investment ($OFD{I}_{it}$) come from the foreign non-financial direct investment in the Statistical Bulletin of China Foreign Direct Investment issued by the Ministry of Commerce each year and the indexes are processed logarithmically. It is worth noting that the proposed treatment of the quantitative part of OFDI is to use the stock of external non-financial OFDI instead of the net investment flow of each year. After the implementation of the economic policy of “going global”, the scale of foreign investment has been continuously expanded. As the impact of OFDI is a continuous process, the outward OFDI in the early stage will also have an impact on future economic activities [33].

The data on regional sustainable innovation ($Inno{v}_{it}$) come from the annual patent applications of each province in the China Statistical Yearbook of Science and Technology. The number of patents applied by new technology inventors to local patent examination institutions is still an important way to intuitively reflect a region’s innovation capability, and the relevant data of patent numbers in various provinces are complete and accurate. In this paper, the chain-on-chain growth rate of patent applications is adopted to represent the dynamic level of continuous innovation [34]. The obtained chain growth index is processed logarithmically. Among them, the formula for calculating the chain growth rate is as shown in Equation (5):

$$inno{v}_{it}=\frac{patent{s}_{i,t}+patent{s}_{i,t-1}}{patent{s}_{i,t-1}+patent{s}_{i,t-2}}\cdot \left(patent{s}_{i,t}+patent{s}_{i,t-1}\right)$$

(5)

The relevant data on regional human capital accumulation ($Huma{n}_{it}$) come from the China Statistical Yearbook of Science and Technology. From the theoretical analysis of the content of human capital accumulation, it can be seen that OFDI promotes the accumulation of regional human capital, which gives the region more people with technical knowledge, and then promotes regional innovation activities to improve innovation output. In order to test the mediating mechanism of human capital accumulation, this paper explores the intermediary effect of human capital accumulation on the innovation effect of public education investment and selects the number of employees engaged in scientific research and development and technical service to measure the level of regional human capital accumulation. Referring to this practice, this paper selects the number of employees engaged in scientific and technological research and development and technical service in the region to measure the accumulation of regional human capital [35]. The indexes are processed logarithmically.

The relevant data of regional absorption capacity ($Absor{b}_{it}$) come from the China Statistical Yearbook of Science and Technology. Research and development activities are regarded as the main driving force of innovation. Research and opening activities not only enable enterprises to produce innovative knowledge but also identify, absorb, and utilize external knowledge. The regional absorptive capacity includes many aspects, and there is no unified standard yet. The R&D investment has dual characteristics in the process of regional innovation, which is not only conducive to directly creating new knowledge but also can improve the effective channels for innovation subjects in the region to absorb new knowledge. Therefore, the greater the R&D investment intensity, the stronger the absorption capacity of the region to new knowledge and technology. The investment intensity of R&D funds is used as a proxy variable to measure regional absorptive capacity [36]. R&D investment intensity is the proportion of regional R&D expenditure to GDP.

The regional financial market level ($Financ{e}_{it}$) measures the level of financial development, with two representative indicators that are often used in empirical studies. One is the ratio of the money stock to regional GDP in a broad sense. The other one is the financial correlation ratio, which is generally the proportion of the value of regional financial assets to GDP. This paper adopts the idea of using the loan balance of regional financial institutions and the proportion of GDP as an index to measure the development degree of the financial market. The larger this indicator is, the larger the regional financial market is, and the more adequate the resource allocation and capital supply of the whole economy will be.

Other factors of each province may have an impact on regional sustainable innovation. Control variables mainly include the following variables. The per capita GDP growth rate ($GD{P}_{it}$) is the proportion of the difference between the per capita GDP of the current year and the per capita GDP of the previous year. Usually, the higher the level of regional economic development, the more it will promote regional R&D activities. The scale of government ($Gov{t}_{it}$) adopts the proportion of general budget expenditure of local government to GDP. Relevant data of control variables come from the Statistical Yearbook of China and the annual statistical yearbooks of various provinces and cities. They also include the level of regional infrastructure construction, measured by highway facilities ($Roa{d}_{it}$); that is, the length of highways per 10,000 square kilometers in the region. In addition, the protection of intellectual property rights ($IP{R}_{it}$) is included in research models as a control variable. The level of protection on intellectual properties positively affects regional capabilities on innovation [37]. It is measured by the proportion of turnovers in the technological market over the nominal GDP [38]. Moreover, international trade ($Trad{e}_{it}$) positively affects innovation [39]. The export dependency ratio is used to measure international trade with the ratio of total imports and exports to the nominal GDP.

3.3. Descriptive Statistics

The sample of this paper is the provincial panel data of China from 2008 to 2019. In this paper, we used STATA 15.0 software to process and analyze the data.

Table 1. Descriptive statistics.

Variable	Description	Obs	Mean	Std Dev	Min	Max
Innov	Regional sustainable innovation	360	10.506	1.521	6.453	14.036
OFDI	The stock of external non-financial OFDI	360	12.522	1.812	6.157	16.814
Human	Number of employees engaged in scientific R&D and technical service	360	11.054	1.183	7.454	13.596
Absorb	R&D investment intensity	360	1.608	1.094	0.230	6.310
Finance	Loan balance of regional financial institutions/GDP	360	131.307	40.575	57.382	252.422
GDP	GDP growth rate per capita	360	9.414	3.091	0.500	30.000
Govt	Government expenditure /GDP	360	25.105	11.320	9.978	75.830
Road	Highway length per 10,000 square kilometers	360	9.104	4.956	0.784	21.177
IPR	Turnover in the technological market/GDP	360	1.260	2.410	0.020	16.100
Trade	Total imports and exports /GDP	360	1.101	0.165	0.059	9.793

contains the descriptive statistics and

Table 2. Correlation matrix.

Variables	Innov	OFDI	Human	Absorb	Finance	GDP	Govt	Road	IPR	Trade
Innov	1.0000
OFDI	0.7182	1.0000
Human	0.6805	0.4677	1.0000
Absorb	0.4239	0.4548	0.6470	1.0000
Finance	0.0789	0.2416	−0.1678	0.3112	1.0000
GDP	−0.2168	−0.2870	−0.1713	−0.1955	−0.4570	1.0000
Govt	−0.4109	−0.2153	−0.7492	−0.4048	0.4965	−0.1003	1.0000
Road	0.4281	0.3730	0.6905	0.5864	−0.0341	−0.0590	−0.6392	1.0000
IPR	0.1267	0.2519	0.2648	0.6743	0.4284	−0.2092	−0.0397	0.2207	1.0000
Trade	0.5816	0.7282	0.4653	0.5974	0.3893	−0.3647	−0.2414	0.4417	0.4379	1.0000

the correlation matrix. The results of correlations between variables are below 0.75, which shows that there are no strong correlations among variables that may affect empirical results.

4. Results of Empirical Analysis

4.1. Regression Analyses of the Effects of OFDI on Sustainable Innovation

4.1.1. Baseline Regression

To test the effect of OFDI on China’s sustainable innovation ability, the empirical part of this paper used STATA15.0 software to make a regression analysis of panel data. Firstly, we used the Hausman test to choose between random effect and fixed effect models. The p value obtained by the Hausman test was less than 0.05, which indicates that the original hypothesis should be rejected at the level of 5%; that is, the hypothesis that the random effect model (RE) is correct and normative should be rejected. Therefore, we selected the fixed effect regression model (FE) to test Equation (1).

Table 3. The regression results of OFDI on sustainable innovation.

Variable	Innov
	(1)	(2)	(3)	(4)	(5)	(6)
OFDI	0.495 ***	0.456 ***	0.428 ***	0.389 ***	0.375 ***	0.394 ***
	(0.014)	(0.019)	(0.021)	(0.0274)	(0.028)	(0.029)
GDP		−0.025 **	−0.025 **	−0.022 **	−0.019 **	−0.019 **
		(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
Govt			0.017 **	0.020 **	0.195 **	0.018 ***
			(0.006)	(0.006)	(0.006)	(0.006)
Road				0.652 *	0.687 *	0.071 *
				(0.289)	(0.288)	(0.029)
IPR					0.608	0.777 **
					(0.264)	(0.270)
Trade						−0.427 **
						(0.164)
_cons	4.306 ***	5.031 ***	4.942 ***	−9.420 ***	4.740 ***	−5.603 **
	(0.174)	(0.302)	(0.301)	(2.730)	(0.320)	(1.199)
R-sq	0.797	0.802	0.806	0.809	0.829	0.847
observed value	360	360	360	360	360	360

Values in () are robust standard error. * p < 0.05, ** p < 0.01, *** p < 0.001.

shows the results of OFDI on sustainable innovation with fixed effect regression.

The results in column (1) show that the regression coefficient of OFDI is 0.495, which has passed the 1% significance level test. The growth rate of GDP per capita is added in the regression as a control variable, and the results are shown in column (2). The regression coefficient of OFDI is positive, and it still passes the 1% significance level test. Then, the coefficient of OFDI is 0.428 and it is significantly positive with adding the proportion of government expenditure to GDP as a control variable in column (3). In column (4), regional highway infrastructure is added as a control variable for regression analysis, and the coefficient of the core explanatory variable OFDI is 0.314, which is significant at 1% significance level. The coefficient of OFDI in column (5) is 0.386, which is statistically significant. Moreover, the coefficient of OFDI is 0.313 after adding all control variables in Equation (1). This shows that China’s outward foreign direct investment plays a role in promoting regional capability in sustainable innovation. The coefficient of highway distance per square kilometer in a region is significantly positive in column (4) and column (5). This helps to explain that the improvement of regional infrastructure represented by highways is conducive to the improvement of regional sustainable innovation capability. Among the regression results on control variables, the growth rate of GDP per capita and net exports negatively affect regional sustainable innovation. The government expenditure and highway infrastructure positively affect regional sustainable innovation in China.

4.1.2. Endogeneity Discussion

Empirical analysis of the impact of OFDI on regional sustainable innovation discussed above shows that OFDI promotes regional sustainable innovation in China, but sustainable innovation may also encourage multinational cooperation to invest abroad; that is, there may be reverse causality between OFDI and sustainable innovation. Therefore, we need to discuss the endogeneity issue. In view of the possible endogeneity issues in the baseline regression, our work referred to Huang and Zhang (2022) in dealing with the endogeneity issue. Aiming at examining the endogeneity issue of OFDI and sustainable innovation, our study adopts instrumental variables to carry out two-stage least squares (2SLS) regression. The flow of OFDI with a one-period lag is selected as the instrumental variable of OFDI. The coefficient of OFDI in

Table 4. The regression results on endogeneity (2SLS).

Variable	Innov	Innov
	(7)	(8)
OFDI	0.676 ***	0.493 ***
	(0.025)	(0.034)
Control	No	Yes
N	360	360
R-squared	0.631	0.798

Values in () are robust standard error. *** p < 0.001.

is compared with the result in column (7) and column (8) of Table 3. The results are consistent. The coefficient of OFDI of 2SLS regression without control variables is significantly positive in column (7). After including all control variables, the coefficient of OFDI is still positive and statistically significant. Overall, the regression results on the positive impact of OFDI on regional sustainable innovation are robust.

4.2. Mediating Effect

For this paper, we used the mediating model to test the transmission mechanism among OFDI, regional human capital accumulation, and regional sustainable innovation capability. Among them, human capital accumulation is a mediating variable. The test results of the mediating mechanism of human capital accumulation under the regression of fixed effect panel data are shown in

Table 5. Regression results of the mediating effects of human capital accumulation.

Variable	Innov	Human	Innov
	(9)	(10)	(11)
OFDI	0.394 ***	1.457 ***	0.332 ***
	(0.029)	(0.180)	(0.030)
Human			0.540 ***
			(0.106)
Control variable	YES	YES	YES
_cons	4.592 **	−7.986 ***	0.027
	(0.316)	(1.701)	(0.957)
R-sq	0.816	0.738	0.829
observations	360	360	360

Values in () are robust standard error. ** p < 0.01, *** p < 0.001.

.

Column (9) of Table 5 is the regression result of Equation (1); that is, the coefficient of the OFDI variable shows that OFDI has a significant positive impact on regional sustainable innovation. By adding control variables, including the GDP growth rate, government expenditure ratio, regional highway facilities, protections on intellectual properties, and net exports, the regression results in column (9) in Table 5 are the same as those in column (6) in Table 2. The coefficient of OFDI is significant, referring to the analysis method of the mediating effect [29]. Second, we tested the relationship between OFDI and human capital accumulation, and carried out a regression analysis on Equation (2). The analysis results are shown in column (10) of Table 5. The coefficient of OFDI of 0.140 is significantly positive; that is, regional foreign direct investment has promoted regional human capital accumulation. The third step, taking OFDI and human capital accumulation as explanatory variables at the same time, we used regression analysis to test their effects on regional sustainable innovation capability. The results are shown in column (11) of Table 5. Among them, the coefficient of human capital accumulation is 0.436 and has passed the 1% significance level test because the coefficient of OFDI in column (10) and the coefficient of human capital accumulation in column (11) are both significant. Therefore, the indirect effect is significant. That is, OFDI’s influence on the ability of sustainable innovation is partly direct and partly through the accumulation of human capital of intermediary variables. In the process of indirect effect, the feedback mechanism of foreign investment promotes the accumulation of human capital in the home country and then promotes the sustainable innovation of the region.

4.3. Threshold Effect

As shown in the related research on OFDI and regional innovation in the existing literature, the impact of OFDI on innovation may be nonlinear. Different regions have different absorptive capacities for OFDI reverse technology spillovers, and there are also differences in the regional institutional environment. Therefore, there are differences in technical knowledge of absorbing and transforming OFDI spillovers among different regions. Because of the difference in threshold variables, independent variables have different effects on dependent variables [37]. Therefore, in this paper we used the threshold effect model for further testing.

Firstly, the bootstrap method was used to test whether there was a threshold effect, and the number of real threshold values had to be determined. The results are shown in

Table 6. Threshold existence test.

	Model	Threshold Value	Variance Ratio	p Value	Critical Value
					1%	5%	10%
Absorptive capacity	Single threshold	1.18	23.62 *	0.027	38.60	25.38	20.59
	Double threshold	2.31	22.76	0.243	26.27	20.69	17.58
	Triple threshold	5.25	22.22	0.433	37.09	23.86	18.07
Financial market	Single threshold	88.94	31.12 *	0.020	32.87	23.39	20.27
	Double threshold	119.08	15.37	0.103	24.67	18.30	15.58
	Triple threshold	204.06	7.99	0.603	33.10	23.88	19.96

* p < 0.05.

.

As can be seen from

Table 7. Regression results of the threshold model.

Innov	Absorptive Capacity	Financial Development Level
OFDI	0.376 ***(0.042) (absorb1.110)≤	0.352 ***(0.040) (finance88.940)≤
	0.404 *** (0.041) (absorb1.110)>	0.378 *** (0.038) (finance88.940)>
GDP	−0.015 (0.008)	−0.017 *(0.008)
Govt	0.014(0.010)	0.013 (0.011)
Road	0.500 (0.341)	0.500 (0.359)
IPR	0.868 *(0.386)	0.787 *(0.390)
Trade	−0.422 *(0.204)	−0.351(0.213)
Cons	4.839 *** (0.526)	5.039 *** (0.038)
Obs	360	360

* p < 0.05, *** p < 0.001.

, the regional absorption capacity of threshold variables passed the single threshold test but failed the double threshold test and the triple threshold test. The threshold value of absorptive capacity was 1.11. Similarly, according to the results in Table 5, the financial market development level passed the single threshold variable but failed the double and triple threshold tests, and the threshold value was 88.94. Furthermore, the regression analysis of the threshold effect of panel data was carried out, and the results are shown in Table 7.

The results of the threshold regression model in Table 7 show that OFDI has a significant positive impact on sustainable innovation. Therefore, the empirical results show that China’s outward foreign direct investment can play a positive role in promoting regional sustainable innovation capability. When the regional absorptive capacity (that is, R&D investment intensity) is lower than 1.11, OFDI has a significant positive impact on sustainable innovation, with a marginal coefficient of 0.287. When the R&D investment intensity is higher than 1.11, the marginal coefficient of OFDI’s significant positive impact on sustainable innovation is 0.313. The conclusion shows that although China’s foreign direct investment has a positive impact on sustainable innovation capability when the regional absorptive capacity is higher than the threshold value, the positive effect of foreign investment on sustainable innovation will also be enhanced. When the development level of the regional financial market (that is, the proportion of loan balance of financial institutions to GDP) is lower than 88.94, the marginal coefficient of OFDI’s impact on sustainable innovation is 0.265. When the level of regional financial development is higher than 88.94, OFDI’s impact on sustainable innovation is still significantly positive, and the marginal coefficient is 0.313. The conclusion shows that China’s foreign direct investment has a significant positive impact on the sustainable innovation capability of domestic regions. When the level of regional financial development exceeds the threshold, the positive role of foreign investment in sustainable innovation will also be strengthened.

The above empirical results show that OFDI’s impact on sustainable innovation is restricted by its own absorptive capacity and the development level of external financial markets. There is a single threshold value for both. The positive impact of OFDI on sustainable innovation is enhanced when the threshold variable exceeds the first threshold value.

5. Discussion

According to Section 4.1, the OFDI significantly positively affects regional sustainable innovation in China. With the increasing amount of OFDI in China, the regional capability for sustained innovation is improved. This result is consistent with studies on the reverse spillover effect on regional innovation in China [17,40]. According to Section 4.2, human capital accumulation mediates the relationship between OFDI and regional sustainable innovation. This finding supplements the study of Yang and Huang [25] by analyzing the mediating effect of human capital accumulation. It implies that OFDI affects regional sustainable innovation in both direct and indirect ways. OFDI benefits human capital accumulation, and the more human capital accumulated, the higher capabilities for regional sustainable innovation.

In the threshold effect analysis, it can be seen that the regional absorptive capacity and development of the financial market have significant threshold effects on the relationship between OFDI and sustainable innovation. With the threshold existence test in Table 6, the single threshold exists. When the regional absorptive capacity exceeds the threshold value, the impact of OFDI on regional sustainable innovation is stronger. The result is similar to those of studies on the threshold effects of absorptive capacity on the impact of OFDI on regional innovation [21,24,41]. When the level of development of regional financial markets exceeds the threshold value, the impact of OFDI on regional sustainable innovation is stronger. The result is similar to that found in studies on the threshold effects of financial development on the impact of OFDI on regional innovation [23].

6. Conclusions

In the context of China’s high-level exploitation of foreign investment, exploring the impact of OFDI on China’s sustainable innovation is an important issue for China to achieve sustainable economic development. Therefore, discussing the effect of the reverse spillover effect brought by OFDI is an important issue in addressing the sustainable development of developing countries. This paper focuses on the impact of Chinese OFDI on regional capabilities of sustained innovation. Although a large number of empirical studies have analyzed the impact of OFDI on innovation, the mechanism of the impact on sustained innovation in China is unclear. Our study contributes to this research area.

6.1. Main Findings

Using provincial-level panel data of China from 2008 to 2019, this paper empirically analyzes the impact of OFDI on regional capabilities of sustained innovation in China and describes its impact mechanisms. The following main conclusions can be drawn.

First, OFDI significantly promoted regional capabilities on sustainable innovation; that is, the larger amount of OFDI, the higher the regions’ capabilities for sustained innovation.

Second, this paper finds that human capital accumulation has a mediating role in the relationship between OFDI and sustainable innovation through the analysis of the mediating effect. It is concluded that OFDI contributes to the accumulation of regional human capital, and then has a positive impact on regional sustainable innovation.

Third, the regional absorptive capacity has a significant threshold effect on the relationship between OFDI and sustainable innovation. By analyzing the threshold effect, when the regional absorptive capacity exceeds the threshold, OFDI’s effect on sustainable innovation will be enhanced.

Finally, the regional financial market also has a significant threshold effect on the relationship between OFDI and sustainable innovation. When the development level of the regional financial market exceeds the threshold value, the positive effect of OFDI on sustainable innovation is stronger.

6.2. Theoretical Implications

Based on the academic discussion related to the impact of OFDI on domestic innovation, this paper provides new insights from the perspective of sustainable innovation. Most articles focus on the impact of OFDI on the capacity and efficiency of regional innovation and have not yet reached a unified conclusion. Moreover, factors such as regional absorptive capacity and the development of the financial market have been rarely considered in related studies. This paper discusses the impact of OFDI from the perspective of sustainable innovation. We also combined factors such as human capital, regional absorptive capacity, and financial market development to discuss the mechanism of OFDI reverse spillover. The analysis of the acquisition and transmission of sustainable innovation provides the basis for our theoretical analysis of OFDI reverse spillover. Our work develops the view that OFDI facilitates regional human capital accumulation, which in turn exerts a positive effect of human capital on the capabilities of sustainable innovation. In addition, based on the nonlinear findings from existing studies, we explored the impact factors of the relationship between OFDI and sustainable innovation by selecting the absorptive capacity and the development of the financial market as threshold variables for empirical analysis. Compared with the existing literature, this paper expands the theoretical framework of the OFDI spillover effect by exploring the mediating effect and threshold effects, which help to reveal the complex relationship between OFDI and regional capabilities on sustained innovation.

6.3. Political Implications

Against the background of the new era, China should always adhere to the “going out” strategy which is a new strategy of continued opening up in China. Based on the research conclusion of this paper, the following policy suggestions are put forward.

First, we should constantly improve the service level of the financial market and perfect the construction of the credit service system. By effectively solving the capital problems of enterprises, we can provide sufficient capital sources for enterprises to invest abroad. At the same time, it provides a good financial market environment for enterprises to absorb OFDI reverse technology spillovers to promote domestic sustainable innovation.

Second, the promotion of domestic sustainable innovation is inseparable from the support of R&D investment. Therefore, our government should actively promote all kinds of enterprises to increase R&D investment and encourage enterprises to actively participate in independent innovation activities through preferential policies and other measures.

Third, we must make full use of and absorb external advanced technical knowledge, actively establish a good research environment for scientific researchers, and attract scientific research talents to increase human capital accumulation. With the deepening of the opening-up strategy, the continuous improvement of the institutional environment, the enhancement of regional absorptive capacity, and the accumulation of human capital, China’s regional sustainable innovation capability have been positively influenced. Therefore, it should be an important strategy to continue to deepen the strategy of foreign investment and utilize its technology spillovers to cultivate the ability of independent and sustainable innovation.

6.4. Future Extensions

Future research can be extended in several ways. First, future research could explore the impact of OFDI on sustainable innovation at the country level. Our work only focuses on the relationship between OFDI and sustainable innovation at the provincial level in China, but country-level data can extend our findings and have important practical implications for sustained development in developing countries. Second, we encourage future research to use more factors of the institutional environment to explore the complex relationship among OFDI, the institutional environment, and sustainable innovation, such as the legal environment.