Time-Varying Effects of China’s Agricultural OFDI on the Sustainability of Bilateral Agricultural Trade Within RCEP Countries: Multiple Analyses Based on the Model of Trade Gravity and TVP-VAR

Abstract

China’s agricultural OFDI has achieved a new promotion of regional opening to the RCEP agricultural field, and the bilateral agricultural trade cooperation has great potential. China’s agricultural OFDI has been steadily increasing in recent years. As significant trade partners for Chinese agricultural products, RCEP countries represent an important region for China’s agricultural OFDI initiatives. However, existing studies have paid little attention to the trade effect and dynamic impact of China’s agricultural OFDI on agricultural products of RCEP countries. Therefore, based on the panel data of agricultural trade between China and RCEP countries from 2004 to 2019, this paper uses the trade gravity model and TVP-VAR model to investigate the time-varying effects of China’s agricultural OFDI on agricultural trade of RCEP countries. The findings are as follows: (1) The impact of China’s agricultural OFDI on bilateral agricultural trade between China and RCEP countries has significant time-varying characteristics, and there are short-, medium-, and long-term positive impacts most of the time. (2) When the bilateral agricultural trade reaches a certain level, the role of China’s agricultural OFDI is limited, but the overall positive and significant impact remains, showing a strong trade effect. (3) The positive impact on China’s agricultural OFDI has a heterogeneous impact on RCEP countries with different income levels and geographical connections, but the overall positive externality effect is significant. The research conclusions expand the research perspective on the trade effect of agricultural OFDI to a certain extent and provide management and policy implications for agricultural enterprises to “move fast” and promote the sustainability of bilateral agricultural trade between China and RCEP countries through government efforts.

1. Introduction

Outward foreign direct investment (OFDI) is an investment made by an investment entity in one country or region to an enterprise in another country or region for the purpose of gaining operational control and sharing long-term profits. For a long time, the interaction between OFDI and international trade has been a hot topic in academic circles. It is generally believed that there are significant trade effects. Early studies have found that if the assumption that factors of production cannot flow internationally is relaxed, OFDI will have a transfer effect on international trade; that is, OFDI will cause changes in bilateral trade patterns [1]. With the development of global economy and in-depth research in the field of investment and trade, OFDI’s transfer theory of international trade has been unable to fully explain the evolution law of international economic activities. OFDI will bring the transfer of production factors in management experience, professional technology, and other aspects, further promote the sustainability of international trade development, and thus produce the trade creation effect of OFDI [2]. Relevant empirical tests also find that not only will monopolistic enterprises’ OFDI produce trade effects (including trade creation and trade transfer effects), but also other enterprises will produce trade effects in the marginal expansion of the entire industry [3]. With China’s accession to the WTO, the “going out” regulation has been relaxed, and the research on the impact of China’s OFDI on international trade has received more attention. Some scholars believe that China’s OFDI belongs to the trade creation type. After China’s economy enters the stage of transformation and upgrading, traditional labor-intensive industries must transfer to meet the requirements of industrial upgrading and other economic growth and quality improvement, and the most important way is foreign investment [4]. Some scholars also believe that China’s OFDI will indeed promote trade exchanges, but this effect is the joint result of trade creation and trade transfer, because investment activities increase the participation rate of Chinese enterprises in international trade and international division of labor [5].

Trade in agricultural products accounts for a large proportion among RCEP member countries, and the RCEP agreement has a far-reaching impact on trade in agricultural products. The RCEP will provide a broader platform for the construction and development of agricultural internationalization, help promote the high-quality agricultural development of economies in the region [6], and accelerate the pace of China’s transformation from a major agricultural country to an agricultural power. OFDI for agriculture can effectively alleviate the problems of China’s agricultural development, consolidate and promote the political, diplomatic, and economic and trade relations between China and other countries, and is an important part of China’s peaceful development strategy [7]. China has maintained long-term and close investment cooperation with RCEP state parties in agricultural trade and agriculture. China’s advantages in market, capital, technology, and business environment will lead the overall agricultural economic and trade cooperation among RCEP countries to a larger scope and a higher level [8].

After the RCEP takes effect, Chinese agricultural enterprises’ active overseas investment activities can promote the import and export trade of agricultural products, and even have an impact on the national economy of the host countries, which is the so-called trade creation effect. China takes advantage of geopolitics and deeply participates in the regional agricultural value chain, which helps to improve China’s position in the value chain of the “Belt and Road” countries, that is, the so-called trade diversion effect. Studies have proved that in the short term, China’s agricultural OFDI has an import transfer effect and presents a long-term equilibrium relationship [9]. Long-term analysis shows that China’s agricultural OFDI has a driving effect on the host countries, especially in ASEAN countries, where there are a large number of investment enterprises, a large amount of investment, a wide range of cooperation fields and industrial chains, and the local agricultural and economic development has been promoted significantly [10]. After reviewing the existing literature, it can be found that the academic circle has established a relatively complete and mature theoretical system for the research on the connection between OFDI and international trade, but in terms of the research content, developing countries are rarely involved. Most domestic scholars study the trade effects of OFDI and foreign trade based on their own national conditions, or explore the impact of direct investment on trade in some regions or sectors from the perspective of a certain field or sector, etc., but there are few studies in the field of agriculture.

In the context of the rapid development of China’s OFDI, what is the relationship between China’s agricultural direct investment in RCEP countries and agricultural trade? From the analysis of empirical evidence, does China’s agricultural OFDI bring about a time-varying impact of “only good and win-win” or “mutual benefit and win-win”? Existing studies have not given a full explanation. As China rapidly develops from a large trading and agricultural country to an investment and agricultural power, it is particularly important to discuss the impact of China’s agricultural OFDI on the agricultural trade of RCEP member countries, which will help better optimize the structure of OFDI and provide certain strategic guidance for Chinese agricultural enterprises to enter the international market, promoting the sustainability of bilateral economic and trade cooperation. Outward investment in agriculture and trade in agricultural products have an important impact on sustainability, involving multiple economic, environmental, and social dimensions. Outward investment in agriculture contributes to technology transfer and productivity gains, especially in developing countries, by introducing advanced technologies such as precision agriculture and water-saving irrigation that not only increase crop yields but also reduce resource consumption, thereby contributing to environmental sustainability. At the same time, agricultural trade improves the efficiency of agricultural production in various regions by optimizing the allocation of global resources and promotes the development of agriculture through changes in market demand. Therefore, studying the impact of agricultural investment on trade at different points in time is helpful for promoting long-term agricultural trade cooperation between countries, paying attention to green investment, promoting fair trade mechanisms, strengthening international policy coordination and technical cooperation, and guiding the rational allocation of agricultural resources, so as to achieve the long-term sustainability of agricultural systems.

The possible marginal contributions of this paper are as follows: Based on the current economic and trade development trend and geo-economic relations between China and RECP countries, China’s agricultural OFDI will have a profound trade impact on RECP member countries, and this study will help to deepen the research in this field. At the same time, it helps to analyze the economic and trade cooperation between China and RECP countries, provide an empirical basis for optimizing international trade policies, and effectively deal with the possible international impact of China’s agricultural OFDI.

2. Theoretical Analysis and Research Hypothesis

Agricultural OFDI is an important way [8] for a country to participate in the global agricultural value chain based on comparative advantages such as resources, technology, and market, and realize cross-border cooperation and international status promotion of the agricultural industry chain. The technology spillover effect generated by agricultural OFDI is reflected in the export of raw materials, agricultural machinery, technology, and other products required by the host country from the mother country for agricultural activities, thus promoting the trade creation effect [9]. Since the beginning of the 21st century, China has accelerated its participation in the agricultural value chain, and the scale of agricultural OFDI has increased significantly. China has entered a new development stage [10] of actively participating in the allocation of global agricultural resources. China’s agricultural OFDI has a positive effect on agricultural imports by promoting product diversification rather than simply raising prices [11]. In addition, it promotes trade liberalization and enhances the ability of developing countries to import and export in global trade [12]. Agricultural OFDI by Chinese enterprises not only increases agricultural investment in host countries and helps alleviate poverty, but also expands supply in local markets [13]. China’s agricultural OFDI has not only solved the problem of overcapacity and promoted the transfer of agricultural technology, but also improved the agricultural production efficiency [14] in developing countries, provided new development impetus for agriculture and rural areas in less-developed countries, and effectively ensured food security [15]. At the same time, the scale of China’s agricultural imports continues to expand, and the rate of food self-sufficiency continues to decline, resulting in a long-term trade deficit in foreign trade. The fundamental reason is that compared with the major agricultural producers in the world, China’s agricultural competitive advantage is generally declining [16]. By exporting technical advantages such as breeding, cultivation, and processing, China’s agricultural OFDI produces agricultural products in the host country, which has a trade transfer effect and can make up for the deficiency of China’s food self-sufficiency rate to a certain extent. On the other hand, enterprises in the host country can rapidly improve their technology and productivity by learning and imitating the production management experience of foreign-funded enterprises. This spillover effect is finally reflected in the total amount of agricultural foreign direct investment, the increase in production, and the improvement of the agricultural import and export trade structure, that is, the technological spillover effect of agricultural OFDI. According to “dry learning” and the “inter-industry spillover effect”, China’s agricultural OFDI is conducive to promoting the development of agricultural planting, management, and technology levels of RCEP countries, thus generating a trade creation effect. Therefore, the first hypothesis is proposed:

H1. 

There is a trade effect between China’s agricultural OFDI and agricultural trade between China and RCEP countries.

The trade effect of OFDI is complicated, and OFDI with different investment motives has different [17] effects on trade. In theory, OFDI can promote the upgrading of product structure between the home country and the host country, so that both sides can produce more products that meet the needs and preferences of consumers and improve the added value and market share [18] of products. At the same time, OFDI can also improve the product quality of the two countries through technology transfer, brand effect, certification effect, and other ways, and further enhance the competitive advantage [19] of both sides in the international market. In fact, the position of Chinese agriculture in the Asia-Pacific production network is rising steadily, but the participation in the value chain is relatively low [20]. The technical complexity of China’s agricultural exports is constantly improving, and the technical complexity of agricultural exports varies greatly [21] among different categories and regions. By providing funds and using agricultural technology to stimulate agricultural production potential, China has promoted the development of agricultural sectors in developing countries [22]. Studies have shown that outward foreign direct investment (OFDI) generally has a positive impact on trade development in developing countries [23]. Taking China as an example, it has shown a significant trade creation effect on the OFDI of countries along the “Belt and Road”, especially in promoting medium and low technology exports [24]. However, the relationship between OFDI and trade may vary depending on a country’s income level. Low-income countries may face substitution effects, while upper-middle-income countries tend to exhibit complementary effects [25]. China’s OFDI can have a positive impact on export trade by expanding market scale, improving product competitiveness, optimizing industrial structure, etc. Huawei’s investment in Africa is a typical case. In general, incorporating OFDI into the framework of international trade theory contributes to a more comprehensive understanding of its interrelationship [26]. The main body of China’s agricultural OFDI is agricultural enterprises. In addition to land leasing, Chinese investors also implement the order contract model to [20] connect overseas bases, farmers, and markets, promote commercial investment, and develop the agricultural development potential of host countries. At the same time, it will bring capital, equipment, and experience to host countries, accelerate technology promotion and innovative application, and significantly improve the agricultural production conditions of RCEP member countries. Therefore, this paper proposes the second hypothesis:

H2. 

China’s agricultural OFDI has a positive impact on the bilateral agricultural trade effect between China and RCEP countries.

3. Basic Situation of Bilateral Agricultural Trade

3.1. China’s Agricultural OFDI Has Developed Rapidly

In recent years, China’s agricultural OFDI has shown a rapid growth trend. According to the statistics of the Ministry of Commerce, from 2006 to 2016, China’s agricultural OFDI increased by 8.5 times and reached USD 4.680 billion in 2019, an increase of 9.80% over 2018. In 2020, affected by the COVID-19 epidemic, China’s agricultural OFDI decreased to USD 4.20 billion. However, from the long-term trend, the scale of China’s agricultural OFDI continues to expand and still maintains a steady growth momentum. In 2022, the investment stock reached USD 30.2 billion, creating 180,000 local jobs in host countries.

From the perspective of investment region, geographical relationship is the main factor when considering the regional distribution of China’s agricultural OFDI, and more than 50% of the investment flows into Asian countries. Political relations are also a key factor in the regional distribution of China’s agricultural OFDI. In Asia, Africa, and Latin America, most of the countries in these regions maintain good diplomatic relations with China and have abundant land, water resources, and labor force, which is in line with China’s agricultural development needs. In terms of the scale of investment, it has covered eight major fields: grain, cotton, oil, sugar, glue, livestock, fishery, agricultural materials, and agricultural machinery, and the industrial categories are constantly enriched. From the perspective of investment methods, the main ones are leasing land overseas and establishing overseas agricultural production bases, or implementing the contract model of order, providing seeds, fertilizers, and agricultural machinery to farmers in the host country, and purchasing agricultural products. There are also some agricultural enterprises participating in cross-border agricultural investment activities through mergers and acquisitions [27].

3.2. Trade in Agricultural Products Between China and RCEP Countries Has Expanded Significantly

RCEP countries are important agricultural trading partners of China, with trade in agricultural products accounting for more than 30 percent of China’s total agricultural trade. Figure 1 shows that, from 2004 to 2023, the import and export volume of China’s agricultural trade with RCEP countries increased from USD 19.34 billion to USD 103.33 billion, a year-on-year increase of 12.8 percent. Among them, the value of exports increased from USD 11.909 billion to USD 416.03 billion, and the value of imports increased from USD 7.430 billion to USD 617.72 billion, further strengthening the agricultural trade relations between China and RCEP member countries.

China and RCEP countries are complementary and synergistic in the trade of agricultural products. China’s major agricultural exports include fruits, aquatic products, tea, and flowers, which have a high demand and competitiveness in RCEP countries. The major agricultural products imported from RCEP countries include soybeans, palm oil, rubber, and rice, which are important raw materials for China’s agricultural production and processing. Through trade in agricultural products, China and RCEP countries can optimize resource allocation and improve economic returns. With the implementation of the RCEP agreement, there will be more opportunities and space for cooperation between China and RCEP countries in agricultural trade. The RCEP will provide more convenient and favorable trade terms for both sides, reduce tariff barriers, standardize trade rules, and promote trade facilitation [28].

The ratio of total agricultural imports and exports of one country to another country’s gross agricultural product can be used as a measure of dependence on agricultural trade. As shown in Figure 2, the average dependence of RCEP countries on China’s agricultural products is greater than the average dependence of China on its agricultural products, indicating that China is in a relatively active position and occupies a dominant position in RCEP countries’ agricultural products trade. From the analysis of the evolution trend of dependency degree, the dependency degree of RCEP countries on China’s agricultural products trade shows an increasing trend year by year, and the growth rate is greater than that of non-RCEP countries in China’s agricultural products trade. It is expected that, in the future, the trade volume of agricultural products between China and RCEP countries will continue to maintain a steady growth, bringing more benefits and impetus to the economic and social development of both sides.

4. Analysis Based on Trade Gravity Model

4.1. Model Setting

The idea and concept of the gravitational model originated from the law of gravitation in physics proposed by Newton, and the gravitational model has been continuously expanded and improved after the introduction of studies to measure bilateral trade flows. After the 1980s, some scholars added factors such as adjacent borders and cultural convergence into the model, mainly because these factors have a strong impact on the characteristics of the agricultural outbound investment network, and then have an impact on a country’s position in the division of labor in the global agricultural value chain. Conversely, the rise of a country’s position in the global agricultural value chain also affects the distribution of a country’s foreign investment in agriculture on a global scale. When setting the model, although different scholars made small adjustments to the explanatory variables of the gravity model, the core explanatory variables generally chose objective factors with short-term stability. Therefore, in this paper, referring to Anderson’s practice [29], OFDI flow, which is not easy to change in the short term, is selected as the core explanatory variable. The specific model is as follows:

(1)$\ln {EX}_i={\alpha }_0+{\alpha }_1\ln {OFDI}_i+{\alpha }_2\ln CGDP+{\alpha }_3\ln {GDP}_i+{\alpha }_4\ln {dist}_i+{\alpha }_5{comlang}_i+{\alpha }_6{contig}_i+{\alpha }_7{FTA}_i+{\alpha }_8{Market}_i+{\sum }_{i=1}^{j=15}year+{\epsilon }_i$

(2)$\ln {IM}_i={\beta }_0+{\beta }_1\ln {OFDI}_i+{\beta }_2\ln CGDP+{\beta }_3\ln {GDP}_i+{\beta }_4\ln {dist}_i+{\beta }_5{comlang}_i+{\beta }_6{contig}_i+{\beta }_7{FTA}_i+{\beta }_8{Market}_i+{\sum }_{i=1}^{j=15}year+{\phi }_i$

4.2. Description of Variables and Data Sources

1. Core explanatory variable: China’s agricultural OFDI flow. According to Chen and Guo’s research [26], agricultural OFDI flow directly reflects the outflow of funds related to agriculture in a country. At the same time, considering the interference of the novel coronavirus epidemic, the regression results of the trade gravity model may be unstable, so the data from 2004 to 2019 are selected in this paper, all of which are taken from the Statistical Bulletin of China’s OFDI in previous years.

2. Explained variables: China’s imports and exports of agricultural products to RCEP countries. According to the research [30], the trade volume of agricultural products can reflect the direct trade of agricultural products between the two countries. The relevant agricultural product data used in this paper are from the UN COMTRADE database and the 3-digit data under the third edition of the United Nations Standard Trade Classification (STIC). According to the WTO definition, agricultural products include categories 0, 1, 2, and 4 of STIC, but do not include categories 27 and 28 of Category 2.

3. Control variables: The geographical distance between China and RCEP capitals, GDP, common official language, territorial border, accession to the WTO, and market share of agricultural products trade were selected as the model-influencing factors in this paper. The above data came from CEPII of the French Institute of International Economics. In order to prevent absorption by fixed effects, DIST is multiplied by Brent international crude oil price in the current year as the measurement data. In the MARKET, the proportion of the trade volume of agricultural products between RCEP countries and China in the trade volume of their own agricultural products is measured, reflecting China’s position in the trade volume of agricultural products between RCEP countries. The larger the value, the greater the possibility for China to carry out trade in agricultural products with RCEP countries.

Due to the lag effect of China’s accession to the World Trade Organization in 2001 and the outlier impact of the epidemic on the data after 2019, the data time span in this paper is from 2004 to 2019. The natural logarithm is taken for all the above data, and the influence of seasonal factors and heteroscedasticity is excluded by using the censusX12 seasonal method. The names and meanings of the specific variables involved in this study are shown in Table 1:

${\mathrm{E}\mathrm{X}}_{\mathrm{i}}$ represents China’s agricultural exports to country i; ${\mathrm{I}\mathrm{M}}_{\mathrm{i}}$ represents the amount of China’s agricultural imports to i. The core explanatory variable, ${\mathrm{O}\mathrm{F}\mathrm{D}\mathrm{I}}_{\mathrm{i}}$, represents the flow of China’s agricultural foreign direct investment to country i. In addition, ${\mathrm{D}\mathrm{I}\mathrm{S}\mathrm{T}}_{\mathrm{i}}$ represents the straight-line distance between China and the capital of country i, and ${\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}}$ represents the gross domestic product of country i. The control variables include YEAR, which refers to the annual dummy variable; ${\mathrm{C}\mathrm{O}\mathrm{M}\mathrm{L}\mathrm{A}\mathrm{N}\mathrm{G}}_{\mathrm{i}}$, which refers to whether China and country i share the same official language; ${\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{T}\mathrm{I}\mathrm{G}}_{\mathrm{i}}$, which refers to whether China and country i share the same territory; ${\mathrm{F}\mathrm{T}\mathrm{A}}_{\mathrm{i}}$, which refers to whether country i has joined the WTO; and ${\mathrm{M}\mathrm{a}\mathrm{r}\mathrm{k}\mathrm{e}\mathrm{t}}_{\mathrm{i}}$, which refers to the proportion of the agricultural trade volume between country i and China in the total agricultural trade volume of country i.

4.3. Baseline Regression Results

According to the sample data, Pooled OLS was selected for analysis in this paper, and the results were shown in Table 2. Model (1) to model (3) tested the import effect of agricultural OFDI, and model (4) to model (6) tested the export effect of agricultural OFDI. The results show that China’s agricultural OFDI has a significant creative effect on the import and export trade of agricultural products in RCEP countries. First, it can be seen from model (3) that the import coefficient is 0.723 and significant at the 1% level, which indicates that after controlling other factors and annual effects, when China’s agricultural OFDI to RCEP countries increases by 1 percentage point, China’s agricultural imports to RCEP countries will increase by 0.723%. Secondly, it can be seen from model (6) that the export coefficient is 0.273 and significant at the 1% level, which means that when other factors remain unchanged, every 1% increase in China’s agricultural OFDI will increase China’s agricultural exports to RCEP countries by 0.273%. This indicates that China’s agricultural OFDI improves the agricultural infrastructure level of RCEP countries, which in turn stimulates the agricultural exports of RCEP countries to China, and ultimately leads to the agricultural trade surplus of RCEP countries with China. Thirdly, the distance between the capitals of the two countries has positive and negative effects on the import and export trade of agricultural products between China and RCEP countries. Models (2) and (3) show that the distance between the capitals of the two countries will have a significant positive impact on the import volume of agricultural products between China and RCEP countries. It can be seen from models (5) and (6) that the greater the distance between the two capitals, the lower the trade volume of China’s agricultural exports to RCEP countries. This indicates that China has a strong preference for obtaining overseas agricultural production resources and is more willing to import agricultural production raw materials from faraway countries with strong resource endowments. However, compared with the export of agricultural products, the long distance will inevitably increase the export cost. If the territory borders have a significant positive impact on the import and export volume of agricultural products, this indicates that China and the RCEP member countries with territorial borders have a more significant trade creation effect, which validates the research conclusion [31] of most scholars. Compared with model (3) and model (6), it can be seen that the increase in the proportion of agricultural trade volume between RCEP member countries and China in their total agricultural trade volume has a significant positive impact on the import of agricultural products from RCEP member countries to China, while it has a significant negative impact on the export of agricultural products from RCEP member countries to China. This is consistent with the fact that China has shown a trade deficit in agricultural products in recent years. Finally, by comparing the coefficients of model (2) with model (5) and model (3) with model (6), it can be found that after controlling other factors, the import transfer effect of agricultural products of China’s agricultural OFDI is greater than the export creation effect of agricultural products. Hypothesis 1 and Hypothesis 2 are confirmed.

4.4. Robustness Analysis

Considering that there are many zeros in the zero-value trade and dummy variables, this paper further selects the PPML (pseudo-maximum likelihood estimation) model for regression, which can effectively correct the zero-value problem and heteroscedasticity problem [32,33]. Table 3 shows the estimation results of the PPML model. From the results of model (7) to model (12), the following can be seen: First, the core explanatory variables are significantly positive at the level of 1%, showing that China’s agricultural OFDI can promote the bilateral agricultural trade between China and RCEP member countries and stimulate the intra-regional agricultural trade circulation. At the same time, by comparing model (9) and model (12), it can be seen that the reverse import-induced effect of OFDI is greater than the export-created effect, which is consistent with the result of the benchmark regression model; that is, China’s agricultural OFDI can promote China’s imports of agricultural products from RCEP countries to a greater extent, possibly because China has a long-term problem of insufficient self-sufficiency rate of agricultural products. In addition, RCEP member countries have always been important agricultural trade partners of China, and the reverse import-induced effect can improve China’s agricultural consumption structure and establish a long-term and stable supply base of agricultural raw materials. Second, for the control variables, consistent with the benchmark regression model, the economic level of RCEP member countries has a significant promoting effect on bilateral trade at the 1% level. Both distance and market have the same effect at a significance level of at least 5%. Both distance and market have a driving effect on China’s import of agricultural products from RCEP member countries, while they have an inhibiting effect on China’s export of agricultural products from RCEP member countries. It can be seen that China imports more agricultural products from member countries that are far away from each other. The possible reason is that RCEP member countries such as Australia, Singapore, and New Zealand have more complementary agricultural products, so China imports the most. At the same time, whether the territory borders or not and whether the RCEP is a member of the WTO will have a significant trade effect. In short, the results of the OLS model and PPML model show the same direction of influence, but the degree of influence is different, which further verifies the robustness of baseline regression.

4.5. Heterogeneity Analysis

A number of studies have shown that a country’s OFDI may either create (increase trade volume) or transfer (decrease trade volume) [34] import and export trade. According to the trade effects of OFDI flows at different levels, it can further provide enterprises with flexible overseas strategic decisions [35]. Therefore, by constructing a panel quantile regression model, this paper examines the impact of China’s agricultural OFDI on the heterogeneity of agricultural trade of RCEP member countries at different loci levels.

The quantile regression results of export trade are shown in Table 4. First of all, most OFDI for agriculture is significantly positive at the 1% level, and its magnitude is around 0.20. At the level of 10% of export trade, agricultural OFDI has the largest positive impact on China’s agricultural exports to RCEP countries, which indicates that keeping other factors unchanged, every one percentage point increase in OFDI will increase China’s agricultural exports to RCEP countries by 0.35%. Secondly, from columns (7) to (9), it can be seen that agricultural OFDI no longer exerts an impact at a significance level of 1%, which indicates that at a higher level of export trade, the export creation effect of China’s OFDI on RCEP countries becomes smaller. Finally, the trade effect at all sub-points is positive, which means that China has an export creation effect on agricultural products of RCEP countries’ agricultural OFDI. However, this effect decreases with the increase in agricultural exports. This indicates that when China’s export trade with RCEP member countries is at a low level, agricultural OFDI can further open China’s trade market with RCEP countries.

Quantile regression results for import trade are shown in Table 5. For the first 30% of the sub-points, the regression coefficients of OFDI reach the maximum, all of which are significantly positive at 1% level, and their sizes are all above 0.70. This shows that when the import volume reaches the top 30% level or above, every one percentage point increase in agricultural OFDI will increase the import volume of agricultural products by 0.70% for China to RCEP countries. In other words, when the level of import trade between China and RCEP countries is low, agricultural OFDI can well promote the import of agricultural products whose “self-sufficiency rate” is insufficient in China, so as to keep meeting the domestic demand for agricultural products. And at the sub-points of 40–80%, agricultural OFDI has a positive effect on agricultural imports at a significance level of at least 10%. However, there is no significant effect at 90% sub-points, indicating that when the import trade between China and RCEP member countries is at a high level, the promotion effect of agricultural OFDI on imports is no longer significant.

According to the results of heterogeneity regression, although the influence of OFDI on imports and exports is limited when they reach a certain level, in general, OFDI has a positive relationship with RCEP countries and most of the sub-points have a significant impact, indicating that China’s agricultural OFDI has a creative effect on the import and export trade of agricultural products in RCEP countries and can expand bilateral agricultural trade.

5. Analysis Based on TVP-VAR Model

The previous empirical results have found that agricultural OFDI has trade effects on agricultural products between China and RCEP countries. But what is the dynamic impact of China’s agricultural OFDI on bilateral agricultural trade? Are there different effects of OFDI in each period? In order to further study the relationship, this paper further uses the TVP-VAR model and the monthly data of 15 RCEP countries from February 2005 to December 2020 as samples to further study the time-varying impact of China’s agricultural OFDI on agricultural trade of RCEP countries.

5.1. Model Setting

The vector autoregressive (VAR) model is a kind of multi-variable prediction algorithm, which is often used to analyze the possible interaction mechanism of variables in a period of time. Based on the VAR model, the TVP-VAR model assumes that the coefficient matrix and covariance matrix are time-varying, and uses variable parameters to calculate the results of the impulse response of each variable to different lag periods at all time points, so that the change characteristics of the relationship between variables can be analyzed over time, and the relationship between variables in various states can be studied regarding whether there are structural mutations. With reference to Nakajima et al.’s research [36], the TVP-VAR model of this paper is set as follows:

(3)$A{y}_t=F_1{y}_{t-1}+\cdots +F_s{T}_{t-s}+{\mu }_t,t=s+1,\cdots n,$

$y_t$ is the $\mathrm{k}\times 1$ dimensional observation vector, $A$ is the $\mathrm{k}\times \mathrm{k}$ dimensional observation vector, $F_1$,... $F_s$ is a $\mathrm{k}\times 1$ maintenance matrix, and the disturbance term is a $\mathrm{k}\times \mathrm{k}$ dimensional structural shock. Assuming ${\mu }_t$~N(0, ${\mathsf{\Sigma }}^2$), Σ = $\left[\begin{array}{cccc}{\mathsf{\sigma }}_1& 0& \cdots & 0\\ 0& \ddots & \ddots & ⋮\\ ⋮& \ddots & \ddots & 0\\ 0& \cdots & 0& {\mathsf{\sigma }}_{\mathrm{k}}\end{array}\right]$, the structural shock of the simultaneous relation is judged by recursion, and the simultaneous parameter matrix A is a lower triangular matrix:

$A=\left[\begin{array}{cccc}1& 0& \cdots & 0\\ {\mathrm{a}}_{21}& \ddots & \ddots & ⋮\\ ⋮& \ddots & \ddots & 0\\ {\mathrm{a}}_{\mathrm{m}1}& \cdots & {\mathrm{a}}_{\mathrm{m}1},{\mathrm{a}}_{\mathrm{m}-1}& 1\end{array}\right],$

The structural VAR model can be simplified as

${\mathrm{y}}_{\mathrm{t}}={\mathrm{B}}_1{\mathrm{y}}_{\mathrm{t}-1}+\cdots {+\mathrm{B}}_{\mathrm{s}}{\mathrm{y}}_{\mathrm{t}-\mathrm{s}}+{\mathrm{A}}^{-1}\mathsf{\Sigma }{\mathsf{\epsilon }}_{\mathrm{t}},{\mathsf{\epsilon }}_{\mathrm{t}}~\mathrm{N}(0,{\mathrm{I}}_{\mathrm{k}})$

where ${\mathrm{B}}_{\mathrm{i}}={\mathrm{A}}^{-1}{\mathrm{F}}_{\mathrm{i}},\mathrm{i}=1,\cdots$,s, the row elements in ${\mathrm{B}}_{\mathrm{i}}$ are stacked to form a ${\mathrm{k}}^2\mathrm{s}\times 1$ dimensional vector. Let ${\mathrm{X}}_{\mathrm{t}}={\mathrm{I}}_{\mathrm{k}}$⨂${({\mathrm{y}}^{\prime }}_{\mathrm{t}-1},\dots ,{\mathrm{y}}_{\mathrm{t}-\mathrm{s}}$), where ⨂ represents the Kronecker score; then, the model can be simplified as

(4)${\mathrm{y}}_{\mathrm{t}}={\mathrm{X}}_{\mathrm{t}}{\mathsf{\beta }}_{\mathrm{t}}{{+\mathrm{A}}_{\mathrm{t}}}^{-1}\sum {\mathsf{\epsilon }}_{\mathrm{t}}(\mathrm{t}=\mathrm{s}+1,\cdots ,\mathrm{n})$

At this time, the original formula becomes an SVAR model whose parameters do not change with time. Assuming that all parameters are time-varying, stack the non-0 and non-1 elements in $A_t$ as ${\mathrm{a}}_{\mathrm{t}}$ = (${\mathrm{a}}_{21,\mathrm{t}},{\mathrm{a}}_{31,\mathrm{t}},{\mathrm{a}}_{41,\mathrm{t}}$,${\mathrm{a}}_{\mathrm{m}1,\mathrm{m}-1}),{\mathrm{h}}_{\mathrm{t}}=({\mathrm{h}}_{1\mathrm{t}},{\mathrm{h}}_{2\mathrm{t}},\cdots ,{\mathrm{h}}_{\mathrm{k}\mathrm{t}}$), and have ${\mathrm{h}}_{\mathrm{j}\mathrm{t}}=\log {{(\mathsf{\sigma }}_{\mathrm{j}\mathrm{t}}}^2$). Let ${\mathsf{\beta }}_{\mathrm{t}}$, ${\mathrm{a}}_{\mathrm{t}}$, and ${\mathrm{h}}_{\mathrm{t}}$ obey the first-order random walk process and have no correlation. In this paper, the dynamic changes in model coefficients are expressed as follows:

(5)${\mathsf{\beta }}_{\mathrm{t}+1}={\mathsf{\beta }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathsf{\beta }\mathrm{t}}$

(6)${\mathsf{\alpha }}_{\mathrm{t}+1}={\mathsf{\alpha }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathsf{\alpha }\mathrm{t}}$

(7)${\mathrm{h}}_{\mathrm{t}+1}={\mathrm{h}}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{h}\mathrm{t}}$

(8)$\left(\begin{array}{c}{\mathsf{\epsilon }}_{\mathrm{t}}\\ {\mathsf{\mu }}_{\mathsf{\beta }\mathrm{t}}\\ {\mathsf{\mu }}_{\mathrm{a}\mathrm{t}}\\ {\mathsf{\mu }}_{\mathrm{h}\mathrm{t}}\end{array}\right)~\mathrm{N}\left(0,\left(\begin{array}{cccc}\mathrm{I}& 0& \cdots & 0\\ 0& \sum \mathsf{\beta }& \ddots & ⋮\\ ⋮& \ddots & \sum \mathsf{\alpha }& 0\\ 0& \cdots & 0& \sum \mathrm{h}\end{array}\right)\right)$

where I represents the identity matrix; ∑${\mathsf{\beta }}_{\mathrm{t}}$, ∑a, and ∑h are positive definite diagonal matrices. Based on the above assumptions, the TVP-VAR model can effectively capture potential structural changes. In order to obtain more accurate and robust empirical results, the Markov Monte Carlo (MCMC) method is used to estimate the parameters of the TVP-VAR model, which can overcome the problem of too many parameters to be estimated in nonlinear estimation.

In order to determine the stationarity of the time series data and avoid the pseudo-regression problem that may occur in the regression analysis, the unit root test is carried out on the logarithmic raw variable series, and the ADF method is used to test the stationarity of the data. It can be seen from Table 6 that the ADF statistics of LNOFDI, LNIMPORT, and LNEXPORT are greater than the critical value at the significance level of 1%, 5%, and 10%, indicating that the original sequence of the three variables is not stable. In order to meet the stationary data required by the self-vector regression model, the unit root test values of the ADF statistics of the three variables after first-order difference are all less than the critical values at the significance levels of 1%, 5%, and 10%, indicating that these three variables are stable after first-order difference and can be used in subsequent models.

5.2. Parameter Simulation Test

The MCMC sampling method was used to simulate 10,000 times to obtain valid samples, and the estimated results are shown in Figure 3. The sample autocorrelation coefficient showed a steady development trend, and the sample data were stable. The simulation path of the parameters showed significant fluctuation aggregation, and the different parameters all converged to the simulated tail. The estimation results of the sample mean once again demonstrated the robustness of the model estimation.

Table 7 shows the index values of the posterior distribution of parameters of the TVP-VAR model. It can be seen that the convergence diagnosis values of Geweke are all less than the critical value at a 95% significance level. At the 95% confidence level, the maximum value of the non-valid factor is 183.64, and it can be calculated that at least 54 (10,000/183.64) uncorrelated samples can be obtained under the condition of 10,000 continuous sampling, which means that the Markov chain can tend to converge in a more concentrated manner through 10,000 pre-sampling. It is proved that the estimation of this model is valid, and the dynamic relationship can be further explored.

5.3. Analysis of Time-Varying Characteristics of Lower Equal Interval Pulse Response in Different Lag Periods

In this paper, lag 4, 8, and 12 periods are selected to represent the short, medium, and long term, respectively, the horizontal axis represents the time node, and the vertical axis represents the impulse response intensity. As can be seen from Figure 4, the impulse response curves of China’s agricultural OFDI to RCEP countries’ agricultural trade in different lag periods all have obvious trends and fluctuations, which indicates that China’s agricultural OFDI has different effects at different time points and has obvious time-varying characteristics. In addition, the response curves of 4 months, 8 months, and 12 months of lag are obviously separated and have similar fluctuation characteristics.

First, when the positive impact of China’s agricultural OFDI is 1 percentage, if the lag period is 4 months, the impact of RCEP countries’ agricultural import trade from 2004 to 2021 will be positive. If the lag period is 8 and 12 months, there is a weak negative impact from 2004 to 2008, and the impact turns positive from 2008 to 2021. The possible reason is that in the early stage of outbound investment, many companies carry out investment activities when they lack sufficient estimation of overseas investment risks and a complex environment in order to seek higher return on capital, resulting in a rise in sunk costs. The return cannot reach the expected effect. After investing a certain cost, many companies pay more attention to the early market assessment, prudently make overseas investment decisions, attach importance to the localization of overseas project construction and risk control, etc., so that the foreign investment has a positive effect and stimulates foreign trade. It can be further concluded from Figure 4 that the short-term effect is greater than the medium-term effect, and the medium-term effect is greater than the long-term effect. This indicates that increasing the level of China’s agricultural OFDI will increase the RCEP countries’ imports of agricultural products from China, and this effect will gradually weaken with the extension of time.

Second, regardless of the short-, medium-, and long-term impulse response, the export trade effect produced by China’s agricultural OFDI shows a basically consistent trend, which means that RCEP countries’ agricultural product export trade to China will show a positive effect over a long period of time, with the short-term effect being the most significant, the medium-term effect being more significant, and the long-term effect being the least significant. The peak of the negative effect appeared between 2018 and 2020. A possible reason is that China’s agricultural products, due to the lack of “self-sufficiency rate” and the emphasis on “food security”, have an inhibitory effect on the agricultural product export trade of RCEP countries, and the outbreak of the novel coronavirus pneumonia in 2020 has encouraged trade protectionism in some countries. Countries have successively introduced restrictive foreign investment and trade policies, resulting in a decline in foreign direct investment in 2020, and a significant decline in the positive impact of their export trade.

Third, when RCEP countries’ agricultural trade is given a positive impact of 1 percentage, the impulse response of China’s OFDI shows a large jagged-like change, indicating that the impact of RCEP countries’ agricultural trade on China’s agricultural OFDI lags behind, and the impulse response with a lag of 4 months is the strongest and the most sensitive to the impact of China’s agricultural OFDI. The cumulative total effect of China’s agricultural OFDI showed positive fluctuations in the short, medium, and long term, reaching a peak in 2010 and 2016. It is further verified that China’s agricultural OFDI and RCEP countries have a two-way promotion effect on agricultural trade, but the import promotion effect is greater than the export promotion effect.

5.4. Impulse Response at Different Time Points

In order to determine the impact of major events, this paper selects three time points in January 2012, September 2013, and June 2018 for simulation, corresponding to the first proposal of the RCEP agreement, the first proposal of the “Belt and Road” development strategy, and the implementation of a series of trade protection policies against China by former US President Donald Trump. Figure 5 shows that the impulse response of China’s agricultural OFDI to RCEP countries’ agricultural trade at the three time points is generally characterized by a rapid response at first and then gradually stabilizing; that is, the impact of a 1 percentage positive impact of China’s agricultural OFDI at the three time points tends to be consistent. This indicates that the impact of China’s agricultural OFDI on agricultural trade of RCEP countries is relatively stable at different economic levels, and there will be no significant difference due to different time points.

The results of impulse response at different time points were further compared. First, the impact on agricultural trade was greater in January 2012, indicating that the proposal of the RCEP agreement is conducive to the bilateral agricultural trade between China and RCEP countries, and the proposal of the “Belt and Road” strategy has a certain promoting effect, but the difference is not obvious at different time points. China’s agricultural OFDI has a negative effect on agricultural trade in the current period. The OFDI began to show a positive effect after a lag of one period. Second, for the import impact of agricultural products, the positive impact effect reaches the maximum in the first phase, but immediately presents a trend of rapid decline, while the negative effect reaches the maximum in the third phase, and then maintains a certain weak positive effect. Thirdly, for the export effect of agricultural products, the impulse response of China’s agricultural OFDI at the three time points is obviously volatile, experiencing a negative effect in the early stage, turning to a positive effect in the later stage, and tending to a weak positive effect after the sixth stage.

5.5. Heterogeneity Analysis of Impulse Response

5.5.1. Impulse Response to RCEP Countries’ Agricultural Imports

Research on China’s OFDI in RCEP member countries reveals significant heterogeneity in the influencing factors across these states. Specifically, market size and labor resources exert a positive influence on OFDI [37], whereas elements such as tax burden and natural resource endowment exhibit certain inhibitory effects. In nations that have successfully attracted substantial Chinese investments, political stability plays a crucial role in shaping OFDI patterns [38]. Geographical and institutional distances generally facilitate OFDI; however, economic distance demonstrates varying impacts among RCEP countries [39]. The increase in GDP growth rates and per capita GDP of host nations has also stimulated Chinese investment, while fluctuations in per capita net national income, per capita GDP growth rate, and trade in goods and services have had a constraining effect on such investments [40]. These findings underscore the intricate interplay of factors affecting China’s OFDI within RCEP countries and emphasize the necessity for developing region-specific investment strategies.

As can be seen from Figure 6, China’s agricultural OFDI has the most significant positive impact on Laos, as well as on Malaysia, Thailand, and Vietnam. In other words, agricultural OFDI further improves trade openness and stabilizes China’s import market of key commodities and commodities. Laos and Vietnam share borders with China, making it more convenient for foreign investment and agricultural trade [41]. In particular, Vietnam is a major importer of fish meal used as feed for key commodities, Thailand is a major importer of rice, chicken, and its by-products, and Malaysia is a major importer of natural rubber. The impact effect of China’s agricultural OFDI on other RCEP countries is obviously volatile. Malaysia, Vietnam, Thailand, Singapore, and Laos show a continuous positive effect, South Korea and Myanmar show a continuous negative effect, and the impact on Cambodia, Japan, the Philippines, and Brunei fluctuates around a value of 0. This indicates that, compared with other countries, China’s agricultural OFDI can significantly promote import trade in Malaysia, Vietnam, Thailand, Singapore, and Laos. In addition, the impact effect of China’s agricultural OFDI on Vietnam, Thailand, Singapore, Laos, and Australia is significant in the short term, while the impact on Malaysia, South Korea, Myanmar, and Japan is significant in the long term.

5.5.2. Impulse Response of Agricultural Exports to RCEP Countries

Similar to the impulse response effect of imports, it can be seen from Figure 7 that China’s agricultural OFDI can easily promote the agricultural product export of RCEP countries, help to improve trade openness, and stabilize China’s export market in the field of important commodities. Among them, the positive impact effect on Thailand and Australia is significant, and it also has a significant positive impact on Vietnam and Indonesia. Thailand is one of the top three markets for China’s export of peanut kernels, cuttlefish and squid, and other commodities, while Indonesia is the main market for China’s export of tobacco, pulp, and palm oil. In terms of the volatility of the impact effect, Vietnam, Thailand, Indonesia, and Japan have a continuous positive effect, South Korea has a continuous negative effect, and the impact of Malaysia, Cambodia, and Brunei shows a significant zigzag effect. At the same time, there are obvious differences in the impact effects of different countries in different periods. China’s agricultural OFDI significantly increases the export volume of agricultural products to Vietnam, Thailand, Indonesia, and Australia in the short term, while the long-term impact effect on the Philippines and Brunei is more significant.

6. Conclusions and Implications

6.1. Conclusions

Based on the trade gravity model, this paper analyzes the sustainability trade effect of China’s agricultural OFDI on the agricultural products of RCEP member countries, and further analyzes the dynamic impact results between China’s agricultural OFDI and RCEP member countries’ agricultural products trade by using the TVP-VAR model. Important conclusions can be drawn as follows:

First, based on the analysis of sample global regression results, China’s agricultural OFDI has significant positive spillover effects on RCEP member countries. A 1% increase in China’s agricultural OFDI has a significant positive impact on the import and export of agricultural products of RCEP countries in the short, medium, and long term, and there is a continuous positive impact effect, which reflects that “China’s agricultural development is good, and the agricultural development of RECP member countries is good” to a certain extent. The reason for the existence of this trade creation effect may lie in that, on the one hand, in order to find low-cost agricultural production means, Chinese agricultural enterprises realize the transfer of agricultural production through international trade and technology export, which brings a trade creation effect to the agricultural trade of RECP countries and generates export demand. On the other hand, China participates in the international agricultural industry chain through the export of capital and technology. The agricultural products of RCEP countries are sensitive to the price of the international market, while China’s huge and stable consumer market produces an import transfer effect. As RCEP countries are mainly developing countries, they still face food security problems, and the problems of poverty and hunger are far from being solved. Bilateral trade and investment in agricultural products is of great significance to ensuring food security and livelihood security for all participants.

Second, according to the quantile regression results, although agricultural OFDI has a limited impact on agricultural trade between China and RCEP countries when it reaches a certain level, in general, agricultural OFDI maintains a positive and significant impact, and agricultural OFDI will still have a trade-creating effect in the future. The reason is that both in the short and long term, the tariff rates and non-tariff barriers between China and RCEP countries are already low, and China’s agricultural OFDI has little impact on the bilateral trade partnership. Therefore, encouraging China’s competitive agricultural enterprises to “go global” through OFDI strategy can reduce trade frictions and explore foreign markets, and deepen cooperation and innovation with RCEP countries in the agricultural industrial chain, supply chain, and value chain.

Thirdly, from the analysis of impulse response results, although China’s agricultural OFDI will show the alternating phenomenon of an export creation effect and import transfer effect in different periods, the total effect of one standard deviation impact from China’s agricultural OFDI causes the import promotion effect to be greater than the export promotion effect; that is, the “spillover effect” has an increasing trend. This means that China’s agricultural OFDI not only improves the ability of Chinese agricultural enterprises to allocate global agricultural resources, but also improves the ability of RCEP countries to shape and expand the agricultural industry, providing new opportunities for it to explore the Chinese consumer market, and forming a “mutually beneficial and win-win” situation. Through the heterogeneity results of impulse response, it is further found that China’s agricultural OFDI has a more significant trade creation effect on low-income developing countries and neighboring Southeast Asian countries. China should pay attention to sharing agricultural development experience and practical agricultural technology with RCEP countries, so as to promote the development of modern agriculture and increase farmers’ income in RCEP countries through OFDI.

The following points can be further discussed in this study: Referring to existing literature, this paper mainly studies the selection variables affected by foreign investment and trade. The impact of policy changes has been taken into account, and non-economic factors such as FTA accession, territorial border, and language have been introduced into the model as control variables. However, there are many influencing factors among international trade, and the influence of policy changes and non-economic factors (such as natural disasters and political conflicts) between countries can be classified for further discussion. In this paper, soybean, corn, wheat, oat, soybean meal, and rice are selected as representatives to analyze the impact of foreign investment on its trade volume and discuss the impact from 2004 to 2019. The data cover a long period of time and have certain reference opinions. However, due to the availability of data, this paper does not adequately discuss the impact in recent years following the outbreak, which could be further supplemented.

6.2. Implication

According to the above research conclusions, this paper has the following policy implications: First, actively adjust the international trade strategy, expand the scope of international trade, take advantage of China’s OFDI advantages in RCEP countries, greatly improve the OFDI level of Chinese enterprises, and make breakthroughs in related industries such as agricultural machinery, commercialization of agricultural science and technology, and cross-border marketing of agricultural products. Second, attach importance to the export creation effect and import transfer effect of China’s agricultural OFDI, and combine with the upgrading and adjustment of agricultural industrial structure of RCEP member countries to form misplaced competition and achieve mutual benefit and a win–win situation. Third, take into account the actual characteristics of RCEP countries to achieve functional complementarity between the two countries, so as to comprehensively stimulate investment in different industrial types. It will have a driving effect on RCEP countries and serve as a positive example, further raising the sustainability of economic growth of China and RCEP member countries.

Author Contributions

Supervision, project administration, funding acquisition, Q.M.; conceptualization, methodology, software, formal analysis, resources, data curation, writing—original draft preparation, X.W. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. Trade in agricultural products between China and RCEP countries (2004–2023). Units: per billion USD.

Figure 2. Trade dependence between China and RCEP member countries (2004–2021).

Figure 3. Parameter estimation results.

Figure 4. Impulse response function of different lag periods. (a) OFDI–OFDI. (b) OFDI–IMPORT. (c) OFDI–EXPORT. (d) IMPORT–OFDI. (e) IMPORT–IMPORT. (f) IMPORT–EXPORT. (g) EXPORT–OFDI. (h) EXPORT–IMPORT. (i) EXPORT–EXPORT. Note: Short and green, long and purple, and solid and red lines indicate short (4 months), medium (8 months), and long (12 months), respectively.

Figure 5. Impulse response of China’s agricultural OFDI to RCEP countries’ agricultural trade at different time points. (a) OFDI–OFDI. (b) OFDI–IMPORT. (c) OFDI–EXPORT. (d) IMPORT–OFDI. (e) IMPORT–IMPORT. (f) IMPORT–EXPORT. (g) EXPORT–OFDI. (h) EXPORT–IMPORT. (i) EXPORT–EXPORT. Note: The plus sign (+), cross ([Forumla omitted. See PDF.]), and triangle (▲) indicate January 2012, September 2013, and June 2018, respectively.

Figure 6. Impulse response to agricultural imports of RCEP countries. (a) Malaysia. (b) Vietnam. (c) Thailand. (d) Cambodia. (e) Burma. (f) Korea. (g) Japan. (h) The Philippines. (i) Indonesia. (j) Australia. (k) New Zealand. (l) Singapore. (m) Laos. (n) Brunei. Note: Short and green, long and purple, and solid and red lines indicate short (4 months), medium (8 months), and long (12 months), respectively.

Figure 6. Impulse response to agricultural imports of RCEP countries. (a) Malaysia. (b) Vietnam. (c) Thailand. (d) Cambodia. (e) Burma. (f) Korea. (g) Japan. (h) The Philippines. (i) Indonesia. (j) Australia. (k) New Zealand. (l) Singapore. (m) Laos. (n) Brunei. Note: Short and green, long and purple, and solid and red lines indicate short (4 months), medium (8 months), and long (12 months), respectively.

Figure 7. Impulse response to RCEP countries’ exports. (a) Malaysia. (b) Vietnam. (c) Thailand. (d) Cambodia. (e) Burma. (f) Korea. (g) Japan. (h) The Philippines. (i) Indonesia. (j) Australia. (k) New Zealand. (l) Singapore. (m) Laos. (n) Brunei. Note: Short and green, long and purple, and solid and red lines indicate short (4 months), medium (8 months), and long (12 months), respectively.

Table 1

Variable names and their meanings.

Variables	Meaning
IMPORT	The annual value of China’s agricultural imports to RCEP members.
EXPORT	China’s annual agricultural exports to RCEP member countries.
OFDI	Annual flow of China’s agricultural direct investment in RCEP member countries
GDP	GDP of RCEP member countries, reflecting the level of economic development of the host country.
DIST	The straight-line distance between Beijing, China, and the capitals of RCEP member countries. Distance is generally considered to reflect the cost of trade. It is calculated by multiplying the straight-line distance between two capitals by the price of Brent crude oil in each year.
CONTIG	Dummy variable, indicating whether China borders the territory of the host country. Yes = 1, no = 0.
COMLANG	Dummy variable that indicates whether the host country shares an official language with China. Yes = 1, no = 0.
MARKET	The proportion of RCEP member countries’ agricultural trade with China in the member countries’ total agricultural trade.
FTA	Dummy variable, indicating whether an RCEP member has joined the WTO. Yes = 1, no = 0.

Table 2

The results of baseline regression.

	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Variables	Import			Export
lnOFDI	0.799 ***	0.718 ***	0.723 ***	0.532 ***	0.263 ***	0.273 ***
	(0.076)	(0.066)	(0.067)	(0.061)	(0.049)	(0.048)
InGDP		0.982 ***	1.007 ***		1.144 ***	1.134 ***
		(0.071)	(0.072)		(0.052)	(0.052)
lnDIST		0.483 ***	0.719 ***		0.883 ***	1.136 ***
		(0.134)	(0.157)		(0.099)	(0.113)
CONTIG		1.048 ***	1.145 ***		0.306 *	0.193
		(0.233)	(0.236)		(0.173)	(0.170)
COMLANG		0.607 *	0.741 **		−0.380	−0.312
		(0.360)	(0.369)		(0.270)	(0.266)
MARKET		14.06 ***	13.09 ***		4.630 ***	4.149 ***
		(1.747)	(1.811)		(1.295)	(1.307)
FTA		2.214 ***	2.157 ***		1.156 ***	1.291 ***
		(0.249)	(0.253)		(0.184)	(0.183)
year	no	no	yes	no	no	yes
Constant	1.247	55.14 ***	53.80 ***	4.506 ***	1.010	−4.910
	(0.8880)	(9.1600)	(11.6900)	(0.7180)	(6.7870)	(8.4310)
Observations	224
R-squared	0.3320	0.8520	0.8610	0.2520	0.8610	0.8760

Note: ***, **, and * indicate a significance level of 1%, 5%, and 10%, respectively.

Table 3

Results of robust regression.

	Model (7)	Model (8)	Model (9)	Model (10)	Model (11)	Model (12)
Variables	Import			Export
LNOFDI	0.081 ***	0.090 ***	0.090 ***	0.052 ***	0.029 ***	0.030 ***
	(0.012)	(0.015)	(0.015)	(0.005)	(0.005)	(0.004)
LnGDP		0.109 ***	0.110 ***		0.113 ***	0.112 ***
		(0.008)	(0.008)		(0.004)	(0.004)
LnDIST		0.029 *	0.039 **		0.073 ***	0.095 ***
		(0.016)	(0.019)		(0.010)	(0.011)
CONTIG		0.137 ***	0.143 ***		0.041 **	0.031 *
		(0.025)	(0.025)		(0.017)	(0.016)
COMLANG		−0.059	−0.061		−0.024	−0.020
		(0.039)	(0.038)		(0.025)	(0.025)
MARKET		1.433 ***	1.400 ***		0.338 ***	0.295 **
		(0.229)	(0.221)		(0.121)	(0.119)
FTA		0.243 ***	0.240 ***		0.126 ***	0.140 ***
		(0.029)	(0.027)		(0.017)	(0.017)
Constant	1.399 ***	9.022 ***	2.480 ***	1.758 ***	1.870 ***	0.181
	(0.153)	(1.719)	(0.370)	(0.070)	(0.578)	(0.157)
R-squared	0.297	0.761	0.862	0.236	0.826	0.873
Log pseudolikelihood			−492.8			−477.9
Wald chi2			436.5			1 479.0
Prob > chi2			0.000			0.000

Note: ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively.

Table 4

Quantile regression results (export trade).

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Variables	10%	20%	30%	40%	50%	60%	70%	80%	90%
LnOFDI	0.356 ***	0.285 ***	0.202 ***	0.181 ***	0.200 ***	0.265 ***	0.122	0.137 *	0.236 **
	(0.088)	(0.081)	(0.075)	(0.067)	(0.075)	(0.093)	(0.086)	(0.071)	(0.104)
LnGDP	0.792 ***	0.804 ***	1.037 ***	1.128 ***	1.159 ***	1.134 ***	1.144 ***	1.162 ***	1.156 ***
	(0.056)	(0.094)	(0.117)	(0.088)	(0.082)	(0.079)	(0.073)	(0.064)	(0.057)
LnDIST	0.931 ***	0.910 ***	0.794 ***	0.782 ***	0.832 ***	0.990 ***	0.806 ***	0.795 ***	0.703 ***
	(0.070)	(0.178)	(0.172)	(0.137)	(0.132)	(0.127)	(0.108)	(0.111)	(0.196)
CONTIG	0.708 *	−0.361	0.307	0.349	0.445	0.322	0.476 ***	0.502 ***	0.399 ***
	(0.360)	(0.382)	(0.333)	(0.301)	(0.313)	(0.242)	(0.181)	(0.133)	(0.113)
COMLANG	0.221	0.347	0.065	0.008	−0.215	−0.395	−0.301	0.815	0.742
	(0.237)	(0.287)	(0.303)	(0.266)	(0.298)	(0.432)	(0.675)	(0.745)	(0.540)
MARKET	0.003	−0.038	0.713 *	1.041 ***	1.313 ***	1.524 ***	1.484 ***	1.670 ***	1.850 ***
	(0.467)	(0.459)	(0.394)	(0.284)	(0.287)	(0.340)	(0.328)	(0.248)	(0.266)
FTA	9.293	10.14	3.850	−3.921	−3.730	4.631	−8.635	−7.148	2.728
	(15.38)	(14.52)	(12.71)	(10.19)	(9.662)	(10.84)	(10.48)	(9.215)	(10.59)
Observations	224

Note: ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively.

Table 5

Quantile regression results (import trade).

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Variables	10%	20%	30%	40%	50%	60%	70%	80%	90%
LnOFDI	1.151 ***	1.002 ***	0.703 ***	0.5270 ***	0.492 **	0.431 **	0.366 **	0.450 ***	0.137
	(0.171)	(0.228)	(0.218)	(0.1890)	(0.200)	(0.175)	(0.154)	(0.161)	(0.211)
LnGDP	1.204 ***	1.181 ***	1.283 ***	1.3080 ***	1.299 ***	1.241 ***	1.143 ***	1.040 ***	1.086 ***
	(0.105)	(0.085)	(0.072)	(0.0820)	(0.102)	(0.108)	(0.116)	(0.091)	(0.058)
LnDIST	0.283	0.556 **	0.905 ***	0.9520 ***	0.834 ***	0.742 ***	0.721 ***	0.806 ***	0.406
	(0.190)	(0.237)	(0.257)	(0.2160)	(0.171)	(0.144)	(0.154)	(0.229)	(0.272)
CONTIG	1.535 ***	1.449 ***	1.596 ***	1.5040 ***	1.198 ***	0.780 **	0.584 **	0.325 *	0.115
	(0.328)	(0.265)	(0.265)	(0.3300)	(0.369)	(0.353)	(0.266)	(0.195)	(0.174)
COMLANG	2.082 ***	2.159 ***	2.151 ***	1.956 ***	1.784 **	2.218 ***	1.813 *	−0.812	1.121 *
	(0.372)	(0.474)	(0.469)	(0.6010)	(0.755)	(0.848)	(1.011)	(0.861)	(0.623)
MARKET	2.841 ***	2.879 ***	2.326 ***	2.1910 ***	2.275 ***	2.297 ***	2.099 ***	1.973 ***	1.385 **
	(0.500)	(0.472)	(0.565)	(0.5840)	(0.661)	(0.581)	(0.497)	(0.623)	(0.581)
FTA	77.23 ***	66.16 ***	32.53 *	16.86	10.66	−3.441	−3.215	3.235	−14.88
	(16.11)	(20.24)	(19.39)	(15.61)	(18.55)	(15.79)	(12.23)	(13.19)	(16.55)
Observations	224

Note: ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively.

Table 6

Results of ADF test.

Variable	ADF Statistics	1% Critical Value	5% Critical Value	10% Threshold	p Value	Smoothness
LNOFDI	−1.813	−4.992	−3.875	−3.388	0.633	Non-smooth
DLNOFDI	−5.767	−4.004	−3.098	−2.690	0.000	Smooth
LNIMPORT	−1.236	−3.461	−2.874	−2.574	0.658	Non-smooth
DLNIMPORT	−15.65	−3.461	−2.874	−2.575	0.000	Smooth
LNEXPORT	−1.859	−3.464	−2.876	−2.574	0.351	nonstationarity
DLNEXPORT	−8.761	−2.577	−1.942	−1.615	0.000	Smooth

Table 7

Parameter estimation results of TVP-VAR model.

Parameter	Mean	Stdev	95%L	95%U	Geweke	Inef.
sb1	0.021	0.002	0.018	0.027	0.135	10.20
sb2	0.018	0.001	0.015	0.021	0.654	8.990
sa1	0.087	0.062	0.042	0.202	0.215	122.7
sa2	0.085	0.045	0.040	0.214	0.771	110.5
sh1	0.736	0.139	0.510	1.032	0.576	183.6
Sh2	0.655	0.117	0.446	0.905	0.492	47.96

References

1. Mundell, R.A. International trade and factor mobility. Am. Econ. Rev.; 1957; 47, pp. 321-335.

2. Kojima, K. International Trade and Foreign Investment: Substitutes or Complements. Hitotsubashi J. Econ.; 1975; 75, pp. 1-12.

3. Dunning, J.H. The eclectic paradigm of international production: A restatement and some possible extensions. J. Int. Bus. Stud.; 1988; 19, pp. 1-31. [DOI: https://dx.doi.org/10.1057/palgrave.jibs.8490372]

4. Benwu, X. The Effects of Outward Foreign Direct Investment on International Trade: Evidence from China. Financ. Trade Econ.; 2009; 4, pp. 77-82.

5. Niu, R.; Zhou, X. China’s Overseas Direct Investment and its Influence on Trade: An Empirical Study Based on Panel Gravity Model. Int. Econ. Trade Res.; 2012; 28, pp. 69-81.

6. Han, J.; Xu, Y.-Y. RCEP and the Integration of Asia-Pacific Trade Agreements—A Quantitative Study on the Agreement Text. China Ind. Econ.; 2021; 7, pp. 81-99.

7. Lin, G.; Chen, T. The Impact of Outward FDI on the Improvement of Resource Allocation Efficiency: Analysis Based on Agricultural Processing Industry. World Econ. Stud.; 2020; 7, pp. 46-59.

8. Wang, Y.; Xie, L.; Zhang, Y.; Wang, C.; Yu, K. Does FDI promote or inhibit the high-quality development of agriculture in China? An agricultural GTFP perspective. Sustainability; 2019; 11, 4620. [DOI: https://dx.doi.org/10.3390/su11174620]

9. Morea, D.; Balzarini, M. Financial sustainability of a public-private partnership for an agricultural development project in Sub-Saharan Africa. Zemed. Ekon.; 2018; 64, pp. 389-398. [DOI: https://dx.doi.org/10.17221/161/2017-AGRICECON]

10. Jiang, X.; Chen, Y. The Impact of Agricultural Foreign Investment Outflow on Home Countries’ Participation in Global Value Chains—From the Perspective of Social Network Analysis. J. Huazhong Agric. Univ.; 2023; 3, pp. 44-53.

11. Fan, S.; Zhu, Y.; Fang, X. Big Food Vision and food security in China. Agric. Rural Stud.; 2023; 1, 0001. [DOI: https://dx.doi.org/10.59978/ar01010001]

12. Lu, Y.; Lu, Y.; Zeng, K.; Li, Y. China’s Outward Foreign Direct Investment and the Margins of Trade: Empirical Evidence from “One Belt, One Road” Countries. China Int. J.; 2018; 16, pp. 129-151. [DOI: https://dx.doi.org/10.1353/chn.2018.0006]

13. Osabuohien-Irabor, O.; Drapkin, I.M. Outward FDI and international trade: The study of causal effects. Proceedings of the Economics and Finance Conferences; Prague, Czech Republic, 29 July 2021; No. 12513347 International Institute of Social and Economic Sciences: London, UK, 2021.

14. Amanor, K.S.; Chichava, S. South–south cooperation, agribusiness, and African agricultural development: Brazil and China in Ghana and Mozambique. World Dev.; 2016; 81, pp. 13-23. [DOI: https://dx.doi.org/10.1016/j.worlddev.2015.11.021]

15. Santangelo, G.D. The impact of FDI in land in agriculture in developing countries on host country food security. J. World Bus.; 2018; 53, pp. 75-84. [DOI: https://dx.doi.org/10.1016/j.jwb.2017.07.006]

16. Du, Y.; Zhang, X.; Liang, T. National Food Security and the Construction of New Agricultural Development Paradigm. Issues Agric. Econ.; 2022; pp. 4-10.

17. Hsiao, F.S.T.; Hsiao, M.C.W. FDI, exports, and GDP in East and Southeast Asia—Panel data versus time-series causality analyses. J. Asian Econ.; 2006; 17, pp. 1082-1106. [DOI: https://dx.doi.org/10.1016/j.asieco.2006.09.011]

18. Li, J.; Strange, R.; Ning, L.; Sutherland, D. Outward foreign direct investment and domestic innovation performance: Evidence from China. Int. Bus. Rev.; 2016; 25, pp. 1010-1019. [DOI: https://dx.doi.org/10.1016/j.ibusrev.2016.01.008]

19. Zhou, Y.; Jiang, J.; Ye, B.; Hou, B. Green spillovers of outward foreign direct investment on home countries: Evidence fromChina’s province-level data. J. Clean. Prod.; 2019; 215, pp. 829-844. [DOI: https://dx.doi.org/10.1016/j.jclepro.2019.01.042]

20. Tang, B.; Chang, Y. The Measurement and Development of China’s Agriculture Value Chain: Based on the Empirical Analysis of Asia-Pacific Region. Issues Agric. Econ.; 2019; 10, pp. 50-62.

21. Tianbao, G.; Xiaolin, W.; Yunfeng, W. Regional Comparison and Development Strategy of Technical Complexity in Exporting China’s Agricultural Products. Rural Econ.; 2022; 1, pp. 10-19.

22. Chichava, S.; Duran, J.; Cabral, L.; Shankland, A.; Buckley, L.; Lixia, T.; Yue, Z. Brazil and China in Mozambican agriculture: Emerging insights from the field. Ids Bull.; 2013; 44, pp. 101-115. [DOI: https://dx.doi.org/10.1111/1759-5436.12046]

23. Qiang, Y. How does OFDI affect trade development in developing countries?. J. Chin. Econ. Foreign Trade Stud.; 2013; 6, pp. 85-90. [DOI: https://dx.doi.org/10.1108/JCEFTS-05-2013-0020]

24. Huangfu, S.; Zhu, J.; Zou, H.; Wang, Y. Decomposition of Extensive and Intensive Trade Margin Impacts of China’s Agricultural OFDI. Adv. Econ. Manag. Political Sci.; 2024; 67, pp. 240-251.

25. Badu-Prah, C.; Agyeiwaa-Afrane, A.; Gidiglo, F.K.; Srofenyoh, F.Y.; Agyei-Henaku, K.A.A.; Djokoto, J.G. Trade, Foreign Direct Investment and Agriculture in Developing Countries. Res. World Agric. Econ.; 2023; 4, pp. 1-14. [DOI: https://dx.doi.org/10.36956/rwae.v4i3.861]

26. Chen, Q.; Guo, P. Outward foreign direct investment in agriculture by Chinese companies: Land grabbing or win–win?. Econ. Political Stud.; 2017; 5, pp. 404-420. [DOI: https://dx.doi.org/10.1080/20954816.2017.1384607]

27. Kaarhus, R. Land, investments and public-private partnerships: What happened to the Beira Agricultural Growth Corridor in Mozambique?. J. Mod. Afr. Stud.; 2018; 56, pp. 87-112. [DOI: https://dx.doi.org/10.1017/S0022278X17000489]

28. Benhai, G.U.; Zixing, W.A.G.; Fei, W.A.G. Research on China’s Export Trade Efficiency of RCEP Agricultural Products from the Perspective of FTA Depth. World Agric.; 2023; pp. 9-23.

29. Anderson, J.E. A theoretical foundation for the gravity equation. Am. Econ. Rev.; 1979; 69, pp. 106-116.

30. Tang, H.; Zhang, H. International Intra-product Specialization and Developing Countries’ Upgrading Value Chain. Econ. Res. J.; 2009; 44, pp. 81-93.

31. Fang, Y.; Fan, Z.; Li, X. The Impact of Trade Facilitation on Global Agricultural Trade and Its Implications for China. Issues Agric. Econ.; 2022; pp. 122-133.

32. Silva, J.M.C.S.; Tenreyro, S. The log of gravity. Rev. Econ. Stat.; 2006; 88, pp. 641-658. [DOI: https://dx.doi.org/10.1162/rest.88.4.641]

33. Liu, J.; Xiong, Y. Energy Price Difference, Comparative Advantage and Chinese Export Trade: On PPML Model with International Panel Data. Financ. Trade Res.; 2018; pp. 76-87.

34. Petri, P.A. The regional clustering of foreign direct investment and trade. Transnatl. Corp.; 1995; 3, pp. 1-24.

35. Lin, C.-W.; Tan, N.; He, C.-T. A Study on Trade Effect of China’s Outward Foreign Direct Investment in ASEAN. Int. Econ. Trade Res.; 2019; 35, pp. 60-79.

36. Nakajima, J. Time-varying parameter VAR model with stochastic volatility: An overview of methodology and empirical applications. Monet. Econ. Stud.; 2011; 12, pp. 107-142.

37. Qing, X. Analysis on Location Choice of China’s OFDI in RCEP Countries. Highlights Bus. Econ. Manag.; 2023; 11, pp. 302-307. [DOI: https://dx.doi.org/10.54097/hbem.v11i.8147]

38. Deng, X.; Tang, W.; Zhu, H.; Xing, Z. Influencing factors of China’s direct investment in RCEP countries: Evidence from panel quantile regression. Appl. Econ.; 2022; 55, pp. 3347-3364. [DOI: https://dx.doi.org/10.1080/00036846.2022.2114996]

39. Galluzzo, N. Emigration and common Agricultural policy in Romanian Rural Areas: An analysis of Patterns of Inefficiency. Agric. Rural Stud.; 2023; 1, 0003. [DOI: https://dx.doi.org/10.59978/ar01010003]

40. Zhong, X. Study on the Influencing Factors of China’s Outward FDI in RCEP Member Countries. Adv. Econ. Manag. Political Sci.; 2023; 51, pp. 12-24. [DOI: https://dx.doi.org/10.54254/2754-1169/51/20230601]

41. Jorgenson, A.K.; Kuykendall, K.A. Globalization, Foreign Investment Dependence and Agriculture Production: Pesticide and Fertilizer Use in Less-Developed Countries, 1990–2000. Soc. Forces; 2008; 87, pp. 529-560. [DOI: https://dx.doi.org/10.1353/sof.0.0064]