Effects of cross-border railway on agricultural value chain linkages: Evidence from the China-Europe Railway Express

Abstract

As an important part of “the Belt and Road” initiative, the CR Express provides a new land route for economic and trade exchanges between economies along the route. Based on the data of Eora26 multi-region and multi-industry input-output database, this paper measures the agricultural value chain linkages between China and foreign countries from the perspective of motion dimension, and uses the difference-in-differences model to explore the agricultural value chain linkages between China and countries along the railway line. The results show that: the operation of CR Express has a significant effect on enhancing bilateral agricultural value chain linkages. From the perspective of value-added inflow, the CR Express can substantially improve agricultural value chain linkages by extending the length of the agricultural industry chain between China and other countries. There is regional heterogeneity in the agricultural value chain linkage effect of the CR Express, compared to non-adjacent countries, the linkage effect is significantly stronger for countries adjacent to the railway's route. Therefore, it is essential to focus on the deep integration of the CR Express with “the Belt and Road” Initiative, segment international markets, and achieve complementary advantages in the agricultural industry through the extension of the agricultural industry chain.

1. Introduction

Since the implementation of “the Belt and Road” initiative, the connectivity of the agricultural product markets between China and Europe economies has been growing increasingly closer, and their price relationships have shown new characteristics and trends [1]. On the one hand, as the division of labor in global value chains deepens, more economies are leveraging China's resources and market advantages to develop agricultural supply chains, and strengthen control over the Global Value Chain [2]. On the other hand, trade frictions have prompted China to strengthen its agricultural cooperation with “the Belt and Road” countries, increasing the correlation of agricultural product prices between them, and significantly promoting the level of regional cooperation with Europe. Furthermore, with the advancement of trade liberalization, the impact of transportation infrastructure on national trade becomes more and more evident, and the economic development of countries increasingly depends on infrastructure [3].

As a key transportation hub connecting the Eurasian continent, the China-Europe Railway Express (CR Express) closely links the economies of East Asia, Europe, and the countries along its route. It not only deepens China's economic and trade cooperation with these countries, but also accelerates the implementation of “the Belt and Road” Initiative. In the context of economic globalization, the foundational role of agriculture is becoming increasingly prominent, and agricultural trade has become a vital part of the global trade network. As a major agricultural country, China needs to integrate its domestic agricultural resources, and explore international agricultural product markets to better integrate into the global agricultural trade network. In this process, the CR Express, as an important carrier of cross-border railway transportation, has become a crucial channel for China to participate in global division of labor and integrate into international markets.

As of October 2023, the CR Express has operated 77,000 trains, with the cumulative trade value exceeding USD 340 billion, connecting more than 200 cities in 25 countries and regions in Europe.1 Especially during the COVID-19 pandemic, the CR Express maintained stable operations, playing a positive role in stabilizing China's import and export trade, and ensuring the normal and orderly production and people's daily life. Although the CR Express currently mainly transports industrial products, its fast, stable, and cost-effective characteristics present immense potential for agricultural product transportation. With the advancement of “the Belt and Road” initiative, the CR Express is expected to optimize agricultural supply chains, reduce transportation costs, and elevate the position of Chinese agricultural products in the global value chain. The operation of the CR Express may reshape the agricultural trade landscape between China and Europe, forming new trade routes and logistics networks, and promoting sustainable growth in bilateral agricultural trade.

Based on the above context, this paper treats the launch of the CR Express as a quasi-natural experiment, and empirically examines the impact of cross-border railways on agricultural value chains, using the difference-in-difference method. It further delves into the potential mechanisms behind this impact. The aim of the paper is to provide an econometric model to measure the influence of the CR Express on bilateral agricultural trade, revealing its actual effects on agricultural trade. This study offers a new perspective on understanding the complex relationship between infrastructure, logistics systems, and agricultural development in the context of globalization. The findings will provide significant references for formulating transnational transportation policies, agricultural policies, and regional development strategies, thereby contributing to China's high-quality advancement of “the Belt and Road” Initiative, promoting international agricultural cooperation, and enhancing agricultural trade.

The contribution of this paper lies in two aspects. First, in terms of metric development, this paper innovatively introduces a cross-border frequency index from a value-added perspective. Based on this index, the paper calculates the value-multiplier effect brought by each unit of value-added in cross-border agricultural trade, thereby representing the depth of bilateral trade. Compared with traditional methods that rely on trade volume (such as import and export quantities) to assess trade growth, this approach better reveals the inherent complexity and dynamic changes in trade relationships between countries, offering greater explanatory power. Second, the research background aligns with China's “Westward Opening” strategy, especially in the context of land transportation serving as a crucial link connecting inland China to the rest of the world, the role of railway infrastructure construction has become increasingly prominent. By examining the sustained trade growth and interaction effects brought about by railway infrastructure constructions, this paper provides strong empirical support for evaluating the economic returns and social benefits of infrastructure investment. It also offers practical guidance for the future construction and optimization of land trade corridors.

2. Literature review

At present, with the ongoing strengthening of the trend of economic globalization, a new international division of labor system based on the Global Value Chain has been continuously developing. There is a mutual interdependence between the value-added of various countries. The division of labor within the Global Value Chain has gradually become the new norm in the international division of labor system [4]. Research on value chain issues has become a hot topic in international trade matters, with many scholars incorporating vertical integration into trade models and conducting a series of studies on value chains.

The research relevant to this paper mainly focuses on two aspects: first, studies on value chain linkages, primarily concerning the measurement of value chain linkage indicators. Scholars using input-output methods to decompose value-added from three dimensions: value, position, and movement, to obtain indicators of value chain linkages [[5], [6], [7], [8]]. In empirical studies, some scholars measured value chain linkages among OECD countries and analyzed the impact of the internet on bilateral value chain linkages using bilateral and bidirectional network linkage data from a value perspective. The research results indicate that the internet can become a significant driver of value chain linkages between two countries by reducing trade costs, shortening transaction times, and extending production steps [9]. From a positional perspective, Xiao et al. measured the position of Chinese manufacturing in the Global Value Chain, empirically analyzed the promoting effects of technological efficiency and total factor productivity on value chain upgrading [10]. Additionally, some scholars analyzed agricultural value chain linkages between China and countries along “the Belt and Road” as well as other major economies, the results suggested that “the Belt and Road” initiative can increase trade flows among participating countries [11], “the Belt and Road” countries also have more Global Value Chain participation [12]. Meanwhile, “the Belt and Road” initiative can increase the degree of value chain linkages for countries (regions) along the route and even for other external countries (regions) [13]. Furthermore, scholars have enriched discussions on the factors influencing the enhancement of value chain linkages. They believe that the functional and spatial separation of the production process can lead to changes in the length of the industry chain [14], resulting in an increase in the number of transaction links required to produce the final product, ultimately leading to a decrease in the trading price of intermediate products. Therefore, compared to internal production, companies are more willing to purchase intermediate products, China's manufacturing sectors can improve Global Value Chain participation by promoting green upgrade mainly through green technology progress. Some scholars also argue that the Global Value Chain is fundamentally driven by the multiple cross-border circulation of intermediate goods [15], which leads to a continuous expansion of the transmission, superposition, and amplification effects of trade costs, thereby affecting corporate earnings.

Second, research on the factors influencing international trade. Factors such as the level of economic development, comparative advantage, trade costs, infrastructure development, institutional quality, and distance all affect agricultural trade. Agricultural trade volume is positively correlated with economic level, institutional quality (such as legal structure index, property rights protection, and trade liberalization), and infrastructure [[16], [17], [18]], while negatively correlated with trade costs and distance [19]. Well-developed infrastructure can promote division of labor by reducing transportation costs, thus expanding import and export trade. Grossman et al. pointed out that, larger market spaces help firms obtain necessary intermediate products, further reducing transaction costs, and infrastructure improvements enhance a country's position in the value chain [20]. Behrens [21] and Shepherd [22] found that, the higher the quality of a country's infrastructure, the larger its foreign trade volume, and the smaller the development gap between regional economies. Other studies have found infrastructure, particularly transportation, broadband, and information communication systems, has a significant impact on product competitiveness and economic growth [17]. Stefan Bojnec and Imre Fertő [23] constructed a gravity model to empirically test the impact of the Internet on food industry trade in OECD countries, finding that the Internet can enhance bilateral trade levels by reducing information asymmetry and related trade costs, helping to mitigate the negative impact of distance on international food industry trade. Additionally, the development of information and communication infrastructure, promotes bilateral agricultural trade flows among OECD countries and helps reduce transaction costs in trade exchanges [24].

Through the literature review above, it has been found that, despite existing research confirming the economic effects of infrastructure on trade growth from various perspectives, there are still three main shortcomings. First, most of the existing studies on the CR Express focuses on manufacturing and enterprises, with few studies examining the impact of transportation infrastructure development on agricultural trade. Second, while many scholars have explored the influence of infrastructure on bilateral trade and economic growth, these studies mainly analyze trade volume [23] and export unit value [16], with few focus on value chain linkages. Additionally, existing research on value chain linkages mostly concentrates on value and positional dimensions, with relatively few studies involving the agricultural sector. Third, there is an inherent interaction effect between transportation infrastructure construction and trade growth. The CR Express is essentially both transportation infrastructure and a substantial trade carrier under “the Belt and Road” Initiative, making it challenging to identify its policy effects.

Based on these limitations and considering that agricultural products are perishable, prone to spoilage, and have low unit value, the construction of agricultural value chains depends more on improvements in transportation, tariffs, and customs efficiency [2]. This paper treats the launch of the CR Express as a quasi-natural experiment, measures the linkages between Chinese and foreign agricultural value chains from a mobility dimension, constructs a difference-in-difference model to identify the agricultural value chain linkage effects of the railway, and further explores the underlying mechanisms. The goal is to provide decision-making references to stimulate the market vitality and creativity of the CR Express.

3. Theoretic analysis

Transportation and logistics have always been critical components of global agricultural trade, and the formation of agricultural value chains increasingly relies on improved external conditions, such as transportation and customs efficiency. Infrastructure development can effectively reduce time delays in the supply chain, allowing agricultural products to enter international markets more quickly, and enhancing their position in the global agricultural value chain. The operation of the CR Express has reshaped the economic geography between China and Europe by addressing factors of “distance, density, and segmentation” [9], laying the groundwork for opening up China's inland and border regions. On one hand, the operation of the railway significantly enhances the trade potential between China and the countries along its route, fostering export growth and further stimulating enterprise innovation activities [8]. On the other hand, the CR Express shortens the time-space distance, reduces transportation costs at infrastructure nodes, enhances the comparative advantages of node cities, and strengthens the potential for international trade specialization [25]. Furthermore, the “point-to-point” export model reduces information asymmetry in export markets, enhancing trade facilitation, which in turn supports the export of high-complexity products [7]. Clearly, the CR Express deepens international specialization, lowers market entry barriers, and facilitates easier integration of countries along its route into the global trade and production network, thus improving inter-country linkages in agricultural value chains. Based on this, we propose the following hypothesis:

H1

The operation of the CR Express will significantly enhance the linkage between bilateral agricultural value chains.

New economic geography theory suggests that, there are pervasive “iceberg costs” in product trade, factors like transportation time and distance significantly impact transaction costs. Transaction costs, in turn, are crucial determinants of a firm's production boundaries [26]. With the increasing segmentation of intermediate product markets, the production process has progressively achieved functional and spatial separation, which influences the length of the industrial chain [19]. On one hand, the operation of the railway improves cross-border supply chain efficiency, enabling closer collaboration among upstream and downstream firms, creating more value across each link in the supply chain. On the other hand, the CR Express promotes regional specialization and collaboration, thereby enhancing the linkage between bilateral value chains. This occurs because changes in industrial chain length can reduce transaction costs for intermediate products, encouraging firms to purchase these products rather than produce themselves. As companies demand more intermediate products, market specialization and product segmentation deepen further, which creates a positive cycle. Additionally, the railway's operation provides opportunities for vertical integration and upgrading within the industrial chain, facilitating regional technology transfer and innovation cooperation, also significantly enhancing the overall efficiency of bilateral value chains. Therefore, increasing the segmentation level of intermediate product markets not only enables each trading entity to leverage its factor endowments more fully, extending the length of the industrial chain, but also promotes bilateral trade and strengthens value chain linkages [14]. Based on this theoretical and empirical foundation, we propose the following hypothesis:

H2

The extension of the industrial chain length is a key mechanism by which the CR Express enhances the linkage between bilateral agricultural value chains. Specifically, the railway's operation increases bilateral value chain linkages by extending the length of the industrial chain.

4. Model construction and data sources

4.1. Econometrics model setting

Currently, as a new organizational form of international railway transportation, the CR Express is increasingly becoming the most closely connected transportation corridor between China and the countries along “the Belt and Road”. This paper constructs a difference-in-differences model, treating the operation of the CR Express as a natural experiment. This is because the decision to operate the railway is usually driven by government initiatives, and influenced by macro policies, such as “the Belt and Road” Initiative, trade demands, and international relations, rather than by local economic conditions or specific industry needs. Therefore, it is reasonable to view the operation of the CR Express as exogenous to the economic development conditions of specific regions, and treat it as a natural experiment.

Due to the ability to exclude the influence of external factors and ease of use, the difference-in-differences method is widely applied in econometric evaluations of public policy implementation effects [27]. Thus, this paper considers countries that have operated the CR Express as the treatment group, and those that have not as the control group, constructing the following difference-in-differences model:

$$\mathit{b}v{c}_{it}=\alpha +\beta \ast CRexpres{s}_i\ast tim{e}_t+X_i\ast \gamma +X_{it}\ast \phi +{\nu }_i+{\nu }_t+{\epsilon }_{it}$$ (1)

In Eq. (1), bvc_it represents bilateral agricultural value chain connectivity (i denotes the country, t denotes the year, the same applies throughout), including inflow and outflow; CRexpress_i represents the treatment group dummy variable, if a country has operated the CR Express, it is assigned a value of 1, otherwise, it is 0; time_t represents a dummy variable used to indicate the period of the treatment effect. Since the first CR Express started operating in 2011, so we consider 2011 as the year when the policy began to have an impact. If it is in or after 2011, it takes a value of 1; otherwise, it is 0; CRexpress_i ∗ time_t represents the interaction, used to estimate the agricultural value chain linkage effects, resulting from the operation of the CR Express; β is the coefficient of primary interest in this paper, used to measure the impact of the CR Express's operation on the domestic and foreign agricultural value chain connectivity; X_i, X_it represent the country characteristics that do not change over time and those change with time; v_i, v_t represent country and year fixed effects, and ɛ_it represents the random disturbance.

4.2. Data sources

The research data in this paper are derived from the United Nations Conference on Trade and Development's multi-region input-output table (MRIO), which encompasses 190 countries across 26 economic sectors. Given the accuracy and availability of the data, this study employs the Eora26-2015 MRIO to calculate bilateral agricultural value chain connectivity, with a sample size consisting of data from 85 countries and regions spanning from 1995 to 2015. Data on the operation of the CR Express is sourced from the CR Express Development Plan (2016–2020). The language similarity index and distance between the two countries' capitals are derived from the CEPII database, which includes global data on geographical, economic, and cultural distances. Data on the trade countries' economic scale, GDP growth rate, population size, agricultural land area, and the proportion of the agricultural population come from the World Bank database. Data on the partner country's government effectiveness index, regulatory quality index, control of corruption index, and rule of law index are sourced from the Worldwide Governance Indicators database, which reports views on governance quality based on surveys of businesses, citizens, and experts worldwide. The indicator for civil aviation and maritime agreements between the two countries is sourced from the treaty documents database on the Chinese Ministry of Foreign Affairs website.

4.3. Dependent variable: bilateral agricultural value chain connectivity

When testing hypotheses, the key issue is how to measure the dependent variable: the linkage of bilateral agricultural value chains. Following the decomposition method of Wang et al. [28,29]、Muradov [7,8], this paper constructs a cross-border frequency indicator to measure the movement characteristics of value-added by various industries in each country's value chains. Additionally, based on the calculation method of Qin & Zhu [13], we decompose the agricultural value-added e_ci resulting from China's exports to country i. The specific calculation formulas are as follows:

$$e_{ci}={\left[E_{ci}\right]}_{c_{-}a\mathrm{g}ri}\times u_{G\times N}=B_{ci}\times \left({\left[V{A}_{ci}\right]}_{c_{-}a\mathrm{g}ri}\times u_{G\times N}\right)$$ (2)

$$E_{ci}=\hat{V}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\times A_{ci}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\times {\left\{{\left[I-A_{ci}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\right]}^{-1}\right\}}^2\times \hat{Y}+\hat{V}\times L\times {\hat{R}}_{ci}$$ (3)

$$V{A}_{ci}=\hat{V}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\times A_{ci}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\times {\left[I-A_{ci}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\right]}^{-1}\times \hat{Y}+\hat{V}\times {\left(I-{\bar{A}}_{ci}\right)}^{-1}\times {\hat{R}}_{ci}$$ (4)

In Eq. (2), G and N represent the number of countries and industry sectors, respectively. E_ci and VA_ci are both (G × N) × (G × N) matrices. E_ci is the matrix for decomposing China's exports to country i, where any element represents the value of exports from China to country i resulting from the flow of value-added from an industry sector in the row country to an industry sector in the column country. VA_ci represents the decomposition matrix of value-added corresponding to China's exports to country i, where any element represents value-added produced by an industry sector in the row country, appearing in China's export process to country i and flowing to the final product in the industry sector of the column country. Eq. (3) and Eq. (4) respectively illustrate the specific forms of E_ci and VA_ci. B_ci represents the cross-border frequency of China's agricultural export value-added, a higher value indicates a greater degree of production fragmentation, and with more transactions between countries, thereby enhancing bilateral agricultural value chain connectivity. []_{c_agri} represents selecting the row corresponding to China's agricultural sector in the matrix. $u_{G\times N}$ is a GN × 1 column vector. $\hat{V}$ is a diagonal matrix, where diagonal elements representing the value-added ratios for each country and sector. I is the identity matrix. A is the direct consumption coefficient matrix (G × N rows, G × N columns), where $A=A_{ci}+{\hat{A}}_{ci}$, A_ci represents the direct consumption matrix retains the row corresponding to China and the column corresponding to country i, with all other elements set to zero. $\hat{Y}$ is a diagonal matrix constructed from the final demand vector. L represents the Leontief inverse matrix. ${\hat{R}}_{ci}$ is The diagonal matrix derived from the final demand vector of country i, keeping the corresponding row for China and setting all other elements to zero.

$$\mathit{b}v{c}_{it}=\left(B_{ci}-1\right)\times {10}^7$$ (5)

Furthermore, in order to obtain more economically meaningful indicators, we draw on the calculation method of Qin & Zhu [13], improves the B_cit indicator and constructs a new value chain connectivity indicator. In Eq. (5), bvc_it means: in bilateral trade, the additional value created by China for every 10 million dollars of agricultural exports to a specific country. Similarly, by decomposing China's agricultural import values from a specific country, one can obtain the corresponding additional value derived from foreign imports.

4.4. The variables related to the operation of the CR express

This paper identifies the micro effects of the opening of the CR Express by constructing a difference-in-differences model. The specific variables used are as follows. According to the “Development Plan for the Construction of the CR Express (2016–2020)," we select 26 countries that have either operated or planned to operate the railway to identity whether the country is along the CR Express. Also, based on official documents, we found 2011 was determined as the year the policy started to have an impact, so we consider the year 2011 as the point in time when the policy took effect. Control variables consist of country characteristics that either change over time or remain constant, variables that do not change over time are the Language Similarity Index and the distance between the capital cities of the two countries. Variables that change over time include the economic size of the two countries, GDP growth rate of trading country, population size and agricultural land size of the two countries, the proportion of agricultural population in the trading country, government effectiveness index of the partner, regulatory level index, corruption control index, and rule of law index. Additionally, whether the two countries have signed civil aviation agreements and maritime agreements are also considered. The descriptive statistics of each variable are shown in Table 1.

Table 1.

Definition of variables and descriptive statistics.

Variables	Description	Mean	S.E	Min	Max
Inflow	Agricultural value chain linkages inflow (dollars)	218.800	631.700	0.0326	5400.000
Outflow	Agricultural value chain linkages outflow (dollars)	203.900	724.400	0.0182	7835.000
CRexpress	Whether it is a country along the “CR Express": Yes = 1; No = 0	0.259	0.438	0.000	1.000
time	Whether after 2011 (including 2011): Yes = 1; No = 0	0.238	0.426	0.000	1.000
lngdp	The logarithm of the product of the domestic and foreign GDPs	54.130	2.153	48.060	60.260
GDPgrow	Trade partner country GDP growth rate (%)	4.119	5.205	−33.100	88.960
lnpopu	The logarithm of the product of domestic and foreign population quantities	9.578	1.703	5.724	14.400
lnagriarea	The logarithm of the product of domestic and foreign agricultural land areas	26.120	2.510	17.340	30.820
ratio	The proportion of the agricultural population in the partner country	21.190	20.230	0.190	85.410
neighbor	Neighbor of China: Yes = 1; No = 0	0.153	0.360	0.000	1.000
distance	Straight-line distance between Beijing, China and partner capitals (km)	6364.409	2401.861	1172.047	17614.300
language	Language similarity index: [0,1]	0.064	0.051	0.000	0.475
gov	Government effectiveness index of the partner country: [-2.5,2.5]	0.294	1.008	−2.232	2.437
reg	Regulatory level index of the partner country: [-2.5,2.5]	0.271	1.017	−2.344	2.261
cor	Corruption control index of the partner country: [-2.5,2.5]	0.164	1.094	−1.673	2.470
law	Rule of law index of the partner country: [-2.5,2.5]	0.200	1.035	−1.991	2.100
air	Whether China has signed a shipping agreement with the partner country: Yes = 1; No = 0.	0.752	0.432	0.000	1.000
sea	Whether China has signed a maritime agreement with the partner country: Yes = 1; No = 0.	0.513	0.500	0.000	1.000

5. Empirical results

5.1. Parallel trend test

The parallel trends assumption is a crucial prerequisite for the validity of the difference-in-differences model. An important condition for using the difference-in-differences method is that, prior to the policy intervention, the trends of the treatment and control groups should be roughly the same. To further accurately assess the impact of the operation of the CR Express on the bilateral agricultural value chain linkages, this paper follows the approach of Liu & Qiu [30], and conduct a further analysis of the trend changes in the treatment and control groups. The regression equation constructed is as follows:

$$\mathit{b}v{c}_{it}=\alpha +{\beta }_k\sum _{k⩾-4}^{4}CRexpres{s}_i\times yea{r}_{2011+k}+X_i\gamma +X_{it}\phi +v_i+v_t+{\epsilon }_{it}$$ (6)

In Eq. (6), year_2011+k is a year dummy variable, which includes a time span of 4 years before and after the operation of the CR Express. For example, if k = −3, this value indicates whether the year is a dummy variable for 2008. The meanings of the other variables are consistent with the baseline model. This section examines the trend changes over a total of 9 years before and after the operation of the CR Express, focusing on whether there are differences between the treatment group and the control group before the policy was implemented (before 2011). The results in Fig. 1 indicate that, whether from the perspective of value-added inflow or outflow, none of the regression results before 2011 are significant. This indicates that before the launch of the railway, the trend changes in the treatment and control groups were consistent, with no significant differences. The parallel trends test is passed, allowing the use of the difference-in-differences method to assess the agricultural value chain linkage effects of the CR Express.

Fig. 1.

Parallel trend test.

5.2. Baseline regression results

Under the assumption that parallel trends are met between the treatment group and the control group, this paper evaluates the impact of the operation of the CR Express on bilateral agricultural value chain linkages, using the constructed difference-in-differences model. The estimation results are presented in Table 2.2 The first two columns and the last two columns of the table report the results of the impact of the CR Express on agricultural value chain linkages, both inflow and outflow. The empirical results show that, the operation of the CR Express significantly enhances bilateral agricultural value chain linkages. From both the perspective of value-added inflow and outflow, the operation of the CR Express positively promotes agricultural value chain linkages between China and the treatment group, with the inflow effect consistently greater than the outflow effect. Columns (2) and (4) present results after adding control variables, the significance and direction of the coefficients have not changed, indicating that the results of baseline regression have certain robustness.

Table 2.

Baseline regression results.

Variables	(1)Inflow	(2)Inflow	(3)Inflow	(4)Outflow	(5)Outflow	(6)Outflow
CRexpress × time	87.167∗∗∗	100.014∗∗∗	65.570∗∗∗	29.157∗∗∗	42.686∗∗∗	61.831∗∗∗
	(10.299)	(19.804)	(18.565)	(5.636)	(13.900)	(15.435)
Controls	No	Yes	Yes	No	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes	Yes
N	1785	1713	1713	1785	1713	1713
R2	0.777	0.821	0.837	0.776	0.808	0.836

Note: Standard errors in parentheses; ∗∗∗, ∗∗, and ∗ indicate significance levels of 1 %, 5 %, and 10 %, respectively, throughout.

Specifically, after the operation of the CR Express, for every additional $10 million in agricultural value added imported by China, there will be an additional import value of $100.014. Likewise, for every additional $10 million in agricultural value added exported by China, there will be an additional export value of $42.686. This is because the operation of the CR Express makes regional transportation more convenient, shifting the trade of agricultural products towards intermediate goods trade, which not only increases the degree of specialization within the region, but also further reduces trade costs. As trading nations benefit from the deepening of specialization, they are more willing to engage in intermediate goods trade, which further promotes and strengthens the trade ties among trading nations. Additionally, due to the Sino-US trade friction, China continuous expands agricultural product import destinations, further deepening cooperation with countries along “the Belt and Road” initiative. Moreover, in agricultural product trade with countries along the route, China primarily acts as a net receiver, meaning that compared to outward spillovers, the Chinese agricultural market is more receptive to price spillovers from agricultural markets, which further validates that the value-added inflow effect is greater than the outflow effect. To reduce the influence of outliers, this paper also employs a 1 % two-sided trimming method on the panel data for testing. The estimation results can be found in columns (3) and (6), where it is observed that the corresponding regression coefficients remain significantly positive.

5.3. Mechanism analysis

The length of the industrial chain refers to the average of production stages that a sector's product goes through from initial input to final consumption, depicting the movement distance of a country's sectoral value-added in the Global Value Chain [31]. To obtain the agricultural industrial chain length, this paper utilizes Eora26 data and follows the calculation method proposed by Wang et al. [32] to calculate the agricultural industrial chain length of the sample countries. The calculation formula is as follows:

$$\mathit{P}{L}_i={\left[\hat{V}BBY/\hat{V}BY\right]}_{i_{-}agri}$$ (7)

In Eq. (7), PL_i represents the length of the industrial chain in which the agricultural sector of country i is situated. Y represents the column vector of final demand. ${\left[\right]}_{i_{-}agri}$ denotes the elements corresponding to the agricultural sector of country i in the column vector. The meanings of other symbols are as mentioned above.

To further examine the effectiveness of the mechanism through which the operation of the CR Express affects the agricultural industrial chain linkage, this paper follows the approach of Mao [33], and constructs the model of Eq. (8) and Eq. (9) for testing:

$$\mathit{P}{L}_{it}=\alpha +\beta CRexpres{s}_i\times tim{e}_t+X_{\mathrm{i}}\gamma +X_{\text{it}}\phi +{\nu }_i+{\nu }_t+{\epsilon }_{it}$$ (8)

$$\mathit{A}{\text{vcc}}_{it}={\rho }_0+{\rho }_{1}CRexpres{s}_i\times tim{e}_t+{\rho }_{2}P{L}_{it}+X_{\mathrm{i}}\gamma +X_{\text{it}}\phi +{\nu }_i+{\nu }_t+{\epsilon }_{it}$$ (9)

This paper analyses the mechanism effect of the CR Express using a stepwise regression approach, and the results are shown in Table 3. Columns (1) to (3) report the impact of the CR Express on the agricultural industrial chain length, the effect of agricultural industrial chain length on value-added inflow, and the effect on value-added outflow respectively. It can be observed that the operation of the CR Express significantly promotes the length of the agricultural industrial chain. From the perspective of value inflow, the agricultural industrial chain length significantly enhances bilateral agricultural value chain linkages, suggesting that in China's agricultural import trade, the operation of the CR Express can promote the linkage of agricultural value chains between China and its trading partners by extending the length of the agricultural industrial chain.

Table 3.

Mediating effect.

Variables	(1) PL	(2) Inflow	(3) Outflow
PL		9.942∗	4.172
		(5.529)	(4.997)
CRexpress × time	0.090∗∗	99.115∗∗∗	42.309∗∗∗
	(0.035)	(19.812)	(13.860)
Controls	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Country FE	Yes	Yes	Yes
N	1713	1713	1713
R2	0.544	0.821	0.808

However, from the perspective of outflow, the effect of agricultural industry chain length on bilateral agricultural value chain linkages is positive but not significant. This suggests that, from the perspective of value-added outflow, extending the agricultural industry chain is not the key mechanism through which the railway promotes agricultural value chain linkages. A possible reason is the decisive role of transaction costs in international trade. When a product's price difference in the international market and domestic price is greater than transaction costs, traders will increase imports. The operation of the CR Express extends the length of the agricultural industrial chain, leading to a finer division of labor in international agricultural product trade. The reduction in transportation costs gradually lowers transaction costs, so that countries along the route can enjoy greater benefits. Additionally, countries with significant agricultural advantages are more willing to expand their agricultural exports, further changing the international trade patterns and leading to noticeable changes in agricultural value chain linkages between countries. Furthermore, given China's trade deficit position, agricultural value chain linkages primarily focus on inflows.

5.4. Placebo test

5.4.1. Placebo test 1: assume that the first train of the CR express was launched earlier than 2011

Based on the approach of Topalova [34], this paper sets the time before 2011 for the first run of the CR Express, and examines whether the enhancement effect of simulated events on the agricultural value chain linkages still holds. Consistent with the earlier analysis, the assumption of the difference-in-differences method is that there are no significant differences between the treatment group and the control group before the policy change. Therefore, if the event is set to occur before 2011, the estimated coefficients of the core variables will no longer be significant. If the estimated results are contrary to expectations, meaning that the simulated event has a promoting effect on agricultural value chain linkages, suggesting that there are indeed some unobservable factors that can drive the linkages between China and the countries along the CR Express. In this paper, an event year is selected every three years between 1996 and 2011, assuming in 1996, 1999, 2002, 2005, and 2008.

Columns (1) to (5) and columns (6) to (10) in Table 4, respectively present the empirical results from the perspectives of agricultural value chain linkage inflows and outflows. According to the results in Table 4, the estimated coefficients of the interaction terms are not statistically significant, thereby excluding the impact of unobservable factors on the promotion effect of bilateral agricultural value chain linkages. In addition, to reduce the influence of outliers, this paper also employs a 1 % two-sided trimming method on the panel data for testing, and the estimation results are consistent with the original model (see Appendix Table 1).

Table 4.

The results of placebo test.

Variables	(1)	(2)	(3)	(4)	(5)
CRexpress × time	4.876	8.504	30.187	60.601	52.288
	(71.970)	(41.101)	(36.327)	(34.461)	(34.595)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes
N	1713	1713	1713	1713	1713
R2	0.181	0.181	0.181	0.182	0.182
Variable	(6)	(7)	(8)	(9)	(10)
CRexpress × time	−0.123	2.288	29.498	66.077	61.850
	(85.675)	(114.266)	(142.238)	(164.261)	(152.905)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes
N	1713	1713	1713	1713	1713
R2	0.158	0.158	0.158	0.159	0.159

5.4.2. Placebo test 2: randomly select the treatment group

Omitted variables are another potential cause of estimation bias in regression results. Following the approach of Cai et al. [35], this paper uses the Bootstrap method to randomly sample the treatment group and control group, and repeat the above process 500 times. Since the treatment group here is randomly generated, the estimated coefficients of the key variables obtained from the baseline regression using this sample should be distributed around zero. If not, it can be considered that the baseline model is subject to identification bias. Fig. 2 reports the kernel density distribution of the estimated coefficients of the key variables for 500 random samples. As shown in Fig. 2, both value chain inflows Fig. 2(a) and value chain outflows Fig. 2 (b), the estimated coefficients are distributed around zero, indicating that the baseline regression results are not seriously biased due to omitted variables.

Fig. 2.

Estimated coefficients from the random assignment of the treatment group.

5.5. Heterogeneity analysis

To assess whether the operation of the CR Express affects the bilateral agricultural value chain linkages differently in regions, this paper divided the samples into countries with land borders with China (neighboring countries) and countries without land borders with China (non-neighboring countries). Subsequently, we conducted group regression analysis, and the results are presented in Table 5.

Table 5.

The results of heterogeneity analysis.

Variables	Inflow		Outflow
	Neighbor	Non-neighbor	Neighbor	Non-neighbor
CRexpress × time	−40.293∗∗	153.105∗∗∗	−4.644	79.945∗∗∗
	(14.610)	(29.194)	(7.575)	(20.554)
Controls	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
N	265	1448	265	1448
R2	0.737	0.830	0.776	0.813

It can be observed that the operation of the CR Express has a significantly positive impact on the bilateral agricultural value chain linkages in non-neighboring countries. In neighboring countries, the operation of the CR Express positively enhances the agricultural value chain inflow at a 1 % level of significance. However, its impact on the outflow is not significant, and it is consistently negative, which indicates that the effect of the operation of the CR Express on enhancing agricultural value chain linkages exhibits regional heterogeneity. The possible reason is that the operation of the CR Express connects international trade between China and developed economies in Europe. Agricultural trade between China and agricultural powerhouses like the Netherlands, Germany, France, and the United Kingdom has become increasingly active, significantly reducing the international trade costs associated with agricultural products. Existing data demonstrates that, in 2020, China became the largest trading partner of the European Union for the first time.

Under the backdrop of trade liberalization, China-Europe trade has shown diversification trends, with complementary trade gradually replacing competitive trade. China mainly exports labor-intensive products, while the European Union primarily exports highly processed, high-value-added products with a focus on capital and technology. Therefore, the complementarity in bilateral agricultural product trade is much greater than competitiveness. Simultaneously, with the implementation of the “China-EU agreement on geographical indications”, trade barriers for agricultural products between China and European countries have been gradually reduced. The growing vitality of agricultural product trade further promotes the friendly and continuous development of trade relations.

Compared with China's neighbors, European countries have more comparative advantages in agricultural trade. Their dynamic foreign investment and advanced technology not only reduce production costs, but also have more advantages in trade price competition. Compared to transportation costs, China enjoys higher price benefits in trade with European countries, making it more inclined to trade with European countries and reducing agricultural product trade with neighboring countries, especially in the case of highly processed, high-value-added agricultural products. Furthermore, transportation infrastructure plays a crucial role in the cost of intercity transportation. The CR Express significantly reduces logistics costs and shortens the transportation time for agricultural products, laying the foundation for agricultural trade between China and developed European economies. The results also indicate that while China's agricultural trade both in exports and imports is generally on an upward trend, the growth rate of exports is slower than that of imports. In the period studied in this paper, it specifically demonstrates that the impact of the operation of the CR Express on agricultural value chain linkages in China's neighbors is relatively small, with the inflow effect consistently outweighing the outflow effect.

Also, in order to reduce the influence of outliers, this paper also employs a 1 % two-sided trimming method on the panel data for testing, and the estimation results are consistent with the original model (see Appendix Table 2).

6. Conclusions and implications

6.1. Conclusion

With the ongoing advancement of “the Belt and Road” Initiative, the CR Express has become an important link for the central and western regions to integrate into the global economic system. This paper views the operation of the CR Express as a quasi-natural experiment, utilizing cross-regional input-output data from 1995 to 2015 and employing a difference-in-difference model to explore the enhancement effect of the CR Express on bilateral agricultural value chain connections. The research findings indicate that:

First, the CR Express significantly enhances the linkage between domestic and international agricultural value chains, and its role in promoting value-added inflows consistently exceeds that of value-added outflows.

Second, from the perspective of value-added inflows, the CR Express enhances agricultural value chain linkages by extending the length of agricultural industrial chains; from the perspective of value-added outflows, the extension of agricultural industrial chains is not the main mechanism through which the railway promotes agricultural value chain linkages.

Third, the enhancing effect of the CR Express on agricultural value chain linkages exhibits heterogeneity, specifically, for non-neighboring countries, the enhancement effect on agricultural value chain linkages is significantly positive, while for neighboring countries, the corresponding effect is negative, and this negative effect is only significant from the perspective of value-added inflows.

The research conclusions of this paper reveal the significant role of the CR Express as a transportation infrastructure in promoting agricultural value chain connections, especially its remarkable enhancement effect on non-neighboring countries, providing empirical evidence for the formulation of multinational transportation policies under “the Belt and Road” Initiative. Compared to existing research, this paper not only expands the discussion on the relationship between infrastructure and trade growth, particularly in the agricultural sector, providing new empirical support, but also elucidates the specific mechanisms through which the CR Express promotes bilateral agricultural cooperation by extending the agricultural value chain and enhancing value chain connectivity from the perspective of value chain relationships, thereby offering a new theoretical framework for future research. This finding not only enriches the understanding of the policy effects of the CR Express, but also deepens the study of the role of infrastructure in international agricultural cooperation. The research conclusions provide strong policy implications for optimizing land transport networks, enhancing the international competitiveness of agricultural value chains, and promoting the deep integration of the central and western regions into the global economic system, holding significant practical guidance.

6.2. Policy implications

First, given the significant promoting effect of the CR Express on the agricultural value chain connections between China and other countries, particularly with the inflow effect of value-added being greater than the outflow, related policies should prioritize supporting the extension of agricultural industrial chains in the central and western regions, especially in the deep processing sector. China should leverage “the Belt and Road” Initiative to engage deeply in regional economic integration, increasing investment in land transportation infrastructure in the central and western regions to improve logistics efficiency. The government can encourage the development of deep processing enterprises for agricultural products in these regions through funding support, technical training, and policy incentives, enhancing the added value of agricultural products, promoting the “going out” of agricultural products and the “bringing in” of high-value agricultural products, thereby enhancing the competitiveness of Chinese agriculture in the global agricultural value chain.

Second, research shows that extending the agricultural industrial chain can effectively enhance the connectivity of the agricultural value chain. Therefore, China should accelerate the adjustment and optimization of its agricultural industrial structure, deepen agricultural product development, and increase the added value of agricultural products. Currently, the short agricultural industrial chain has become a significant barrier to the enhancement of China's value chain. To address this issue, emphasis should be placed on supporting the construction of multinational agricultural industrial parks, developing deep processing of agricultural products, leveraging the demonstration effect of industrial parks, and enhancing the organization and marketization levels of domestic agriculture. At the same time, the government should promote innovative models in the agricultural supply chain, exploring diversified supply chain combinations to enhance the resilience and overall international competitiveness of the agricultural value chain.

Third, given the significant positive effect of the CR Express on the agricultural value chains of non-neighboring countries, policymakers should increase investment in related land transportation infrastructure in these countries and further optimize the logistics network and operational model of the CR Express. In particular, a reasonable layout of train stops should be established, and freight organization modes should be adjusted to improve transportation efficiency. Through these measures, deeper agricultural trade cooperation between China and non-neighboring countries can be promoted, expanding the coverage of Chinese agricultural products in international markets.

Although this paper systematically analyzes the promoting effect of the CR Express on agricultural value chain connections and achieves certain results, some limitations still exist. The paper uses cross-regional input-output data from 1995 to 2015, which covers important time points in the early stages of the CR Express's operation. However, with the ongoing advancement of the Belt and Road Initiative, the current data may not fully capture the latest trade dynamics. Therefore, future research should incorporate updated multinational trade data to analyze the impact of the CR Express on agricultural value chains in subsequent periods.

Funding statement

This work was supported by the China Association for Science and Technology under Grant [CASTBR201614], the Ministry of Science and Technology of the People's Republic of China under Grant [DL2022172001L].

CRediT authorship contribution statement

Yaqiu Su: Writing – review & editing, Writing – original draft, Software, Methodology, Conceptualization. Yuchun Zhu: Writing – review & editing, Writing – original draft, Project administration, Conceptualization. Guoqing Qin: Writing – review & editing, Writing – original draft, Methodology, Data curation, Conceptualization.

Ethical approval

This paper does not contain any studies with human participants performed by any of the authors.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Table A1.

The results of placebo test

Variables	(1)	(2)	(3)	(4)	(5)
CRexpress × time	1.350	2.423	16.917	38.795	22.860
	(87.306)	(107.839)	(133.793)	(144.587)	(124.223)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes
N	1713	1713	1713	1713	1713
R2	0.198	0.198	0.198	0.199	0.198
Variable	(6)	(7)	(8)	(9)	(10)
CRexpress × time	20.540	27.226	55.366	87.799	79.185
	(73.146)	(97.844)	(125.357)	(147.538)	(134.562)
Controls	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes
N	1713	1713	1713	1713	1713
R2	0.210	0.210	0.211	0.214	0.213

Note: Standard errors in parentheses; ∗∗∗, ∗∗, and ∗ indicate significance levels of 1 %, 5 %, and 10 %, respectively.

Table A2.

The results of heterogeneity analysis

Variable	Inflow		Outflow
	Neighbor	Non-neighbor	Neighbor	Non-neighbor
CRexpress × time	−40.291∗∗	110.112∗∗∗	−4.644	96.600∗∗∗
	(14.610)	(27.005)	(7.575)	(21.291)
Controls	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
N	265	1448	265	1448
R2	0.737	0.848	0.775	0.843

Note: Standard errors in parentheses; ∗∗∗, ∗∗, and ∗ indicate significance levels of 1 %, 5 %, and 10 %, respectively.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.