The impact of China railway express on foreign direct investment inflows in Chinese central and western cities

Abstract

Capital needs transportation channels. Following successful communication and cooperation with Central Asian and European states, the China railway express (CRE) has been built by central and western cities. This new international freight service will greatly enhance the transport conditions in the central and western cities. Therefore, it is necessary to conduct an in-depth analysis of the impact of CRE operations on foreign direct investment (FDI) flows in the central and western regions. For analyses, a panel data of 152 Chinese cities from 2008 to 2020 is used and staggered difference in differences (DID) model is applied as a quasi-natural experiment. The results demonstrated that the operation of the CRE had a positive and significant impact on FDI inflows in the central and western cities, particularly in western cities, large cities, and non-resource based cities in China. Mechanism analysis shows that CRE operations can enhance the ability to attract foreign investment in the central and western regions through the promotion of industrial agglomeration and the expansion of market size. Therefore, the government should actively optimize the layout of CRE transportation routes and establish an inter-regional coordinating mechanism for freight sources, thus allowing the radiating effect of CRE central cities to reach out to peripheral cities.

1. Introduction

Since the reform and opening up, foreign direct investment (FDI) has played a crucial role in China's socio-economic development. It has not only provided sufficient funds for initial economic construction, but also promoted China's faster integration into the global economic division of labor. In 2021, China ranks second with 11.4 % of global FDI inflows. However, China's FDI expansion has also led to uneven regional distribution. Objectively, the country's large geographical size has led to regional disparities between the eastern, central and western regions, due to differences in geographical location and natural resource endowments. In terms of policy, as part of its efforts to increase openness and actively participate in economic globalization, China has implemented a comprehensive opening-up strategy along its coasts, rivers and borders. The development benefits of policy preferences have made the eastern coastal region, with its advantage of international ports, more attractive to FDI than the central and western regions. Over the past two decades, the geographical constraints in the central and western regions have been greatly eased with the development of better transportation infrastructure. In order to further open up the inland areas in the western region, China has promoted the construction of the Belt and Road Initiative to strengthen economic and trade cooperation with countries in Central Asia, West Asia, Central and Eastern Europe and Western Europe. In this context, the China railway express (CRE) was born. Major cities in the central and western cities (e.g. Chongqing, Wuhan, Changsha, Chengdu, etc.) have taken the lead in developing a new transnational rail freight cooperation mechanism with Central Asian and European countries. Based on this, this study attempts to analyze the impact of improving the level of transportation infrastructure on the cross-border flow of capital factors through a quasi-natural experiment of CRE operation. Considering that the cities opened by CRE are mainly located in the interior, 152 cities in the central and western regions are included as the research objects of this paper.

The theory of international production fragmentation suggests that the fragmentation of production is achieved by dividing the entire production process into different stages according to different types of factor intensity, and then allocating the different stages of production to the least costly locations [1]. In particular, low labor costs are considered to be a critical factor in attracting FDI to developing countries, although poor transportation and logistics infrastructure can increase the transportation costs of transnational corporations [2]. Therefore, improved transportation conditions not only maximize low labor costs, but also imply lower logistics costs [3] and shorter production-supply-marketing cycles [4]. Ozcan [5], utilizing provincial data from Turkey, indicates that provinces with higher air traffic volume and dense road networks tend to attract more FDI. Cheng [6] finds that the improvement of transportation infrastructure positively affects the Japanese direct investments in the western regions of China. Most scholars agree that FDI inflows can be attracted by improving transport and logistics infrastructure. As an important practical platform for the infrastructure connectivity of the Belt and Road Initiative, the CRE has been gradually normalizing its operation since the first international service “Chongqing-Xinjiang-Europe” in 2011 and the start of the first round trip in 2014. In 2016, it adopted a unified brand. Refer to Table 1 for specific details. Under the vigorous advocacy and support of the Chinese government, three transport corridors connecting the east, central, and west have been successfully developed. This has transformed the trade pattern, previously reliant solely on maritime transportation, effectively addressing the locational disadvantage of the central and western regions being distant from the ocean. According to statistics from the China National Railway Group Limited as of January 2022, the cumulative number of CRE exceeded 50,000, connecting 180 cities in 23 European countries. The proportion of goods trade through the CRE to the total China-Europe goods trade increased from 1 % in 2015 to 7 % in 2020. It is evident that the most significant impact of the CRE on the opening cities is the enhancement of international freight transportation capabilities.

Table 1.

Operation of major CRE in the central and western regions.

Opening time	Departure City	Train subsidies	Main cargoes	Return time of the liner (year)	Number of columns operating in 2020 (column)
2011	Chongqing	Subsidies to local enterprises and logistics platforms through land allocation, corporate income tax rebates and optimization of the logistics environment.	Laptops, machinery, auto parts, clothing.	2	2603
2012	Wuhan	The Foxconn special liner is subsidized in the name of enterprise tax rebate, and the public liner is subsidized according to the standard of box volume from USD 2500 to USD 3000 per container.	Electronics, fiber optic cables, etc.	3	215
2013	Chengdu	The total annual subsidy is about RMB 1.2 billion.	Electronics, auto parts, wine, etc.	2	2431
2013	Zhengzhou	Land and capital subsidies for liner operating companies.	Light textile, machinery, electronic products.	1	1126
2013	Xi'an	Subsidized at a rate of approximately $3000 per cabinet.	Machinery, auto parts, photovoltaic products, plastic products, household goods.	2	3720
2014	Changsha	Subsidized at a rate of approximately $3000 to $4000 per cabinet.	Electronic products, machinery, auto parts, textiles, ceramics.	1	528
2014	Hefei	Subsidized at a rate of about $3000 per container in the name of improving the logistics environment.	Machinery, industrial supplies, intelligent equipment.	2	568

Note: Based on the official website of the Belt and Road, the city's CRE subsidy documents and public information, and media reports.

In analyzing the impact of transportation and logistics infrastructure on cross-country capital flows, researchers have conducted empirical analyses of single and multiple countries. While many previous studies have analyzed at the country level, few studies have examined the impact at the city level. Based on the existing studies as important references, this paper attempts to address the multiple empirical challenges in identifying causal effects and to accurately identify the impact of transport and logistics infrastructure on FDI inflows. The opening of the CRE coincidentally provides an appropriate quasi-natural experiment for the study in this paper. Firstly, the nature of the CRE opening is distinct from transportation infrastructure construction. It is essentially a transnational freight cooperation mechanism formed through multilateral negotiations on existing railway facilities, a cooperative mechanism not independently determined by the opening cities but subject to policy standards set by the countries for the CRE. Secondly, the CRE operates on a “point-to-point” basis, meaning that cargo handling occurs only at the origin and destination cities, and theoretically has no impact on other cities along the route. These two characteristics make the opening of the CRE an appropriate exogenous shock to urban transport conditions, and provide a crucial foundation for the study's identification strategy.

The contributions of this study are as follows: First, in terms of research perspective, this study examines the impact of improved transport and logistics infrastructure development on FDI inflows based on a cross-border transport corridor perspective. Existing literature analyzes the improvement of domestic transportation networks in terms of improving domestic transportation networks, but neglects the impact of cross-border transportation corridors on FDI inflows. Therefore, this study not only enriches the discussion on transport logistics infrastructure and international capital flows, but also provides empirical insights into the unique characteristics of cross-border capital flows in developing countries. Second, in terms of research methodology, this paper mitigates the problem of endogeneity between transport logistics infrastructure and FDI in the previous literature by analyzing the causal relationship between CRE operations and FDI using a double difference model. Again, in terms of the path of influence, this study examines how CRE operation affects the path of FDI in the central and western regions, based on the idea that the development of transport infrastructure changes the spatial distribution of economic activities, taking into account the factors of industrial concentration and market size. Finally, the paper examines in more detail the spatiotemporal effects of opening the CRE. In terms of temporal considerations, the paper evaluates the impact of increased cumulative operating time and operating intensity of the CRE on FDI inflows. In terms of spatial considerations, this paper examines the spatial spillover effects of CRE operations and reasonably predicts the approximate extent of such effects, so as to further provide guidance for policymaking on CRE operations at the city level.

The rest of the paper is organized as follows. Sections 2, 3 present the literature review and theoretical hypotheses, respectively. Sections 4, 5 provide the econometric strategy and the empirical results, respectively. The last section concludes the paper.

2. Literature review

This study is based on two main categories of literature: the first examines the influence of transportation infrastructure on FDI, with a particular emphasis on research in the field of high-speed rail. The second category is concerned with the economic impact of the CRE, including an analysis of its effects on trade and factors of production.

2.1. Study on the impact of transportation infrastructure on foreign direct investment

Research on the relationship between transportation infrastructure and investment is of great importance for an in-depth investigation in this paper. In the existing literature, the analysis based on the social practice of high-speed railway construction and operation is extensively explored, covering aspects such as expanding market potential, improving economic operating efficiency, and promoting the coordination of regional development. Similar to high-speed railways, CR express serves as an important manifestation of infrastructure connectivity under the “Belt and Road” initiative, and functions as a land transport channel for China-Europe trade. Therefore, a thorough examination of relevant literature holds crucial significance for this paper. Franklin and Ahmed [7] were the first to incorporate infrastructure into considerations for FDI location selection, demonstrating that more sophisticated infrastructure is conducive to attracting foreign capital. Hilber and Voicu [8] found that the improvement of infrastructure levels redistributes economic activities such as trade and industry spatially, fostering economic growth and enhancing locational advantages. Vickerman [9] posits that transportation infrastructure leads to a convergence trend in spatial relationships between regions, improving regional accessibility. The study by Zhang et al. [10] indicates that high-speed rail promotes the aggregation of production factors such as capital and labour from peripheral cities to central cities, resulting in a suction effect that negatively impacts the productivity of peripheral enterprises. Focusing on the multidimensional perspective of distance costs, Chong et al. [11] elucidate the significant impact of domestic transportation systems on FDI. Duan et al. [12] discovered that high-speed rail construction can reduce information transmission costs, enhance investor growth expectations, and consequently facilitate capital flow and innovation. Zhang and Ou [13] demonstrates the positive impact of high-speed rail networks on the introduction of FDI into urban agglomerations, with a suction effect of capital from peripheral to central cities within a certain distance range after the opening of high-speed rail. This study indicates the unevenness of the investment attraction capability of railways between regions, with production factor resources flowing from the periphery to the central region. Similar conclusions are also found in the research by Wei and Sun [14], where the western regions exhibit significantly higher investment benefits than the eastern and central regions after the opening of high-speed rail, with larger cities benefiting more than smaller cities. In 2023, a study was conducted to update the research related to Infrastructure and Investment related issues. The study applied a new research methodology to assess the impact of productive capacity on FDI inflows [15]. The results showed that the transportation sector has a beneficial impact on the host country's ability to attract more FDI, as the development of domestic and international business increases the capacity to transport commercial products. Studies have shown that the development of transportation is an effective means of attracting FDI. Specifically, for every 1 % increase in transportation efficiency, FDI inflows increased by 23.76 % during the COVID-19 period. It is important to note that the content of the improved text must be as close as possible to the source text, and therefore no new aspects have been added. Ono and Sekiyama [16] verified the impact of official development assistance on FDI and found that transportation development in recipient countries is a potential factor in promoting interregional capital flows. In summary, numerous studies suggest that the development of transportation infrastructure becomes a potential factor in promoting the flow of capital between regions.

2.2. Economic effects of CR express

Due to the many differences in economic development, political systems, cultural customs, geographical climate and other aspects, the impact of the CRE as a cross-border freight carrier on the economic and social aspects of the regions along its route is particularly complex. Research in this area mainly revolves around two dimensions: promoting trade development and facilitating the free flow of factors. Most scholars believe that the opening of the CRE can foster trade development. According to Zhang et al. [17], the CRE has established a new land transport route between Asia and Europe. This new route is likely to increase trade, improve industrial development and factor mobility, and ultimately narrow the urban-rural income gap in cities using the CRE. It has altered the primarily seaborne economic model of coastal areas, promoting bidirectional circulation between the East and the West [18]. Additionally, the CRE significantly enhances the trade openness of cities and reduces customs clearance time [19,20]. The increasing number of CRE routes rapidly boosts China's export trade with economies along the route [21]. From the perspective of trade modes, the CRE is more conducive to the growth of processing trade compared to general trade [22].

On the other hand, some scholars argue that the opening of the CRE promotes the free flow of internal factors within regions, significantly contributing to innovation. In their study, Wang et al. [23] investigated the impact and mechanism of CRE operations on green total factor productivity (GTFP). They concluded that CRE operations improve the efficiency of high-end resource flows, which has a significant effect on innovation, and ultimately has a positive impact on GTFP. Li et al. [24] explored the impact of the CRE on urban innovation, discovering that the CRE, by promoting economic growth, financial development, and local government fiscal support, enhances the urban innovation system and thereby improves urban innovation levels. Wang and Bu [25] found that enterprises engaging in trade using the CRE are more likely to engage in foreign investment or attract foreign capital, continuously enhancing their innovation activities and driving innovative development. Moreover, the opening of the CRE can drive the industrial structure upgrading of cities along the route, with this effect being more pronounced in the central and western regions [26]. Some scholars have also explored the role of the CRE in promoting high-quality economic development from the perspective of total factor productivity [27]. A few researchers have investigated the relationship between the CRE and investment, mainly focusing on FDI. Huang and Wo [28] found that China's overall level of green investment efficiency is not high and exhibits significant country differences. The opening of the CRE alleviates corporate financing constraints and improves the efficiency of private enterprises' foreign investment [29]. In summary, despite certain achievements in the existing research on the CRE and its relationship with foreign trade, the overall focus remains on the mechanisms of trade growth resulting from the opening of the CRE. There is a scarcity of research on the investment issues that the opening of the CRE brings to cities. This paper seeks to explore and verify this aspect in the field.

3. Theoretical analysis and research hypothesis

The theoretical framework of neoclassical economics does not take into account transportation costs, which leads to the neglect of the study of the impact of transportation on socio-economic development. According to the trade cost theory, improving transportation infrastructure can effectively reduce transportation costs, minimize losses in the trade process, and indirectly increase the productivity of enterprises [30]. This in turn can promote regional economic development, facilitate the agglomeration and diffusion of regional economic activities, and change the spatial arrangement of economic activities. In the field of new economic geography, “first nature” refers to exogenous factors such as the geographic location of ports and the innate heterogeneity of the distribution of natural resources, while “second nature” refers to acquired factors such as the conditions of transportation infrastructure and the rules of regional agreements [31]. Both of these factors have a significant impact on economic geography and its spatial distribution. The launch of the CRE has improved the transportation infrastructure between China, countries and regions along the Belt and Road, and Europe. This has reduced transportation and time costs associated with cross-border trade and logistics, provided a more efficient mode of international transportation, and increased participation in the international industrial division of labor. CRE has also helped to overcome unfavorable natural conditions and different logistical endowments of parties involved in the international industrial division of labor. In addition, it has effectively improved the level of transportation infrastructure in inland areas [32]. Meanwhile, the opening of the CRE had a significant spillover effect on FDI inflow. The knowledge and technology spillover resulting from FDI inflow will benefit similar enterprises. Additionally, the decrease in time cost after the opening of the CR express allows similar enterprises to learn and imitate from FDI enterprises in a wider region, thereby promoting the improvement of their own production capacity. FDI can promote technological innovation and economies of scale for local upstream enterprises by requiring higher product standards and greater demand for intermediate inputs. Local firms can also benefit from FDI in upstream industries as buyers. These spillover effects may attract the inflow of other FDI enterprises. The introduction of the CRE as a new mode of transportation may have an impact on the attraction of FDI to the central and western regions through clustering and diffusion effects.

The opening of CRE provides a new option for domestic section and international section transportation and improves the construction of transportation infrastructure in the central and western regions [33]. The following is an explanation of the impact of CRE on attracting FDI to the Midwest based on the Footloose Capital (FC) model in the New Economic Geography [34]. It is assumed that there exist two regions A and B, and resource endowments, preferences, market size, and trade policies within the two regions are kept consistent. The quantity of labor within the two regions is $L$ and $L^{*}$, $L+L^{*}=L^T$; the quantity of capital is $K$ and $K^{*}$, $K+K^{*}=K^T$. The share of capital used in production within the two regions is $S_n$ and $S_n^{*}$. The types of products produced in the two regions are $n$ and $n^{*}$, $n+n^{*}=n^w$.

Suppose that the consumer's utility function is Equation (1):

$$U=C_M^{\mu }{C}_A^{1-\mu },Cn={\left({\int }_{i=0}^{n^w}{C}_i^{\left(\sigma -1\right)/\sigma }\right)}^{\sigma ∕\left(\sigma -1\right)},0<\mu <1<\sigma$$ (1)

$C_M$ and $C_A$ are the consumption of industrial and agricultural goods, respectively, $\mu$ is the share of consumer spending on industrial goods, and $\sigma$ is the elasticity of substitution between industrial goods.

The price index of industrial products is $P_M={\left({\int }_{i=0}^{n^w}{C}_i^{1-\sigma }\mathrm{d}i\right)}^{\sigma ∕\left(\sigma -1\right)}$,Average price of industrial goods purchased by consumers $\Delta =\left({\int }_{i\sum 0}^{n^w}{P}_i^{\sigma }\mathrm{d}i\right)∕n^w$。

For the industrial sector, the demand for industrial goods in this region as well as in region B by consumers in region A is given by Equation (2)

$$c_A=\mu E\frac{P_A^{-\sigma }}{P_M^{1-\sigma }},c_A^{*}={\mu E}^{*}\frac{{{(P}_A^{*})}^{-\sigma }}{{{(P}_M^{*})}^{1-\sigma }}$$ (2)

The demand for industrial goods in the region as well as in region A by consumers in region B is given by Equation (3)

$$c_B={\mu E}^{*}\frac{P_B^{-\sigma }}{{{(P}_M^{*})}^{1-\sigma }}，c_B^{*}=\mu E\frac{{{(P}_B^{*})}^{-\sigma }}{P_M^{1-\sigma }}$$ (3)

In the above two equations, $P_A$ and $P_A^{*}$ denote the sales prices of industrial products in region A in this region as well as in region B, respectively, $P_B$ and $P_B^{*}$ denote the sales prices of industrial products in region B in this region as well as in region A, respectively, and $E$ and $E^{*}$ denote the total expenditures in regions A and B, respectively.

In the D-S framework, the firm's operating profit is equal to a fixed share of sales, i.e., $\pi =px∕\sigma$. The firm's optimal pricing strategy is marginal cost-plus pricing, and due to the presence of iceberg costs, the product prices in the two regions is given by Equation (4)

$$p=\frac{w_L{\alpha }_m}{1-1∕\sigma },p^{*}=\frac{{\tau w}_L{\alpha }_m}{1-1∕\sigma }$$ (4)

The sales price of the firm in region A is $p$ and the sales volume is $c$; the sales price and sales volume of the firm in region B are $p^{*}=\tau p$ (due to the existence of iceberg transaction costs, the prices of the same goods in the two regions are different, and the price ratio is set to $\tau$) as well as $c^{*}$. The output of the firm is $x=c+\tau c^{*}$ and the total revenue given by Equation (5).

$$pc+p^{*}{c}^{*}=px$$ (5)

Assuming that the cost of the firm consists of fixed and variable costs, the fixed cost is one unit of capital, and the income from capital is $\pi$. The variable cost is labor, the output of the firm is $x$, each unit of output requires $a_M$ units of labor, and the income from labor is $w_L$, then the cost function of the firm $\pi +a_M{w}_{L}x$. According to the profit maximization principle, Equation (6) is derived.

$$px=\pi +a_M{w}_{L}x$$ (6)

The FC model tells us that the main determinant of capital flows is the difference in the rate of return on capital across regions, which leads to Equation (7).

$$\pi -{\pi }^A=b\frac{E^w\left(1-\varnothing \right)}{K^w\Delta {\Delta }^{*}}\left[\left(1+\varnothing \right)\left(s_E-\frac{1}{2}\right)-\left(1-\varnothing \right)\left(s_E-\frac{1}{2}\right)\right]$$ (7)

From equation (7), it can be seen that the effect of $\left(1+\varnothing \right)\left(s_E-\frac{1}{2}\right)$ on regional factor flows is positive and there is a facilitating effect, while the effect of $\left(1-\varnothing \right)\left(s_E-\frac{1}{2}\right)$ on regional factor flows is negative and there is a hindering effect. Since $\left(1+\varnothing \right)/\left(1-\varnothing \right)>1$, the increase of free trade degree $\varnothing$ will promote the free flow of factor (capital) resources. And it is known from $\varnothing ={\tau }^{1-\sigma }$ that the decrease of transportation cost due to the opening of CRE will increase the regional trade freedom. And the increase of trade freedom will promote the cross-regional flow of factor resources, thus promoting the regional ability to attract FDI. In other words, when CRE is opened, it will significantly increase the trade freedom of region A, thus promoting the attractiveness of city A to FDI when CRE is opened. Based on this analysis, this paper proposes the following hypothesis.

H1

CRE can facilitate the attraction of FDI in central and western China.

The ‘spatio-temporal convergence effect’ of CRE can enhance inter-regional connectivity and reduce barriers to factor flows, thereby catalyzing the reshaping of the spatial layout of regional industries [35]. As a junction station, the improvement of the opening city location conditions will be greater than the peripheral areas, attracting resources and factors to the center [36], thus reducing the cost of local enterprises to obtain information and production factors. Enterprises in peripheral areas, aiming to be close to factor markets and reduce costs, will move to the center, forming industrial clusters [37]. CRE will bring the expansion of market scale to the opening city, mainly in the form of increased population inflow and consumption level. On the one hand, the construction of CRE itself can create a large number of jobs and attract labor inflow [20]. On the other hand, the improvement of transportation infrastructure can reduce trade costs and indirectly improve the productivity of enterprises [30], which is conducive to the economic efficiency. The overall economic development of the region leads to an increase in consumption level, thus expanding the market scale. Past studies affirm the importance of industrial agglomeration and market size in the choice of location for FDI [8]. Therefore, this paper proposes the following hypothesis.

H2

CRE promotes the growth of foreign investment scale in the central and western regions by driving industry aggregation and expanding market scale.

The impact of substantial transportation infrastructure on the regional economy typically becomes progressively evident and strengthened over time. As some cities have been engaged with CRE for an extended duration, it has evolved into a standardized system with cost reduction and increased transportation frequency. These long-term collaborations have allowed for the optimization of regional cargo sourcing, making these cities more mature in their development compared to those that joined the network at a later stage. Simultaneously, as trade volume continues to expand, the operational intensity of CRE has risen. High operational intensity implies a substantial transportation capacity and efficient cargo collection and distribution. Frequent international trade activities further deepen regional openness, accelerate cross-regional factor and resource flows, and amplify the agglomeration effect. In conclusion, the longer and more intensively CRE operates, the more significant its contribution to enhancing the appeal of the region for foreign investment. This paper proposes the following hypothesis.

H3

There is a cumulative effect of time on the investment promotion effect played by CRE, and the increase in the intensity of the operation has a positive moderating effect.

CRE operates under an “axis-width” model [38]. The “axis” represents the hub node of CRE, serving as the focal point for cargo distribution and radiating its effects within the “width network.” The initiation of CRE expedites factor flow, and the subsequent gains and externalities encourage factor migration toward the center, resulting in an agglomeration effect. However, as time progresses, the marginal returns on production factors decrease, leading to the possibility of uneconomic agglomeration in the center. This scenario prompts the dispersion of various factors and resources to the periphery, giving rise to a diffusion effect (Wang et al., 2020). CRE not only brings agglomeration effects to the cities in which it operates but also generates externalities in economic activities, leading to knowledge spillover effects [39]. The movement of people facilitates the dissemination of valuable knowledge and experience from the central hub to the periphery, thereby promoting economic development in peripheral regions. In the later stages of development, relying on the factors and resources influenced by the diffusion effect and incorporating advanced knowledge and experience through spillover effects, the peripheral regions leverage their latecomer advantages. They improve their location conditions via the radiation-driven effect of the central hub [40,41], thereby enhancing their attractiveness to FDI. Therefore, this paper proposes the following hypotheses.

H4

The opening of CRE has a positive radiation effect on attracting FDI to the surrounding area.

4. Econometric strategy

4.1. Estimation framework

The DID model is a commonly used method for assessing the implementation effect of a policy, as it can avoid the interference of external factors to a certain extent. The basic idea behind the DID method is to compare the difference between the experimental group and the control group before and after the policy shock to form a double difference statistic, which can then be used to identify the policy shock effect. To assess the net effect of policy implementation, the test process will be conducted twice: once with the treatment group and once with the control group, both before and after the policy is implemented. The difference between the treatment and control groups is then calculated, which is known as the ‘double difference’ method. Compared with traditional DID method, which has the characteristics of uniform policy implementation time, staggered DID method is suitable for the gradual implementation of the same policy among the affected groups. Following the methodology of Yu et al. [42], this paper considers the opening of CRE as a quasi-natural experiment, with ‘CRE cities' as the treatment group and other cities as the control group. Since the opening time of CRE in different cities is not consistent, this paper adopts the staggered DID method to construct the baseline regression model as follows (Equation (8)):

$$lnfd{i}_{it}={\beta }_0+{\beta }_{1}Polic{y}_{it}+{\beta }_2{X}_{it}+{\theta }_i+{\mu }_t+{\epsilon }_{it}$$ (8)

As above, $fd{i}_{it}$ is the size of foreign investment in city $i$ in period $t$; $Polic{y}_{it}$ is the policy shock, indicating whether city $i$ has opened CRE at time $t$; $X_{it}$ is the control variable; ${\theta }_i$ is the city fixed effect, ${\mu }_t$ is the time fixed effect; ${\epsilon }_{it}$ is an error term.

4.2. Data and variables

The data for the empirical analysis presented in this section are collected by the authors from various official statistical publications and public databases. Using these data sources, we construct a dataset of 152 Chinese cities above prefecture level belonging to the central and western regions for the period 2008 to 2020. The sample includes more than 82 % of the prefecture-level cities in central and western China (186). Efforts were made to ensure that the sample covered almost all representative cities in each province, such as provincial capitals.

4.2.1. Explained variable: FDI($lnfd{i}_{it}$)

The explanatory variable of this paper is the level of utilized foreign capital of cities, and the logarithmic value of the total amount of actual utilized foreign capital of cities in that year is chosen to express it. Since the value recorded in the statistical yearbook is in USD 10,000, this paper converts it according to the average exchange rate of the year after organizing the data, and finally converts it to the value in RMB 10,000.

4.2.2. Explanatory variable: CRE ($Polic{y}_{it}$)

The core explanatory variable is the policy shock of CRE opening, which is set to 1 in the year of opening and later years, and 0 otherwise. This paper only counts the opening of CRE in central and western cities by the end of 2019. Considering that the policy effect of CRE has a certain lag, this paper refers to the treatment in Xiao et al. (2022), which lags the opening of CRE by one year for cities that open in July of the year and later.

4.2.3. Control variables

This paper draws on the ideas of scholars such as Wong et al. [43], Hu et al. [44] to select control variables. The level of economic development ($lnpgdp$) is measured by the logarithm of GDP per capita; accessibility ($lntra$) is measured by the logarithm values of urban road area per capita; industrial development ($ser$) is measured by the share of tertiary sector in GDP; financial development ($fin$) is measured by the total deposits and loans of financial institutions as a share of GDP at the end of the year, urbanization ($urb$) is measured by the urban resident population as a proportion of the city's resident population; informatization ($lnnet$) is measured by the urban resident population as a proportion of the city's resident population.

The descriptive statistics of the variables are shown in Table 2.

Table 2.

Table of descriptive statistical analysis.

Variables	N	Mean	Medians	SD	Minimum	Maximum
$lnfdi$	1976	20.77	20.85	1.784	12.22	25.23
$Policy$	1976	0.053	0	0.224	0	1
$lnpgdp$	1976	10.43	10.44	0.596	4.595	12.46
$lntra$	1976	2.686	2.671	0.426	0.315	4.096
$ser$	1976	0.389	0.379	0.094	0.148	0.727
$fin$	1976	2.453	1.911	3.815	0.731	70.85
$urb$	1976	0.504	0.488	0.133	0.223	0.937
$lnnet$	1976	13.01	12.98	0.952	7.958	16.47

5. Empirical results

5.1. Baseline estimates

Regression using formula (8), the results are shown in Table 3. As can be seen from Table 3, the regression results of Policy are all significantly positive, indicating that the opening of the CRE has a significant role in promoting the growth of FDI in the central and western cities. Taking column (7) as the base result to estimate the meaning of the coefficients, it means that compared with the central and western cities without the opening of the CRE, the FDI scale of the central and western cities with the opening of the CRE will increase by 36.6 %. Similarly, Zhou and Hao [45] analyses that the FDI scale of the cities with the opening of the CRE will increase by 24.5 % more than that of the cities without the opening of the CRE; and the FDI greenfield investment scale of the enterprises in the cities with the opening of the CRE will increase by 36.3 % more than that of the enterprises in the cities without the opening of the CRE. Therefore, Hypothesis 1 is verified.

Table 3.

Baseline regression results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
$Policy$	0.343***	0.311**	0.315***	0.314***	0.315***	0.333***	0.366***	0.144*
	(2.77)	(2.56)	(2.58)	(2.60)	(2.61)	(2.76)	(3.03)	(1.82)
$lnpgdp$		0.926***	0.923***	1.106***	1.097***	1.048***	0.972***	0.912***
		(8.01)	(7.97)	(9.32)	(9.21)	(8.79)	(8.00)	(16.17)
$lntra$			0.0588	0.0253	0.0258	0.00541	−0.00631	−0.109**
			(0.58)	(0.25)	(0.26)	(0.05)	(-0.06)	(-2.56)
$ser$				3.886***	3.907***	3.910***	3.812***	−1.115***
				(5.98)	(6.01)	(6.04)	(5.89)	(−4.93)
$fin$					−0.0103	−0.00861	−0.00750	−0.00689
					(-1.02)	(-0.86)	(-0.75)	(-1.42)
$urb$						2.616***	2.455***	0.391*
						(3.77)	(3.53)	(1.68)
$lnnet$							0.254***	−0.132***
							(3.11)	(−3.03)
L.lnfdi								−0.0548***
								(-6.29)
City fixed effects	YES	YES	YES	YES	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES	YES	YES	YES	YES
Constant	20.25***	11.24***	11.12***	8.083***	8.179***	7.580***	5.438***	14.86***
	(285.9)	(9.96)	(9.72)	(6.5)	(6.56)	(6.05)	(3.81)	(34.74)
N	1976	1976	1976	1976	1976	1976	1976	1672
R2	0.090	0.121	0.121	0.138	0.138	0.145	0.150

Note: Values in parentheses are values of the t-statistic at the robust standard error; *, **, and *** are significant at the 10 %, 5 %, and 1 % levels, respectively. Same for the following tables.

In the regression results of control variables, the level of informatization, urbanization, industrial development, and economic development have a significant positive effect on the FDI inflow in the central and western regions; the coefficient of the degree of financial development is negative and non-significant, which may be due to the fact that financial development in the central and western regions lags behind, and it fails to significantly affect the inflow of foreign capital for the time being; the coefficient of the degree of transportation convenience is negative and non-significant, which is probably due to the fact that the transportation effect of the CRE has crowded out the original traffic of the city. In addition, dynamic estimation and interpretation of the investment climate is needed to understand how it explains future FDI in China. The systematic GMM model regression results are reported in column 8 of Table 3. It can be found that the coefficient of CRE opening on attracting FDI to the central and western regions is still significantly positive after considering the FDI lag term, but the strength of the effect is gradually weakening. Therefore, this study concludes that the opening of the CRE has a certain time lag in attracting FDI.

The possible reasons for the positive effect of the opening of CRE on the attraction of FDI to the central and western regions are: Firstly, the commencement of the CRE serves as a signalling factor for the location selection of multinational enterprises' investments. For multinational enterprises, investing and operating in cities where the CRE operates allows for lower costs and shorter timeframes to reach the Eurasian market. This undoubtedly presents significant new opportunities for market expansion and profit growth. Secondly, the impact of the CRE on the open economy of cities in the central and western regions is primarily reflected in the changes to international transportation channels. Being inland and distant from coastal ports, the central and western regions, which previously relied mainly on maritime transportation for international trade, magnified the geographical disadvantage. This significantly hindered inland areas from engaging in foreign exchanges and trade activities. The opening of the CRE not only improves the level of transportation infrastructure in the central and western regions but also enhances their external accessibility. The point-to-point transport characteristic effectively improves transportation timeliness, reduces transport distances, and enhances cost structures. On one hand, the opening of the CRE shortens the temporal and spatial distance to foreign markets. On the other hand, with government subsidies as a foundation, the scale effects gradually demonstrated by the CRE will further compress marginal costs, reflecting outstanding price advantages. Thirdly, the enhancement of logistics distribution quality and the improvement of the business environment in cities where the CRE operates will attract more foreign investment and business activities. Fourthly, cities where the CRE operates also serve as distribution centre for goods. The improvement of transportation infrastructure can reduce market segmentation and promote market integration. The increase in market correlation contributes to leveraging economies of scale and agglomeration effects, attracting inflows of foreign capital.

5.2. The dynamic impacts of the CRE opening on FDI attraction in the central and western regions

The baseline regression may violate the “parallel trend” and lead to unstable results. To better observe the dynamic effects of CRE policy, this paper uses event analysis to analyses the time-varying effects of the policy in different periods and to test the parallel trend hypothesis. As shown in Fig. 1, in the period before the opening of CRE, the regression results are all insignificant, indicating that the treatment and control groups show basically the same trend of change, which can be concluded to be consistent with the premise of the parallel trend hypothesis. In the period after the operation of CRE, the regression results are significant, indicating that the operation of CRE affects FDI attraction in the central and western regions.

Fig. 1.

Parallel trend test.

5.3. Heterogeneity analysis

The current uneven regional development in China, with significant differences in the level of sophistication of supporting facilities and economic development in different regions, may lead to spatial heterogeneity in the investment attraction effect of CRE opening. The sample cities are further divided into central and western cities according to the administrative division of China. Columns 1–2 of Table 4 show the heterogeneity results of the FDI attraction effect of CRE in central and western cities, respectively. In column 1, the coefficient of the core explanatory variable is not significant. However, in column 2, the coefficient of the core explanatory variable is significant at the 1 % level, indicating that the capital attraction effect of CRE mainly works in the western region. On the one hand, the western region is affected by geographic location and historical policies, and is lagging behind in infrastructure construction and economic development. The opening of the CRE improves the transportation disadvantage of the western region, and promotes the western region's “bringing in” and “going out” to attract foreign investment. On the other hand, the central region is closer to the coastal ports than the western region, and it mainly focuses on low-cost and low-technology products, which makes it less dependent on the CRE, or the transportation infrastructure in the central region is relatively perfect, which squeezes out the impact of the CRE on attracting foreign investment, which may be the reason why the core explanatory variables of the central region are not significant.

Table 4.

Heterogeneity analysis of the impact of CRE on FDI.

Variables	Based on Geographical Location		Based on City Characteristics
	(1) Central Region	(2) Western Region	(3) Large and Above Cities	(4) Small and Medium-sized Cities	(5) Resource-Based Cities	(6) Non-Resource-Based Cities
$Policy$	0.0264	0.840***	0.357**	0.371	0.178	0.315***
	(0.22)	(3.13)	(2.29)	(1.09)	(0.41)	(2.70)
$lnpgdp$	1.766***	0.568***	1.398***	0.672***	0.637***	1.675***
	(11.93)	(2.74)	(6.91)	(3.77)	(3.80)	(8.03)
$lntra$	0.0514	0.0109	0.0938	−0.0633	−0.0150	−0.0972
	(0.45)	(0.06)	(0.81)	(−0.36)	(−0.09)	(−0.79)
$ser$	4.480***	0.235	4.015***	3.341***	3.683***	4.289***
	(6.87)	(0.17)	(5.13)	(3.08)	(3.29)	(5.59)
$fin$	−0.0193**	0.194***	−0.00961	−0.00433	0.0359	−0.00562
	(−2.34)	(3.64)	(−1.18)	(−0.07)	(0.58)	(−0.62)
$urb$	2.497***	−2.157	1.805*	2.331**	1.584*	3.974***
	(4.20)	(−0.85)	(1.66)	(2.30)	(1.69)	(3.47)
$lnnet$	0.362***	0.221	−0.0738	0.634***	0.650***	0.0453
	(3.85)	(1.49)	(−0.81)	(3.82)	(4.00)	(0.51)
City fixed effects	YES	YES	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES	YES	YES
Constant	−3.740***	11.76***	5.453***	3.956*	4.220***	0.693
	(−2.32)	(4.18)	(2.59)	(1.67)	(1.87)	(0.32)
N	1287	689	1092	884	845	1131
R2	0.323	0.106	0.257	0.122	0.143	0.200

The differences in city size and resource endowment may have different effects on the attraction effect of CRE. In terms of city size, this paper analyzes the heterogeneity of the sample based on the size of the resident population in urban areas into large and above cities, and small and medium-sized cities as required by the document “Notice on the Criteria of City Size Classification” issued by the China in 2014. Columns 3 and 4 of Table 4 show that the coefficient of the core explanatory variable is significantly positive in large and above cities, but insignificant in small and medium-sized cities, indicating that the FDI attraction effect of CRE opening is mainly in large and above cities. The reason is that large and above cities have better supporting infrastructure and business environment. In terms of resource endowment, this paper categorizes the sample cities into resource-based cities and non-resource-based cities according to the list of national resource-based cities announced by China in the National Sustainable Development Plan for Resource-based Cities (2013–2020) in 2013. Columns 5–6 of Table 4 show that the FDI attraction effect of CRE mainly works on non-resource-based cities. Many resource-based cities are currently facing resource depletion, which may result in the outflow of factors, and the slow and difficult transformation of resource-based cities, so the impact of CRE opening on resource-based cities is not obvious.

5.4. Robustness checks

5.4.1. PSM-DID

Due to the endogeneity problem of sample self-selection, the propensity score matching method (PSM) is chosen in this paper. The control variables in the baseline regression are used as covariates and robustness analysis is performed using the one-to-one nearest neighbour matching method with a matching radius of 0.01. The PSM equilibrium test showed that the absolute values of standard deviations for all variables decreased significantly after matching, and the p-values were all greater than 0.05, indicating that there were no significant systematic differences between the treatment and control groups, thus demonstrating that the equilibrium test was passed. After that, DID regressions were estimated for the samples matched by the PSM method. The results of the PSM-DID regressions in column 1 of Table 5 show that the coefficient of the policy is still significantly positive, verifying the robustness of the baseline regression results.

Table 5.

Robustness test results.

Variables	(1) PSM-DID	(2) Excluding Other Policy Interferences During the Same Period	(3) Exclusion of the financial crisis and the Covid-19 pandemic
$Policy$	0.471***	0.302**	0.381***
	(2.73)	(2.49)	(2.95)
$FTA\times Post$		0.558***
		(3.86)
$lnpgdp$	2.502***	0.950***	0.887***
	(5.12)	(7.84)	(6.76)
$lntra$	0.388	0.00962	0.073
	(1.59)	(0.10)	(0.66)
$ser$	8.254***	3.720***	3.360***
	(4.19)	(5.77)	(4.60)
$fin$	0.0807	−0.00695	−0.003
	(0.75)	(−0.70)	(−0.21)
$urb$	4.839***	2.412***	3.310***
	(2.83)	(3.49)	(4.19)
$lnnet$	0.0576	0.262***	0.253***
	(0.35)	(3.21)	(2.61)
City fixed effects	YES	YES	YES
Year fixed effects	YES	YES	YES
Constant	−10.68**	5.568***	5.665***
	(-2.30)	(3.92)	(3.42)
N	325	1976	1672
R2	0.3874	0.157	0.138

5.4.2. Excluding other policy shocks in the same period

China has been building free trade zones in several places to expand the level of opening up to the outside world. Since 2016, several provincial zones in central and western China have gradually become FTAs. In order to control the impact of “FTA establishment”, this paper defines the dummy variable (FTA) to indicate whether it is a pilot city of FTA, and the dummy variable takes 1 if it is, otherwise it takes 0. Post is a time dummy variable, taking 1 for the year after the implementation of the FTA pilot cities and 0 for the rest. FTA × Post is the interaction term of “FTZ” policy and time. As shown in Column 2 of Table 5, the coefficient of Policy is still significantly positive, which again confirms the robustness of the results.

5.4.3. Exclusion of the financial crisis and the Covid-19 pandemic

Two typical external financial shocks that existed during the sample period studied in this paper are the 2008 global financial crisis and the 2020 Covid-19 pandemic. Considering that such factors are not easy to be measured by specific variables, this paper adopts the exclusion of the 2008, 2009 and 2020 samples to exclude the interference of the 2008 financial crisis and the 2020 Covid-19 pandemic. The 2010–2019 samples are included in the regression equation and the results are shown in column (3) of Table 5. The results show that the coefficient of Policy is significantly positive at the 1 % level in both scenarios, which indicates that the opening of CR express still has a significant effect on the city's attraction of FDI.

5.4.4. Placebo test

In order to prevent unobservable city characteristics factors from biasing the empirical results, this paper conducts a placebo test by repeating a random sample 500 times for regression simulation. Fig. 2 shows the distribution of P-values for the coefficients of the 500 regression simulations, which roughly follows a normal distribution, with the estimated coefficients concentrated around 0. In addition, the vertical dashed lines in the figure represent the true regression coefficients, which are clearly abnormal to the regression coefficients of the placebo test, and it can be seen that the results are not obtained randomly. In summary, the policy effect of CRE is not seriously biased by unobservable factors, and passes the placebo test.

Fig. 2.

Placebo test.

5.5. Mechanism analysis

This paper refers to the method of Wen and Ye [46] to test the mediating effect through stepwise regression. The model is constructed as follows (Equations (9), (10))).

$$M_{it}={\alpha }_0+{\alpha }_{1}Po\text{lic}{y}_{it}+{\alpha }_2{X}_{it}+{\theta }_i+{\mu }_t+{\epsilon }_{it}$$ (9)

$$lnfd{i}_{it}={\gamma }_0+{\gamma }_{1}Polic{y}_{it}+{\gamma }_2{M}_{it}+{\gamma }_3{X}_{it}+{\theta }_i+{\mu }_t+{\epsilon }_{it}$$ (10)

In the above equation, ${\mathrm{M}}_{\text{it}}$ represents the mediating variables, which denote industrial agglomeration ($lnfirm$) and market size ($lnmarket$), respectively. Industrial agglomeration is measured using the logarithm of the number of above-scale industrial enterprises in the city; market size is measured by taking the logarithm of the total retail sales of consumer goods in the city, which can effectively reflect the degree of realization of consumers' purchasing power and better show the size of the market.

As shown in Table 6, Columns 1 and 3 demonstrate the regression results of the mediating effect of industrial agglomeration. Based on the significantly positive coefficients of the core explanatory variables in column 1 and the industrial agglomeration ($lnfirm$) in Column 3 and the coefficient of Policy decreases compared to the baseline regression results, it indicates that CRE does increase the attractiveness of FDI by promoting industrial agglomeration. Columns 2 and 4 show the regression results of the mediating effect of market size. The coefficients of the core explanatory variables in Column 2 and the market size ($lnmarket$) in Column 4 are significantly positive, and the coefficient of Policy decreases compared to the baseline regression results, indicating the presence of partial mediating effects. H2 is verified.

Table 6.

Regression results of Mechanism analysis.

Variables	(1) $lnfirm$	(2) $lnmarket$	(3) $lnfdi$	(4) $lnfdi$
$Policy$	0.0641***	0.0529**	0.300**	0.338***
	(2.59)	(2.17)	(2.55)	(2.81)
$lnfirm$			1.077***
			(9.62)
$lnmarket$				0.529***
				(4.56)
$lnpgdp$	0.326***	0.343***	0.621***	0.790***
	(13.06)	(13.99)	(5.01)	(6.21)
$lntra$	0.0237	−0.0489**	−0.0318	0.0196
	(1.15)	(−2.40)	(−0.32)	(0.19)
$ser$	−1.161***	0.406***	5.062***	3.598***
	(−8.75)	(3.11)	(7.86)	(5.58)
$fin$	−0.00440**	−0.00270	−0.00276	−0.00607
	(−2.14)	(−1.33)	(−0.28)	(−0.61)
$urb$	−0.298**	−0.358**	2.776***	2.644***
	(−2.09)	(−2.55)	(4.09)	(3.82)
$lnnet$	0.161***	0.0569***	0.0810	0.224***
	(9.59)	(3.44)	(0.99)	(2.75)
City fixed effects	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES
Constant	1.583***	19.75***	3.733***	−5.018*
	(5.41)	(68.61)	(2.66)	(−1.86)
N	1976	1976	1976	1976
R2	0.390	0.887	0.191	0.159

5.6Expanded analysis: spatio-temporal effects test.

5.5.1. Time effect test

Since the opening time of each city in the treatment group is not consistent, the cumulative operating time of each opening city is also different. This paper will introduce the time variable $Policy\_t$, the cumulative opening year of CRE, to examine the cumulative effect of CRE on FDI inflows over time. The regression results are shown in Column 1 of Table 7. The coefficient of $Policy\_t$ is significantly positive, indicating that the positive effect of CRE on the introduction of foreign investment into the region gradually increases with the increase of opening time, indicating that there is a time cumulative effect of this effect.

Table 7.

Time cumulative effect and inspection based on operation intensity.

Variables	(1) Time Cumulative Effect	(2) Operation Intensity
$Policy\_t$	0.0843***
	(2.94)
$Policy\times Number$		0.0353**
		(2.37)
$lnpgdp$	0.982***	0.991***
	(8.09)	(8.16)
$lntra$	0.00136	−0.00616
	(0.01)	(-0.06)
$ser$	3.812***	3.837***
	(5.89)	(5.93)
$fin$	−0.00743	−0.00710
	(−0.74)	(−0.71)
$urb$	2.376***	2.411***
	(3.42)	(3.47)
$lnnet$	0.253***	0.239***
	(3.09)	(2.92)
City fixed effects	YES	YES
Year fixed effects	YES	YES
Constant	5.363***	5.447***
	(3.76)	(3.81)
R2	0.150	0.150

In addition, due to the differences in city size and cargo base, the intensity of CRE operations varies widely even among cities operating in the same year. In order to effectively identify whether the operating intensity exerts an influence on the FDI attraction effect of CRE, this paper introduces the interaction term of CRE ($Policy$) and operating intensity ($Number$) for verification, and constructs the model as follows (Equation (11)).

$$lnfd{i}_{it}={\beta }_0+{\beta }_{1}Polic{y}_{it}\times Numbe{r}_{it}+{\beta }_2{X}_{it}+{\theta }_i+{\mu }_t+{\epsilon }_{it}$$ (11)

$Number$ represents the intensity of liner operation, which is measured by using the logarithmic value of the number of CRE operating cities in the calendar year. This paper focuses on the core coefficient ${\beta }_1$. If the regression result of ${\beta }_1$ is positive, it indicates that the greater the intensity of CRE operation, the stronger the attractiveness to FDI.

The coefficient of the interaction term $Policy\times Number$ is significantly positive at the 5 % level, as shown in Column 2 of Table 7, indicating that the intensity of the operation has a positive moderating effect on the promotion effect played by CRE. In summary, the longer the CRE runs and the greater the intensity of operations, the greater the promotion effect on enhancing the attractiveness of regional FDI. Thus, H3 can be verified.

5.5.2. Spatial effect test

CRE facilitates the connection of “Axis” sites and the source points within the “Network” by rail and various transport modes. This will further drive the development of export-oriented economy in the surrounding areas. This paper will verify whether the “axis-width” organization mode of the CRE will have a radiating effect on the surrounding cities' ability to attract FDI and make a rough projection of its radiation range.

In this paper, we refer to Wang and Bu (2019) and set up a set of dummy variables for the axial amplitude distance DIST. The sample of cities that have opened CRE during the experimental statistical period is removed from the full sample, and the spherical shortest distance from the location of the unopened city to the nearest CRE originating station is calculated using the latitude and longitude coordinates. When the shortest distance between the two lies within the intervals of [0,100 km], [0,150 km], [0,200 km], [0,250 km], [0,300 km], and [0,350 km], the city is included in the treatment group, i.e., the axial amplitude distance DIST is taken as 1, otherwise it is taken as 0; AFTER is the nearest CRE The opening time of the originating station is taken as 0 before opening and 1 after opening; the interaction term AFTER × DIST forms a policy variable, which reflects the radiation of CRE to the surrounding area FDI.

Table 8 shows the regression results based on different axis distances. Columns 1–6 show that the coefficient of the interaction term AFTER × DIST is significantly positive when the axis distance is within 100 km, 150 km, 200 km, 250 km, and 300 km, indicating that the opening of CRE has a positive radiative effect on attracting foreign investment to the surrounding area, which can verify H4. Column 6 shows that the coefficients of the explanatory variables are no longer significant when the axis amplitude distance exceeds 300 km. Therefore, the regression results in the following table can confirm that CRE can radiate to attract FDI in the surrounding areas, and the effective range of radiation is roughly within 300 km.

Table 8.

Inspection of spatial effect and radiation range.

	(1)DIST<100	(2)DIST<150	(3)DIST<200	(4)DIST<250	(5)DIST<300	(6)DIST<350
AFTER × DIST	0.257**	0.277***	0.285***	0.244***	0.209***	0.0256
	(2.35)	(3.31)	(3.62)	(3.13)	(2.69)	(0.32)
$lnpgdp$	0.882***	0.865***	0.844***	0.864***	0.890***	0.910***
	(6.72)	(6.59)	(6.41)	(6.57)	(6.80)	(6.95)
$lntra$	−0.0909	−0.0649	−0.0439	−0.0503	−0.0633	−0.0774
	(−0.82)	(−0.59)	(−0.40)	(−0.46)	(−0.57)	(−0.70)
$ser$	3.878***	3.695***	3.598***	3.693***	3.792***	3.895***
	(5.54)	(5.27)	(5.12)	(5.26)	(5.41)	(5.56)
$fin$	−0.00797	−0.00719	−0.00717	−0.00711	−0.00700	−0.00816
	(−0.76)	(−0.69)	(−0.68)	(−0.68)	(−0.67)	(−0.77)
	(3.25)	(3.25)	(3.11)	(3.08)	(3.24)	(3.30)
$lnnet$	0.310***	0.313***	0.298***	0.301***	0.318***	0.301***
	(3.25)	(3.29)	(3.14)	(3.16)	(3.33)	(3.14)
City fixed effects	YES	YES	YES	YES	YES	YES
Year fixed effects	YES	YES	YES	YES	YES	YES
	(3.66)	(3.75)	(3.98)	(3.84)	(3.54)	(3.51)
R2	0.135	0.139	0.140	0.138	0.137	0.133

6. Conclusions and policy implications

This paper utilized panel data spanning from 2008 to 2020, covering 152 cities above the prefecture level in central and western China, and applied the staggered DID method for empirical analysis. The research findings reveal the following: First, the opening of CRE has a positive effect on attracting FDI to the central and western regions of China, and the results are still consistent with the benchmark regression after the robustness test using the matching method (PSM-DID), the placebo test, and the adjustment of the sample size. The opening of CRE has realized the “portization” of inland areas, improved the level of international logistics facilitation, expanded the opening-up level of the central and western regions, and improved the location conditions, thus enhancing the attractiveness of foreign direct investment. Secondly, there is regional heterogeneity in the promotion effect of CRE. From the perspective of geographic location, the opening of CRE has a significant effect on attracting foreign investment in the western region, but not in the central region; from the perspective of city scale, the opening of CRE has a significant effect on attracting foreign investment mainly in large and above cities, and the effect on small and medium-sized cities is not significant. Thirdly, CRE mainly improves the regional capital attraction capacity by promoting regional industrial aggregation, expanding market scale and other paths. This paper conducts the mediation effect test through the distribution regression method, and the results show that the above two transmission paths are significant at the empirical level. The opening of CRE improves the degree of inter-regional access, promotes the flow and optimal allocation of production factors between regions, which brings about the expansion of market scale and industrial aggregation, enhances the location advantage, and thus improves the attractiveness to foreign investment. Fourthly, from the perspective of time, the promotion effect of CRE on FDI inflow has the cumulative effect of time, and the intensity of operation of CRE has a positive adjustment effect on the promotion effect. With the increase of the opening time of the CRE, the development gradually becomes mature, the operation intensity is also increasing, and the driving effect on the development of export-oriented economy in the region will also be enhanced accordingly. The empirical results show that the longer the operation time and the more the number of columns of CRE trains, the greater the promotion effect on enhancing the attractiveness of regional foreign investment. Fifth, from a spatial point of view, the opening of CRE has a positive radiation effect on the attraction of FDI in the surrounding areas, and the radiation is roughly within 300 km. The “axis-width” organization mode of CRE can play a radiating and driving effect on the investment attraction capacity of the surrounding cities, which is of great significance in promoting the balanced development of the region.

The opening of the CRE has had a significant impact on the economic development of the central and western regions of China, particularly in attracting FDI. The CRE has acted as a catalyst in enhancing the regions' attractiveness to foreign investors by improving connectivity and streamlining logistics. As a result, inland areas have been effectively transformed into attractive investment destinations. The liner reduces trade barriers and boosts investor confidence in the region's accessibility and market potential by providing faster and more reliable transportation of goods between China and Europe. It is important to note that the impact of the CRE on attracting FDI varies by region. While the western region has experienced a significant positive impact, the central region seems to have been less affected. Regional differences may be attributed to various factors, such as the level of economic development, industrial structure, and geographic proximity to the CRE terminal. Policymakers must understand these differences to formulate effective investment promotion strategies. The CRE has multifaceted mechanisms that enhance the ability of the central and western regions to attract foreign investment. The liner primarily promotes regional industrial agglomeration and market scale expansion, attracting capital to the region. The CRE facilitates the flow and optimal allocation of production factors, making the region more attractive to foreign investors seeking strategic locations. The cumulative effect of the CRE on foreign capital inflows highlights the importance of long-term infrastructure investment. As the duration and intensity of CRE operations increase, their positive impact on foreign investment attraction becomes more pronounced. This emphasizes the importance of sustained efforts in developing and optimizing rail infrastructure. Additionally, the spatial dimension of the CRE's impact is not limited to the immediate area served by the rail line. The radiative effects observed within approximately 300 km of the CRE routes highlight the role of the train in attracting investments to the surrounding areas. These spatial spillover effects underscore the importance of regional cooperation and coordinated development initiatives to fully realize the potential of the CRE in driving economic growth and prosperity. In summary, the opening of the CRE has significantly increased the attractiveness of central and western China to foreign investment. The CRE trains have played a significant role in driving economic development in these regions by improving connectivity, facilitating industrial clustering, and stimulating market expansion. This presents promising prospects for future growth and prosperity.

The paper concludes that the opening of the CRE has effectively attracted foreign investment in the central and western regions. To further utilize the economic driving effect of the CRE on the inland areas, it is suggested to narrow the gap between the east and the west and help the formation of the ‘double-circle’ development pattern by improving infrastructure in central and western regions, with a focus on accelerating construction around the CRE. Furthermore, enhance the liner's capacity to attract foreign investment to the region. The efficient operation of the CRE depends on the support of various infrastructures. For example, the collection and distribution of goods at the liner's departure station necessitates the connection and supplementation of a multimodal logistics network. In addition, standardized park warehouses are required for the temporary storage of goods. There is a notable difference in infrastructure development between the eastern regions and the central and western regions, with the latter being relatively underdeveloped. To address this issue, the next step in the deployment of work should focus on increasing investment in infrastructure in the central and western regions. This can be achieved by improving the quality of infrastructure from two perspectives. The logistics network system should aim to improve traffic conditions, enhance the response speed of urban logistics, and optimize the efficiency of cargo sourcing and distribution for the CRE. Furthermore, it is important to perfect the hardware and software facilities for the operation of the CRE, and strengthen the application of cloud computing, mobile Internet, and other transportation technologies to improve customs clearance efficiency and reduce transportation costs. Improving the operational quality of the CRE will drive the development of the regional export-oriented economy, enhance its location advantage, and further narrow the gap between the east and the west. Additionally, promoting the growth of the CRE and planning the spatial layout of liner train stations in a rational manner are crucial. Currently, the stable development of the CRE route has provided numerous policy benefits and development opportunities for the central and western regions. It is necessary to improve the construction of the CRE route in these regions and increase its growth in terms of operating cities and frequency of operations. When launching cities, it is crucial to plan the spatial layout of liner sites in a logical and efficient manner. This will ensure a balanced and well-structured urban environment. Additionally, it is important to maximize the peripheral radiation effect of the ‘axis-width’ organization mode of CRE. To prevent unfair competition, such as acquiring cargo sources at low prices due to an excessive number of axle stations in the ‘amplitude network’, it is recommended to implement an opening audit system and suspend or avoid opening regions with weak industrial foundations and small cargo sources. Furthermore, the number of liner trains should not be increased blindly, but rather raised based on stabilizing cargo sources. To achieve the annual target for the development of the CRE, it is important to limit the vacancy rate to control transportation costs. Supporting the development of local industries and stabilizing the supply of suitable iron cargo sources are also crucial. Establishing a scientific and efficient normalized operation frequency is necessary to achieve high-quality scale growth as the construction goal. Thirdly, policies should be formulated to optimize the business environment and enhance its attractiveness to foreign investment by relying on the influence of CRE on industrial aggregation and market scale. The launch of the CRE has reduced the economic distance between regions, lowered the cost of factor flow, and facilitated the spatial aggregation of industries, thereby expanding the market scale. This has increased the demand for a favorable business environment in the central and western regions. Compared to the eastern region, the market system in the central and western regions is less developed, and local protectionism is a prominent issue. Therefore, the central and western regions should take the opportunity to stabilize the domestic economy, proactively remove administrative protection of local industries, improve marketization, and promote optimal resource allocation. Efforts should be made to create a business-friendly environment that is conducive to institutional growth. This goal can be accomplished by enhancing the legal and financial support systems, as well as by formulating policies that provide tax exemptions and subsidies to increase the appeal of foreign investment.

This paper acknowledges certain limitations due to the lack of available data, which prevents the study from being refined to the micro-firm level. Scholars currently use the fdi Market database, which contains information on global corporate investment, to conduct quantitative analyses of international capital flows. However, the existing data have a large number of missing values when collecting specific FDI amounts and cannot constitute balanced panel data. Therefore, it indirectly prevents the use of the staggered DID model. However, the staggered DID model is the most commonly used and reasonable empirical model for assessing policy effectiveness. Therefore, it is worthwhile to analyze whether the CRE has a positive effect on promoting enterprises in the central and western regions.

Data availability statement

Data will be made available on request.

Funding

This work was supported by the National Social Science Fund of China under Grant [21BJL107].

CRediT authorship contribution statement

Jingjing Liu: Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Zongbin Zhang: Writing – review & editing, Visualization, Validation, Supervision. Tingwei Chen: Writing – review & editing, Writing – original draft, Supervision, Software, Resources, Project administration.

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.

Appendix A. Supplementary data

The following are the Supplementary data to this article: