The Effect of High-Speed Rail on Cropland Abandonment in China

Abstract

This study analyzed Chinese labor-force survey data to determine the relationship between cropland abandonment and high-speed rail (HSR) infrastructure. A PSM-DID approach was employed to examine 2014 and 2016 data from the China Labor-force Dynamics Survey and estimate the impact of HSR, from which it was found that HSR accessibility promoted cropland abandonment in local farm households with a coefficient of 0.206, that is, HSR projects led to a 20.6% increase in area of cropland abandonment and these impacts were found to be greater in hilly areas and lower in plain areas. The results also suggested that HSR accessibility could have a “pull” effect, which resulted in rural labor force shifts to non-agricultural sectors in the local region. Countermeasures and policy suggestions are given to reduce cropland abandonment.

1. Introduction

The relationship between high-speed rail (HSR) and land use has received significant global research attention. The construction of China’s HSR started in 2005 and began operations in 2008, after which the network developed rapidly [1]. By the end of 2020, China’s HSR network had reached 38,000 km, ranking it first in the world, and had connected administrative divisions in 31 provinces. China’s “eight vertical eight horizontal” railway plan estimated that the HSR network would reach 45,000 km and connect 250 cities by 2030 (China’s “eight vertical eight horizontal” railway plan was proposed in the medium and long term railway network plan of China published in 2016). This plan will come true in the near future. Evidence from Asia and Europe shows that the large-scale HSR cross-regional transportation network has shortened the space-time distance between regions, improved city accessibility and inter-regional and inter-city population, information, and technology mobility [2,3,4], and made labor migration easier by decreasing long-distance information decay and market adjustment costs [5]. Furthermore, the reduction in transport costs and the increase in information and labor flows has benefited local economies and spurred urbanization growth [6], which in turn has resulted in greater rural labor force migration to the cities.

China’s HSR provides 2-h travel time services to approximately 74% of the population [7], which has increased China’s internal migration, especially short-term (one or two years) rural-to-urban migration [8,9]. Because of this decrease in the agricultural labor force, significant village cropland has been abandoned, which has resulted in rural shrinkage [10,11,12].

Since the beginning of the 20th century, 385–472 million km² of global cropland has been abandoned [11,13,14,15], which has brought new challenges to food security [16,17,18]. Because cropland preservation is essential to guaranteeing food security [19,20], there have been many agricultural economics, human geography, and land management studies exploring the driving factors associated with cropland abandonment, such as environmental conditions [21], farm stability and viability [22], the rural labor force [23], agricultural production costs [22], and agricultural technology [24]. However, few studies have examined the effects of transport infrastructure on cropland abandonment.

The rapid construction of HSR in China has indeed promoted the development of the regional economy, but it has led to changes in land uses, which may have negative impacts. Therefore, these adverse impacts need to be fully considered when planning the construction of HSR, so that the construction of HSR can promote regional development more comprehensively. This study analyzed the causal effects of HSR projects on cropland abandonment to deepen the understanding of China’s transportation improvement effects on land use. By revealing the heterogeneous impacts of HSR projects in the different regions, it is hoped that the results of this study can promote rural revitalization and sustainability.

The remainder of this article is organized as follows. Section 2 details the theoretical analytical framework and discusses the possible HSR accessibility effect mechanisms on cropland abandonment, Section 3 describes the data and the research methods and variables used in the empirical analyses, Section 4 presents and discusses the results, and Section 5 provides the research conclusions and some policy implications.

2. Conceptual Background

The HSR accessibility effect mechanisms affecting cropland abandonment are complex. First, the HSR has optimized accessibility between cities and reduced the cost of long-distance travel between cities [2,25], which has increased population migration [26], especially rural-to-urban population migration [9], which has contributed to the reduction in rural labor and led to significant cropland abandonment [12,27,28], that is, the HSR developments have had a possible “push” effect. Second, HSR transport infrastructure has enhanced economic development, regional consumption growth [29,30,31], and urbanization, which in turn has motivated a labor force shift to non-agricultura off-farm employment and cropland abandonment [22,32,33], that is, it has had a possible “pull” effect. This paper focused on these “push” and “pull” effects to assess the effect of the HSR on cropland abandonment. Figure 1 shows the HSR project impact mechanism on farm household cropland abandonment.

As shown in Figure 1, the HSR “push” effect improves urban accessibility by decreasing transport costs, shortening intercity and long-haul travel times [34], and improving regional transportation network efficiency [35,36,37]. The increased speed of the HSR can be converted into idle time and activity space [38], which is called the ‘space-time compression’ theory [39].The time and space compression reduces the time and space cost for rural workers to go to cities to obtain jobs. The decrease in transport costs has improved urban accessibility, making labor migration more frequent [5], that is, the opening of HSR stations directly impacts the abandonment of local agricultural activities and cropland because the rural labor force can more easily seek better urban employment [12,17,40].The greater financial support the migrant workers provide reduces the need for other family members to continue farming [41].

As shown in Figure 1, the “pull” effect of the HSR projects improves non-agricultural industrial development and promotes urbanization in HSR-linked regions. Specifically, urbanization is the process of transforming rural agricultural communities into urban communities based on industry [42] by occupying farmland for urban development or encouraging rural workers to abandon their farmland to seek better non-agricultural opportunities [43,44]. As the HSR interconnected regional network structure promotes logistics, there are also additional opportunities for industrial development in small or medium-sized cities, which again encourages increased urbanization [45]. Therefore, as the HSR network has led to significant changes in regional industrial and factor input structures, there has been a commensurate increase in the labor and service markets in HSR connected cities [46], which has motivated rural workers to abandon cropland and seek better lives in non-agricultural industries [47]. Because of the constraints in the existing Chinese land system, the agricultural land property rights trading market is not perfect, which has also led to out-migration by rural workers to non-agricultural industries [48,49].

In summary, as the global food security situation becomes increasingly tense, especially in developing countries, reducing cropland abandonment is vital; however, as rural labor force availability falls, cropland abandonment has been increasing [12,43]. Several studies have examined the relationship between food security and transport infrastructure [50], with most finding that transport infrastructure had a positive effect on food security [51,52].

While previous studies have analyzed the impact of the HSR on population mobility and urbanization, few have comprehensively considered the HSR impacts on abandoned farmland. Many studies have only focused on the effects of HSR on urban land use [1,53] but not on the effects on rural land use. Therefore, this study considered the cropland abandonment associated with HSR projects and explored the heterogeneous HSR effects on cropland abandonment. Furthermore, as HSR has been found to affect cropland abandonment through labor force out-migration and urbanization, the empirical analyses also examined these potential mechanisms.

3. Data, Variables, and Method

3.1. Data

3.1.1. High-Speed Rail in China

HSR is a railway system that operates at speeds faster than 250 km/h (according to the definition presented by the National Railway Administration of the People’s Republic of China). The first HSR went into operation in China in 2008, and by the end of 2020, had a 38,000-km network, which accounted for two-thirds of the world’s HSR networks [54]. HSR infrastructure construction has strengthened social and economic activities between connected cities [55]. This study collected 2012 to 2014 HSR operations data (including site and opening time) from the National Railway Timetable released by the Ministry of Railways of the People’s Republic of China. The HSR projects were treated as a quasi-natural experiment in this paper, with the linked HSR network cities being the treated regions, and the non-HSR-linked cities being the control regions. Because the household data collected in 2014 and 2016 reflected the farmer household situations in 2013 and 2015, 2012 and 2014 HSR data in the prefecture-level cities were collected. Figure 2 shows China’s HSR network and the cities that had HSR services at the end of 2014.

3.1.2. Household Labor Transfer and Cropland Abandonment

To estimate the HSR impact on cropland abandonment, farm household data were collected from the China Labor-force Dynamics Survey (CLDS) organized by the Center for Social Science Survey at Sun Yat-sen University, which focused on household social-economic status and changes in China, such as urbanization, labor force migration, and land use. To ensure that the sample was representative, the CLDS collected data in 2014 and 2016 from 14,000 households in approximately 400 villages in 29 provinces or municipalities directly under the central government. The CLDS used a multistage cluster, stratified, probability proportional to size (PPS) sampling method, and was the most recent continuous data available for this study. As this study was focused on rural cropland abandonment, urban family samples were not considered. After data cleaning, 4880 valid household questionnaire data from 27 provinces in 2014 and 2016 were used in the analysis.

3.2. Variables

3.2.1. Dependent Variable

The core dependent variable was the total abandoned farm household cropland area in 2013 or 2015 [12,23], which was defined as cropland plots in which no crops were cultivated throughout an entire year. To cater to the differences in land size in the studied regions, the cropland abandonment ratio was used to test the robustness of the benchmark estimate.

3.2.2. Independent Variable

The focus dummy variable was whether the city in which the households lived had been linked to the HSR network or not [56]; when the city had been linked to the HSR network, HSR = 1; otherwise, HSR = 0.

3.2.3. Control Variables

Consistent with the existing studies, the main control variables considered to influence cropland abandonment were householder, household-level, and village-level characteristics [12,23,57]. The householder characteristics included householder age [17] and education level [58], the household characteristics included off-farm employment ratio [23], cropland area [59], total rural household members [60], and the number of family members over 64 years old [61], and the village-level characteristics included soil quality, and water availability [10,62]. The variables are detailed in Table 1.

3.3. Methods

High-speed railway construction is part of the national strategic plan and is not regionally controlled. Therefore, by comparing the abandoned cropland area before and after the cities became connected to the HSR network, the impact of HSR accessibility was regarded as a quasi-natural experiment. Because of the differences in the railway foundations and natural conditions, the HSR accessibility completion was different between the cities; therefore, a PSM-DID analysis method was employed to estimate the HSR impact on cropland abandonment. The DID method is widely acknowledged as the best method to study quasi-natural experiments, or to evaluate the influences of external shocks. The DID method requires a completely random selection between the treatment group and control group, otherwise the result can be largely biased. To relieve the endogenous problems caused by the selection bias, we adopt the propensity score matching (PSM) method before applying DID, so as to ensure the accurate estimation of the HSR’s influence.

3.3.1. Difference in Differences Method

As the main objective of this study was to investigate the effects of HSR accessibility on cropland abandonment, difference-in-difference (DID) models were used to estimate the periods and the treatment group’s effects from the household panel data. The measured dependent variable was the rural household abandoned cropland area, with the treatment variable being the households linked to the HSR network.

To allow for an examination of the time-invariant unobserved heterogeneity and choices by changing the unbiased outcomes, the DID method was employed to estimate the policy effect by separating the time effect from the policy-treated implications [63,64]. To distinguish the HSR impact, the households living in towns with a linked HSR network were the treated group and the households in towns not linked to the HSR network were the control group. Therefore, all household samples were divided into four different subsamples, including the treated group households linked to the HSR network in 2013; the treated group households linked to the HSR network in 2014; the control group households not linked to the HSR network in 2013; and the control group households not linked to the HSR network in 2014. The significant HSR impact on the cropland abandonment was analyzed based on the following DID model:

(1)${\mathrm{Y}}_{\mathrm{it}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{HSR}}_{\mathrm{it}}\ast {\mathrm{After}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$

where ${\mathrm{Y}}_{\mathrm{it}}$ was the size of rural household cropland abandonment, in which subscript i represented farm household i and subscript t represented year t, ${\mathrm{HSR}}_{\mathrm{it}}$ denoted whether or not the households were linked to the HSR network (HSR = 1 meant the city was linked to the HSR network; otherwise, HSR = 0), ${\mathrm{After}}_{\mathrm{it}}$ = 1 indicated 2016 (after the city was linked to the HSR network), ${\mathrm{After}}_{\mathrm{it}}$ = 0 indicated 2014, ${\mathsf{\alpha }}_1$ signified the HSR accessibility effect on the rural household abandoned cropland areas; ${\mathsf{\epsilon }}_{\mathrm{it}}$ was the residual.

3.3.2. PSM-DID Methods

An important underlying assumption of the DID method was that if the external effect (HSR connection) did not exist, the development trends for the dependent variable (cropland abandonment) in the treatment and control groups would be parallel, that is, there would be no apparent systematic differences in the cropland abandonment trends between the households linked to the HSR network and the other households. However, as the HSR was first built in provincial capital cities and cities that had railway foundations, the probability of the HSR network being in these cities was significantly higher. As only two panel data periods were consulted, it was difficult to ascertain the common trend hypothesis. To avoid any estimation biases resulting from the above assumptions, the PSM-DID method was applied [65,66].

The PSM method can solve matching problems by placing the pre-treatment features of any subject into a single index variable and matching the treatment groups to the control groups based on the propensity scores. Therefore, the PSM method was used before the DID method to allow for the selection of the appropriate control groups from the cities without a HSR connection, which helped solve the possible endogenous problems and ensured that the DID estimation results were unbiased [67,68]. The probability of assigning treatment depended on the following PSM model preprocessing variables:

(2)$\mathrm{p}\left(\mathrm{X}\right)=\mathrm{Pr} [\mathrm{D}=1|\mathrm{X}]$

where $\mathrm{D}$ represents all the cities including both treatment and control groups; X represents the control variables.

While the PSM method can address sample selection bias problems, it cannot solve the endogeneity problems caused by missing variables; however, the DID model is unable to explain the sample deviation problem but can determine the endogeneity problem and the policy-treatment effect. Therefore, the following PSM-DID model (Equation (3) was the baseline model that was used to accurately evaluate the HSR accessibility impact:

(3)${\mathrm{Y}}_{\mathrm{it}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{HSR}}_{\mathrm{it}}\ast {\mathrm{After}}_{\mathrm{it}}+{\mathsf{\beta }}_1{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$

where ${\mathrm{X}}_{\mathrm{it}}$ was the control variable set affecting the cropland abandonment factor.

4. Results

4.1. Summary Statistics

The summary statistics for the key variables are shown in Table 2. In 2014, most sample areas were not linked to the HSR network; however, in 2016, more than half of the sample areas were linked. The broad coverage of the HSR-linked cities also allowed for the use of the DID model. It was found that 10.3% of the rural households in the total sample had abandoned their cropland, with the average abandonment size being 0.42 mu in 2014 and 0.38 mu in 2016. In 2014, the average cropland abandonment area in the HSR accessible samples was 0.56 mu, which was over twice as high as in the HSR inaccessible samples (0.24 mu). In 2016, the average cropland abandonment area in the HSR accessible samples was also higher than in the HSR inaccessible samples. For the control variables, the average age for the total sample was 54 years, and only about 10% of householders had a high school education and above. In both 2014 and 2016, the average number of household members in the HSR accessible samples was lower than in the HSR inaccessible samples, the average share of off-farm employment labor in the total labor in the HSR accessible samples was lower than in the HSR inaccessible samples in all window periods, and the degree of urbanization in the HSR accessible areas was 13%, higher than in the HSR inaccessible regions in 2014; however, in 2016, the situation was reversed.

4.2. PSM Results

First, the PSM method was used to process the variables that affected cropland abandonment. To test the reliability of the matching results, a logit model was employed to estimate a set of control variables. The nearest neighbor matching method was used, for which the caliper range was set to 0.05 to match the samples, after which a matching check was conducted to determine the covariate distributions between the processing group and the control group. The results are shown in Table 3.

Table 3 shows that after the propensity score matching, the standardized deviation for most covariates was reduced, the bias values of the matched samples were significantly less than 10%, and the p-values were all greater than 0.1, which indicated that the matched treatment and control groups had no systematic differences and satisfied the balance test. As can be observed in Figure 3, there were few out-of-support untreated samples and most observed values for the samples were supported. Therefore, it was appropriate to use the PSM method because the results were reliable. After deleting the samples that were off the common support domain, the DID model was then applied to precisely estimate the HSR impact.

4.3. DID Model Results

The estimated OLS, DID and PSM-DID cropland abandonment results for the HSR accessibility sample are shown in Table 4. Columns (1)–(6) show the comparative benchmark analysis for the different approaches with no control variables and when the individual, household, and village characteristics were respectively added. The results in column (1) show the correlations between HSR accessibility and cropland abandonment using the OLS method after controlling for the individual, household, and village characteristics. While column (2) shows a similar result for the effect of the HSR accessibility on cropland abandonment using the OLS approach, no significant correlation between HSR accessibility and cropland abandonment size was observed. The results in columns (2) and (3), which respectively show the correlations between HSR accessibility and cropland abandonment with and without the household and village characteristic controls, indicate that the HSR accessibility impact on cropland abandonment was significantly positive. Columns (5) and (6) show the PSM-DID method results, and the benchmark estimate for Equation (3) is shown in column (6). With a coefficient of 0.206, the benchmark regression results indicated that HSR accessibility had a significant effect on cropland abandonment size and was statistically significant at the 5% level, that is, compared with the cropland abandonment size of the HSR inaccessible sample, the cropland abandonment size in the HSR accessible sample was higher by about 20.6%, on average. The results in columns (5) and (6) show that the estimated results remained robust, regardless of whether or not there were control variables. Compared with the OLS and DID method results, column (6) indicates that the coefficient for HSR*After had a slight increase, which proved that the sample selection bias was reduced after the treatment group and control group were matched. All the above results suggested that HSR accessibility increased cropland abandonment likelihood and size.

The results in column (6) in Table 4 also show that some control variables had a significant impact on cropland abandonment. Specifically, total rural household members and cropland abandonment were closely related, which was consistent with the conclusions in previous research on the effects of the household labor force on cropland abandonment [12,28]. The number of family household members indicates the richness of the family labor force to a certain extent, with fewer family members indicating that the family does not have enough labor to farm effectively [69]. It was also found that the household off-farm work ratio affected the abandoned cropland area, which was also consistent with previous research [23]. The village-level characteristics, soil quality and urbanization, were found to be negative and significant, suggesting that improving soil quality could reduce cropland abandonment [70,71], that is, high-quality cropland was less likely to be abandoned than low-quality cropland.

4.4. Robustness Test

To further verify the positive effect of HSR accessibility on cropland abandonment, a robustness test was conducted. First, the robustness of the results was tested by using different dependent variables. However, even when the dependent variable measuring methods were changed, the cropland abandoned ratio was also an essential indicator for land quantity protection. Therefore, the cropland abandoned ratio was used to replace the cropland abandoned area [56]. The results in column (1) in Table 5 show that the regression coefficients for HSR*After were still significantly positive, which suggested that HSR accessibility had a significant positive effect on cropland abandonment.

Another concern was the first-order nearest-neighbor matching that was used in the PSM method; therefore, to test the robustness of the nearest neighbor matching DID results, kernel and radius matching were also used to process the experimental and control groups, after which a DID analysis was conducted. As columns (2) and (3) in Table 5 show, regardless of which matching method was used, the HSR*After regression coefficients remained significantly positive, that is, the regression results still showed that HSR accessibility significantly promoted cropland abandonment, which confirmed that the PSM-DID estimation was robust.

4.5. Placebo Test

To further verify whether there were any unobservable factor biases, a virtual HSR access time was randomly assigned to households to conduct the placebo test, which was equivalent to constructing a “false” HSR accessibility variable. Then, the benchmark model was repeatedly estimated 500 times using the “false” HSR accessibility variable and the estimated coefficients stored.

Figure 4 shows the empirical cumulative distribution function and density for the estimated HSR accessibility coefficients. The estimated coefficients for the HSR accessibility placebo variable were distributed around zero. However, the real estimated value of the benchmark was 0.206, as shown in column (6) in Table 4. The benchmark HSR accessibility estimate coefficients clearly lay outside the range of the coefficients estimated for the virtual HSR accessibility. This placebo test method has been widely used in DID processing studies [72,73].

4.6. Regional Heterogeneity Analysis

Because of China’s regional heterogeneity, the HSR accessibility effect on cropland abandonment could differ because of regional topographic and economic development differences [74,75]. For example, because towns and cities on a plain have relatively good locational and topographical conditions, the cost of building the HSR infrastructure is lower, the land mechanization rates relatively higher, the land scale operations easier, and the cropland abandonment probability lower [23]. Therefore, the topographic effects of HSR accessibility on cropland abandonment were examined.

As Table 6 shows, in hilly and plain areas, the impact of HSR accessibility on cropland abandonment was significantly positive; however, in mountainous regions, the HSR accessibility coefficients were not significant. This could have been because HSR construction and operations in mountainous areas are costly, which means that many cities in mountainous areas are not linked to the HSR network and HSR accessibility was not the main influencing factor for cropland abandonment. Higher HSR infrastructure building costs in mountainous regions also mean higher HSR ticket prices. The marked development of low-cost air and coach services over the past decades suggests that not everyone can afford HSR travel [76]. Therefore, the high HSR ticket price also meant that HSR accessibility was not the main factor for the population migration in these areas [77] and was not the main reason for cropland abandonment. However, because the HSR construction and operating costs in hilly and plain areas are lower than in mountainous areas, HSR accessibility is higher and labor force mobility easier, which could increase the probability of cropland abandonment.

The coefficient estimates in the hilly areas were substantially higher than in the plain areas. Compared with hilly areas, plain areas have flatter terrain and higher mechanization utilization rates, which could weaken the impact of labor migration and cropland abandonment [23] and the HSR accessibility effects.

4.7. Mechanism Analysis

The benchmark model results indicated that HSR accessibility influenced cropland abandonment and the out-migration of the rural labor force, that is, there was a possible “push” effect. Conversely, it is possible that HSR accessibility affected the probability of off-farm employment and there was a “pull” effect. These potential mechanisms were, therefore, examined.

First, the “push” effect mechanism that HSR accessibility affected the out-migration of the rural labor force was explored. Labor force migration is closely related to cropland abandonment, as the out-migration of the rural labor force decreases the number of people available for farm work and could result in households having insufficient labor to look after the farm. Therefore, the number of people working outside the rural household was evaluated to determine the effects of the out-migration on the rural labor force. The regression results are summarized in column (1) in Table 7. It was found that HSR accessibility decreased the rural labor force out-migration. This result may have been because HSR projects promote local economic development and urbanization, which provide more local employment opportunities.

Then, the “pull” effect mechanism on off-farm employment was examined. To determine whether this HSR accessibility effect mechanism improved cropland abandonment, the effects of HSR accessibility on off-farm employment were explored. Off-farm employment was defined as the share of off-farm employment labor in total labor. If HSR accessibility promoted local economic development and urbanization, then rural laborers were more likely to find non-agricultural jobs [28]. The results are shown in column (2) of Table 7. It was found that HSR accessibility significantly affected the share of off-farm employment labor in total labor, with all estimated coefficients being positive, that is, HSR accessibility promotes off-farm employment. Previous research has also shown that HSR accessibility can speed up the urbanization process, which in turn results in a labor shift to the non-agricultural sector [32].

Therefore, the results confirmed that HSR projects promoted the abandonment of croplands by increasing off-farm employment, which supported the “pull effect” hypothesis for non-farm sector employment.

5. Conclusions and Policy Implications

5.1. Conclusions

HSR promotes population migration and leads to economic and urbanization growth. Because HSR accessibility increases the ability of the labor force to move between industries, it has a positive effect on cropland abandonment. Taking HSR accessibility as a quasi-natural experiment, the DID method was used to investigate the effects of HSR accessibility on cropland abandonment. To avoid the traditional limitations in DID models, a PSM approach was first applied to ensure the DID estimation results would be unbiased. Using CLDS data from waves 2014 and 2016, the HSR accessibility cropland abandonment coefficient was estimated to be 0.205. When the ratio of cropland abandonment was taken as the dependent variable to estimate the HSR accessibility effect, the robustness of the conclusion improved. The heterogeneity analysis found that HSR accessibility in both hilly and plain areas had a significant effect on cropland abandonment. Several other potential variables were also examined to understand the underlying mechanisms for the impact of HSR accessibility on cropland abandonment. It was found that HSR accessibility decreased the out-migration of the rural labor force and that HSR accessibility promoted off-farm employment. The mechanistic analysis indicated that HSR accessibility promoted local economic development and urbanization growth, which resulted in local farm labor moving to the non-agricultural sector and abandoning their cropland.

5.2. Policy Implications

This study quantified the effects of HSR accessibility on cropland abandonment. Based on the above findings, some policy implications are proposed. Safeguarding food security remains a significant challenge in developing countries. While the development of the HSR network has increased production factor flows, which is important to China’s urbanization and industrialization goals, it has also resulted in a large-scale loss of agricultural labor forces to non-agricultural employment, which in turn has resulted in cropland abandonment. Therefore, local and regional governments need to consider the potential rural hollowing and cropland abandonment risks of HSR projects when formulating relevant policies to improve transportation facilities. Furthermore, as China’s rural labor force is continuing to decrease, labor substitution technologies, such as mechanization and no-tillage technologies, should be gradually employed in HSR-accessible areas. Local governments could also formulate relevant support policies to encourage land leasing to balance the withdrawal of agricultural personnel and cropland abandonment in HSR-accessible areas. For example, this could include building an efficient information platform for the cropland leasing market, or providing corresponding financial subsidies for infrastructure improvement to the cropland leaseholders. Encouraging non-farm personnel to contract land and invest in characteristic agriculture in these HSR-accessible areas could also inhibit cropland abandonment. For example, this could include providing corresponding financial subsidies to characteristic agricultural operators to help them improve agricultural infrastructure, etc.

Author Contributions

J.S. wrote the manuscript text and processed the data; F.W. provided the methodology and useful suggestions. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures and Tables

Figure 1. HSR project impact mechanism on cropland abandonment.

Figure 2. HSR lines and stations in the prefecture-level cities.

Figure 3. PSM validity test.

Figure 4. Land HSR and cropland abandonment: placebo test.

Table 1

Model variables.

Variable	Definition and Assignment
Abandonment	Abandoned rural household cropland area (mu)
Abandonment Ratio	Share of abandoned cropland in total cropland (%)
Age	Age of the household head in the survey year
Education	Does the household head have a high school diploma or above (0 = no; 1 = yes)
Family	Total number of rural family members (number)
Elder	Number of household members over 64 years old (number)
Fixed assets	Current market value of per capita fixed assets owned by households (104 CNY/person)
Soil quality	Whether the quality of soil in the village is good (0 = no; 1= yes)
Water availability	Whether there is an irrigation facility in the village (0 = no; 1 = yes)
Urbanization	Share of urban households in total households with the same county in the sample (%)
Land size	Land area the rural households managed (mu)
Off-farm work	Share of off-farm employment labor in total labor (%)
Out-migration	Number of migrant workers in the household (number)

Table 2

Summary statistics for the crucial variables.

Variable	2014				2016
	HSR		Non-HSR		HSR		Non-HSR
	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.
Abandonment	0.56	1.25	0.24	2.30	0.39	2.52	0.37	1.33
Abandonment Ratio	0.06	0.23	0.10	0.23	0.06	0.21	0.05	0.19
Age	54.96	12.63	53.79	12.69	55.83	12.53	53.77	12.64
Education	0.10	0.29	0.10	0.29	0.12	0.32	0.10	0.30
Family	4.28	2.07	4.85	2.14	6.57	3.87	6.91	3.79
Elder	0.13	0.41	0.18	0.47	0.15	0.44	0.17	0.45
Fixed assets	4.99	15.30	2.86	4.27	3.33	12.42	2.12	5.01
Soil quality	0.91	0.29	0.92	0.27	0.91	0.29	0.91	0.29
Water availability	0.52	0.50	0.54	0.50	0.56	0.50	0.52	0.50
Urbanization	0.13	0.20	0.12	0.20	0.08	0.19	0.09	0.18
Land size	5.65	10.62	5.60	10.65	5.88	25.79	8.31	94.63
Off-farm work	0.70	0.34	0.66	0.33	0.74	0.32	0.64	0.34
Out-migration	0.77	1.30	1.10	1.46	0.69	1.12	0.99	1.31
Number of observations	2005		2875		2882		1998

Table 3

PSM validity test.

Variable	Unmatched/Matched	Mean		%Bias	%Reduct|Bias|	t-Test
		Treated	Control			t	p > t
Age	U	55.71	53.97	14.40		4.99	0.00
	M	55.74	56.39	−5.40	62.40	−1.88	0.16
Education	U	0.11	0.09	5.10		1.77	0.08
	M	0.10	0.11	−3.00	41.40	−0.98	0.33
Family	U	5.55	5.49	1.90		0.65	0.52
	M	5.55	5.62	−2.20	−19.90	−0.76	0.45
Elder	U	0.17	0.18	−2.40		−0.83	0.41
	M	0.17	0.18	−1.50	36.50	−0.53	0.60
Land rent-out	U	1.07	0.33	7.90		2.73	0.01
	M	0.73	1.00	−3.00	62.30	−1.06	0.29
Fixed assets	U	3.65	2.65	9.60		3.33	0.00
	M	3.14	3.0	0.90	90.30	0.52	0.61
Soil quality	U	0.92	0.91	5.00		1.72	0.09
	M	0.92	0.93	−3.30	32.80	−1.23	0.22
Water availability	U	0.56	0.55	3.70		1.27	0.20
	M	0.56	0.57	−1.30	65.40	−0.44	0.66
Urbanization	U	0.07	0.12	−25.00		−8.66	0.00
	M	0.07	0.08	−2.20	91.00	−0.85	0.40
Land size	U	7.43	5.63	8.60		2.99	0.00
	M	6.70	5.97	3.50	59.70	2.87	0.56
Off-farm work	U	0.66	0.64	6.50		2.27	0.02
	M	0.66	0.65	1.00	85.20	0.33	0.74
Out-migration	U	0.72	1.12	−29.80		−10.32	0.00
	M	0.72	0.70	1.30	95.50	0.55	0.59

Table 4

OLS, DID, and PSM-DID estimation results for the HSR effect on cropland abandonment.

	(1)OLS	(2)OLS	(3)DID	(4)DID	(5)PSM-DID	(6)PSM-DID
HSR	0.028	−0.016
	(0.103)	(0.109)
HSR*After			0.132 **	0.183 **	0.162 **	0.206 **
			(0.061)	(0.075)	(0.075)	(0.096)
Age		0.001		−0.000		−0.001
		(0.006)		(0.006)		(0.008)
Education		−0.021		−0.035		0.064
		(0.172)		(0.171)		(0.230)
Family		−0.029		−0.058 ***		−0.067 **
		(0.019)		(0.021)		(0.027)
Elder		0.207 **		0.216 **		0.155
		(0.100)		(0.100)		(0.127)
Land rent-out		0.004		0.004		0.002
		(0.004)		(0.004)		(0.009)
Fixed assets		−0.005		−0.004		−0.026
		(0.005)		(0.005)		(0.017)
Soil quality		−0.155		−0.157		−0.268 *
		(0.123)		(0.123)		(0.155)
Water availability		−0.061		−0.065		−0.073
		(0.082)		(0.082)		(0.101)
Urbanization		−1.291 ***		−0.944 **		−1.629 **
		(0.474)		(0.470)		(0.661)
Cropland size		−0.004 **		−0.004 **		0.004
		(0.002)		(0.002)		(0.005)
Off-farm work		0.329 **		0.325 **		0.314 *
		(0.140)		(0.140)		(0.170)
Out-migration		−0.003		−0.009		−0.005
		(0.029)		(0.029)		(0.036)
Constant	0.398 ***	0.637 *	0.380 ***	0.767 **	0.424 ***	1.124 **
	(0.055)	(0.374)	(0.025)	(0.377)	(0.029)	(0.512)
Wave FEs	Y	Y	Y	Y	Y	Y
City FEs	Y	Y	Y	Y	Y	Y
Observations	7668	5907	7668	5907	4819	4819
R-squared	0.000	0.015	0.002	0.018	0.004	0.023
Number of no	4828	4219	4828	4219	3639	3639

Remarks: standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 5

Robustness checks for different indicators and the matching method.

	(1)Abandonment Ratio	(2)Radius Matching	(3)Kernel Matching
HSR*After	0.030 **	0.209 **	0.207 **
	(0.015)	(0.096)	(0.092)
Constant	0.114	1.067 **	1.081 **
	(0.082)	(0.511)	(0.503)
Control variables	Y	Y	Y
Wave FEs	Y	Y	Y
City FEs	Y	Y	Y
Observations	4796	4803	4806
R-squared	0.037	0.023	0.023
Number of no	3628	3629	3629

Remarks: standard errors in parentheses; ** p < 0.05.

Table 6

HSR and cropland abandonment: regional heterogeneity.

	(1)Mountain	(2)Hill	(3)Plain
HSR*After	0.363	0.355 **	0.128 *
	(0.288)	(0.184)	(0.081)
Constant	0.902	0.721	0.667
	(0.986)	(0.845)	(0.451)
Control variables	Y	Y	Y
Wave FEs	Y	Y	Y
City FEs	Y	Y	Y
Observations	1317	1684	2906
Number of no	984	1236	1999
R-squared	0.139	0.047	0.028

Remarks: standard errors in parentheses, ** p < 0.05, * p < 0.1.

Table 7

Mechanism analysis: impact of the HSR on rural labor force out-migration and off-farm employment.

	(1)Out-Migration	(2)Off-Farm Employment
HSR*After	−0.226 ***	0.019 ***
	(−0.049)	(0.010)
Constant	0.954 ***	0.682 ***
	(0.237)	(0.046)
Control variables	Y	Y
Wave FEs	Y	Y
City FEs	Y	Y
Observations	9760	9760
R-squared	0.009	0.002
Number of no	4880	4880

Remarks: standard errors in parentheses; *** p < 0.01.

References

1. Huang, Z.; Du, X. How does high-speed rail affect land value? Evidence from China. Land Use Policy; 2021; 101, 105068. [DOI: https://dx.doi.org/10.1016/j.landusepol.2020.105068]

2. Wang, L.; Liu, Y.; Sun, C.; Liu, Y. Accessibility impact of the present and future high-speed rail network: A case study of Jiangsu Province, China. J. Transp. Geogr.; 2016; 54, pp. 161-172. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2016.05.015]

3. Shao, S.; Tian, Z.; Yang, L. High speed rail and urban service industry agglomeration: Evidence from China’s Yangtze River Delta region. J. Transp. Geogr.; 2017; 64, pp. 174-183. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2017.08.019]

4. Ravazzoli, E.; Streifeneder, T.; Cavallaro, F. The Effects of the Planned High-Speed Rail System on Travel Times and Spatial Development in the European Alps. Mt. Res. Dev.; 2017; 37, pp. 131-140. [DOI: https://dx.doi.org/10.1659/MRD-JOURNAL-D-15-00051.1]

5. Haynes, K.E. Labor markets and regional transportation improvements: The case of high-speed trains. Ann. Reg. Sci.; 1997; 31, pp. 57-76. [DOI: https://dx.doi.org/10.1007/s001680050039]

6. Motamed, M.J.; Florax, R.J.G.M.; Masters, W.A. Agriculture, transportation and the timing of urbanization: Global analysis at the grid cell level. J. Econ. Growth; 2014; 19, pp. 339-368. [DOI: https://dx.doi.org/10.1007/s10887-014-9104-x]

7. Wang, J.; Jiao, J.; Du, C.; Hu, H. Competition of spatial service hinterlands between high-speed rail and air transport in China: Present and future trends. J. Geogr. Sci.; 2015; 25, pp. 1137-1152. [DOI: https://dx.doi.org/10.1007/s11442-015-1224-5]

8. Wang, F.; Wei, X.; Liu, J.; He, L.; Gao, M. Impact of high-speed rail on population mobility and urbanisation: A case study on Yangtze River Delta urban agglomeration, China. Transp. Res. Part A Policy Pract.; 2019; 127, pp. 99-114. [DOI: https://dx.doi.org/10.1016/j.tra.2019.06.018]

9. Mohabir, N.; Jiang, Y.; Ma, R. Chinese floating migrants: Rural-urban migrant labourers’ intentions to stay or return. Habitat Int.; 2017; 60, pp. 101-110. [DOI: https://dx.doi.org/10.1016/j.habitatint.2016.12.008]

10. van Leeuwen, C.C.; Cammeraat, E.L.; de Vente, J.; Boix-Fayos, C. The evolution of soil conservation policies targeting land abandonment and soil erosion in Spain: A review. Land Use Policy; 2019; 83, pp. 174-186. [DOI: https://dx.doi.org/10.1016/j.landusepol.2019.01.018]

11. Li, Y.; Wu, W.; Liu, Y. Land consolidation for rural sustainability in China: Practical reflections and policy implications. Land Use Policy; 2018; 74, pp. 137-141. [DOI: https://dx.doi.org/10.1016/j.landusepol.2017.07.003]

12. Xu, D.; Deng, X.; Huang, K.; Liu, Y.; Yong, Z.; Liu, S. Relationships between labor migration and cropland abandonment in rural China from the perspective of village types. Land Use Policy; 2019; 88, 104164. [DOI: https://dx.doi.org/10.1016/j.landusepol.2019.104164]

13. Campbell, J.E.; Lobell, D.B.; Genova, R.C.; Field, C.B. The Global Potential of Bioenergy on Abandoned Agriculture Lands. Environ. Sci. Technol.; 2008; 42, pp. 5791-5794. [DOI: https://dx.doi.org/10.1021/es800052w]

14. Levers, C.; Schneider, M.; Prishchepov, A.V.; Estel, S.; Kuemmerle, T. Spatial variation in determinants of agricultural land abandonment in Europe. Sci. Total Environ.; 2018; 644, pp. 95-111. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2018.06.326] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29981521]

15. Long, H.; Tu, S.; Ge, D.; Li, T.; Liu, Y. The allocation and management of critical resources in rural China under restructuring: Problems and prospects. J. Rural. Stud.; 2016; 47, pp. 392-412. [DOI: https://dx.doi.org/10.1016/j.jrurstud.2016.03.011]

16. Liu, Y.; Zhou, Y. Reflections on China’s food security and land use policy under rapid urbanization. Land Use Policy; 2021; 109, 105699. [DOI: https://dx.doi.org/10.1016/j.landusepol.2021.105699]

17. Subedi, Y.R.; Kristiansen, P.; Cacho, O. Drivers and consequences of agricultural land abandonment and its reutilisation pathways: A systematic review. Environ. Dev.; 2022; 42, 100681. [DOI: https://dx.doi.org/10.1016/j.envdev.2021.100681]

18. Ye, L.; van Ranst, E. Production scenarios and the effect of soil degradation on long-term food security in China. Glob. Environ. Chang.; 2009; 19, pp. 464-481. [DOI: https://dx.doi.org/10.1016/j.gloenvcha.2009.06.002]

19. Hervas, A. Mapping oil palm-related land use change in Guatemala, 2003–2019: Implications for food security. Land Use Policy; 2021; 109, 105657. [DOI: https://dx.doi.org/10.1016/j.landusepol.2021.105657]

20. Jiang, L.; Guo, S.; Wang, G.; Kan, S.; Jiang, H. Changes in agricultural land requirements for food provision in China 2003–2011: A comparison between urban and rural residents. Sci. Total Environ.; 2020; 725, 138293. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2020.138293]

21. Alonso-Sarría, F.; Martínez-Hernández, C.; Romero-Díaz, A.; Cánovas-García, F.; Gomariz-Castillo, F. Main Environmental Features Leading to Recent Land Abandonment in Murcia Region (Southeast Spain). Land Degrad. Dev.; 2015; 27, pp. 654-670. [DOI: https://dx.doi.org/10.1002/ldr.2447]

22. Lieskovský, J.; Bezák, P.; ŠPulerová, J.; Lieskovský, T.; Koleda, P.; Dobrovodská, M.; Bürgi, M.; Gimmi, U. The abandonment of traditional agricultural landscape in Slovakia—Analysis of extent and driving forces. J. Rural. Stud.; 2015; 37, pp. 75-84. [DOI: https://dx.doi.org/10.1016/j.jrurstud.2014.12.007]

23. Deng, X.; Xu, D.; Qi, Y.; Zeng, M. Labor Off-Farm Employment and Cropland Abandonment in Rural China: Spatial Distribution and Empirical Analysis. Int. J. Environ. Res. Public Health; 2018; 15, 1808. [DOI: https://dx.doi.org/10.3390/ijerph15091808] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30135363]

24. Shively, G.; Martinez, E. Deforestation, irrigation, employment and cautious optimism in southern Palawan, the Philippines. Agric. Technol. Trop. Deforest.; 2001; pp. 335-346. [DOI: https://dx.doi.org/10.1079/9780851994512.0335]

25. Abrate, G.; Viglia, G.; García, J.S.; Forgas-Coll, S. Price Competition within and between Airlines and High-Speed Trains: The Case of the Milan—Rome Route. Tour. Econ.; 2016; 22, pp. 311-323. [DOI: https://dx.doi.org/10.5367/te.2016.0549]

26. Guirao, B.; Lara-Galera, A.; Campa, J.L. High Speed Rail commuting impacts on labour migration: The case of the concentration of metropolis in the Madrid functional area. Land Use Policy; 2017; 66, pp. 131-140. [DOI: https://dx.doi.org/10.1016/j.landusepol.2017.04.035]

27. Shui, Y.; Xu, D.; Liu, S. Research on Role Cognition and Employment Strategy of Rural Female Laborers in Sichuan, China. Sustainability; 2019; 11, 1708. [DOI: https://dx.doi.org/10.3390/su11061708]

28. Zhang, Y.; Li, X.; Song, W. Determinants of cropland abandonment at the parcel, household and village levels in mountain areas of China: A multi-level analysis. Land Use Policy; 2014; 41, pp. 186-192. [DOI: https://dx.doi.org/10.1016/j.landusepol.2014.05.011]

29. Chen, Z.; Xue, J.; Rose, A.Z.; Haynes, K.E. The impact of high-speed rail investment on economic and environmental change in China: A dynamic CGE analysis. Transp. Res. Part A Policy Pract.; 2016; 92, pp. 232-245. [DOI: https://dx.doi.org/10.1016/j.tra.2016.08.006]

30. Cheng, Y.S.; Loo, B.P.; Vickerman, R. High-speed rail networks, economic integration and regional specialisation in China and Europe. Travel. Behav. Soc.; 2015; 2, pp. 1-14. [DOI: https://dx.doi.org/10.1016/j.tbs.2014.07.002]

31. Rodrigue, J.P. The Economics and Politics of High-Speed Rail: Lessons from Experiences Abroad. AAG Rev. Books; 2016; 4, pp. 17-18. [DOI: https://dx.doi.org/10.1080/2325548X.2016.1117342]

32. Emerick, K. Agricultural productivity and the sectoral reallocation of labor in rural India. J. Dev. Econ.; 2018; 135, pp. 488-503. [DOI: https://dx.doi.org/10.1016/j.jdeveco.2018.08.013]

33. Xie, Y.; Jiang, Q. Land arrangements for rural–urban migrant workers in China: Findings from Jiangsu Province. Land Use Policy; 2016; 50, pp. 262-267. [DOI: https://dx.doi.org/10.1016/j.landusepol.2015.10.010]

34. Zhu, P. Does high-speed rail stimulate urban land growth? Experience from China. Transp. Res. Part D Transp. Environ.; 2021; 98, 102974. [DOI: https://dx.doi.org/10.1016/j.trd.2021.102974]

35. Bentlage, M.; Lüthi, S.; Thierstein, A. Knowledge creation in German agglomerations and accessibility—An approach involving non-physical connectivity. Cities; 2013; 30, pp. 47-58. [DOI: https://dx.doi.org/10.1016/j.cities.2012.07.003]

36. Cao, J.; Liu, X.C.; Wang, Y.; Li, Q. Accessibility impacts of China’s high-speed rail network. J. Transp. Geogr.; 2013; 28, pp. 12-21. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2012.10.008]

37. Shaw, S.L.; Fang, Z.; Lu, S.; Tao, R. Impacts of high speed rail on railroad network accessibility in China. J. Transp. Geogr.; 2014; 40, pp. 112-122. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2014.03.010]

38. Spiekermann, K.; Wegener, M. The shrinking continent: New time-space maps of Europe. Environ. Plan. B Plan. Des.; 1994; 21, pp. 653-673. [DOI: https://dx.doi.org/10.1068/b210653]

39. Harvey, D. The Condition of Postmodernity: An Enquiry into the Origins of Cultural Change. Contemp. Sociol.; 1989; 20, [DOI: https://dx.doi.org/10.17323/1726-3247-2020-2-62-75]

40. Maharjan, A.; Kochhar, I.; Chitale, V.S.; Hussain, A.; Gioli, G. Understanding rural outmigration and agricultural land use change in the Gandaki Basin, Nepal. Appl. Geogr.; 2020; 124, 102278. [DOI: https://dx.doi.org/10.1016/j.apgeog.2020.102278]

41. Ojha, H.R.; Shrestha, K.K.; Subedi, Y.R.; Shah, R.; Nuberg, I.; Heyojoo, B.; Cedamon, E.; Rigg, J.; Tamang, S.; Paudel, K.P. et al. Agricultural land underutilisation in the hills of Nepal: Investigating socio-environmental pathways of change. J. Rural. Stud.; 2017; 53, pp. 156-172. [DOI: https://dx.doi.org/10.1016/j.jrurstud.2017.05.012]

42. Ahani, S.; Dadashpoor, H. A review of domains, approaches, methods and indicators in peri-urbanization literature. Habitat Int.; 2021; 114, 102387. [DOI: https://dx.doi.org/10.1016/j.habitatint.2021.102387]

43. Hou, D.; Meng, F.; Prishchepov, A.V. How is urbanization shaping agricultural land-use? Unraveling the nexus between farmland abandonment and urbanization in China. Landsc. Urban Plan.; 2021; 214, 104170. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2021.104170]

44. Rimal, B.; Sharma, R.; Kunwar, R.; Keshtkar, H.; Stork, N.E.; Rijal, S.; Rahman, S.A.; Baral, H. Effects of land use and land cover change on ecosystem services in the Koshi River Basin, Eastern Nepal. Ecosyst. Serv.; 2019; 38, 100963. [DOI: https://dx.doi.org/10.1016/j.ecoser.2019.100963]

45. Verma, A.; Sudhira, H.S.; Rathi, S.; King, R.; Dash, N. Sustainable urbanization using high speed rail (HSR) in Karnataka, India. Res. Transp. Econ.; 2013; 38, pp. 67-77. [DOI: https://dx.doi.org/10.1016/j.retrec.2012.05.013]

46. Blum, U.; Haynes, K.E.; Karlsson, C. Introduction to the special issue the regional and urban effects of high-speed trains. Ann. Reg. Sci.; 1997; 31, pp. 1-20. [DOI: https://dx.doi.org/10.1007/s001680050036]

47. MacDonald, D.; Crabtree, J.; Wiesinger, G.; Dax, T.; Stamou, N.; Fleury, P.; Gutierrez Lazpita, J.; Gibon, A. Agricultural abandonment in mountain areas of Europe: Environmental consequences and policy response. J. Environ. Manag.; 2000; 59, pp. 47-69. [DOI: https://dx.doi.org/10.1006/jema.1999.0335]

48. Cai, F. Has China’s labor mobility exhausted its momentum. Chin. Rural Econ.; 2018; 9, pp. 2-13.

49. Tombe, T.; Zhu, X. Trade, Migration, and Productivity: A Quantitative Analysis of China. Am. Econ. Rev.; 2019; 109, pp. 1843-1872. [DOI: https://dx.doi.org/10.1257/aer.20150811]

50. Blimpo, M.P.; Harding, R.; Wantchekon, L. Public Investment in Rural Infrastructure: Some Political Economy Considerations. J. Afr. Econ.; 2013; 22, (Suppl. S2), pp. ii57-ii83. [DOI: https://dx.doi.org/10.1093/jae/ejt015]

51. Lang, T.; Barling, D. Food security and food sustainability: Reformulating the debate. Geogr. J.; 2012; 178, pp. 313-326. [DOI: https://dx.doi.org/10.1111/j.1475-4959.2012.00480.x]

52. Yacamán Ochoa, C.; Ferrer Jiménez, D.; Mata Olmo, R. Green Infrastructure Planning in Metropolitan Regions to Improve the Connectivity of Agricultural Landscapes and Food Security. Land; 2020; 9, 414. [DOI: https://dx.doi.org/10.3390/land9110414]

53. Niu, F.; Xin, Z.; Sun, D. Urban land use effects of high-speed railway network in China: A spatial spillover perspective. Land Use Policy; 2021; 105, 105417. [DOI: https://dx.doi.org/10.1016/j.landusepol.2021.105417]

54. Guo, Y.; Cao, L.; Song, Y.; Wang, Y.; Li, Y. Understanding the formation of City-HSR network: A case study of Yangtze River Delta, China. Transp. Policy; 2022; 116, pp. 315-326. [DOI: https://dx.doi.org/10.1016/j.tranpol.2021.12.014]

55. Jia, S.; Zhou, C.; Qin, C. No difference in effect of high-speed rail on regional economic growth based on match effect perspective?. Transp. Res. Part A Policy Pract.; 2017; 106, pp. 144-157. [DOI: https://dx.doi.org/10.1016/j.tra.2017.08.011]

56. Chen, F.; Hao, X.; Chen, Z. Can high-speed rail improve health and alleviate health inequality? Evidence from China. Transp. Policy; 2021; 114, pp. 266-279. [DOI: https://dx.doi.org/10.1016/j.tranpol.2021.10.007]

57. Deng, X.; Xu, D.; Zeng, M.; Qi, Y. Does Internet use help reduce rural cropland abandonment? Evidence from China. Land Use Policy; 2019; 89, 104243. [DOI: https://dx.doi.org/10.1016/j.landusepol.2019.104243]

58. Che, Y. Off-farm employments and land rental behavior: Evidence from rural China. China Agric. Econ. Rev.; 2016; 8, pp. 37-54. [DOI: https://dx.doi.org/10.1108/CAER-09-2014-0086]

59. Zheng, L.; Qian, W. The impact of land certification on cropland abandonment: Evidence from rural China. China Agric. Econ. Rev.; 2021; ahead of print [DOI: https://dx.doi.org/10.1108/CAER-12-2020-0292]

60. He, Y.; Xie, H.; Peng, C. Analyzing the behavioural mechanism of farmland abandonment in the hilly mountainous areas in China from the perspective of farming household diversity. Land Use Policy; 2020; 99, 104826. [DOI: https://dx.doi.org/10.1016/j.landusepol.2020.104826]

61. Lee, J.; Oh, Y.G.; Yoo, S.H.; Suh, K. Vulnerability assessment of rural aging community for abandoned farmlands in South Korea. Land Use Policy; 2021; 108, 105544. [DOI: https://dx.doi.org/10.1016/j.landusepol.2021.105544]

62. Díaz, G.I.; Nahuelhual, L.; Echeverría, C.; Marín, S. Drivers of land abandonment in Southern Chile and implications for landscape planning. Landsc. Urban Plan.; 2011; 99, pp. 207-217. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2010.11.005]

63. Meyer, B.D. Natural and Quasi-Experiments in Economics. J. Bus. Econ. Stat.; 1995; 13, pp. 151-161. [DOI: https://dx.doi.org/10.1080/07350015.1995.10524589]

64. Sunak, Y.; Madlener, R. The impact of wind farm visibility on property values: A spatial difference-in-differences analysis. Energy Econ.; 2016; 55, pp. 79-91. [DOI: https://dx.doi.org/10.1016/j.eneco.2015.12.025]

65. Heckman, J.J.; Ichimura, H.; Todd, P. Matching as an Econometric Evaluation Estimator. Rev. Econ. Stud.; 1998; 65, pp. 261-294. [DOI: https://dx.doi.org/10.1111/1467-937X.00044]

66. Heckman, J.J.; Ichimura, H.; Todd, P.E. Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Programme. Rev. Econ. Stud.; 1997; 64, pp. 605-654. [DOI: https://dx.doi.org/10.2307/2971733]

67. Dehejia, R.H.; Wahba, S. Propensity Score-Matching Methods for Nonexperimental Causal Studies. Rev. Econ. Stat.; 2002; 84, pp. 151-161. [DOI: https://dx.doi.org/10.1162/003465302317331982]

68. Locke, C.M.; Butsic, V.; Rissman, A.R. Zoning effects on housing change vary with income, based on a four-decade panel model after propensity score matching. Land Use Policy; 2017; 64, pp. 353-362. [DOI: https://dx.doi.org/10.1016/j.landusepol.2017.03.019]

69. Zimmerer, K.S. Soil erosion and labor shortages in the Andes with special reference to Bolivia, 1953–1991: Implications for “conservation-with-development”. World Dev.; 1993; 21, pp. 1659-1675. [DOI: https://dx.doi.org/10.1016/0305-750X(93)90100-N]

70. Feranec, J.; Jaffrain, G.; Soukup, T.; Hazeu, G. Determining changes and flows in European landscapes 1990–2000 using CORINE land cover data. Appl. Geogr.; 2010; 30, pp. 19-35. [DOI: https://dx.doi.org/10.1016/j.apgeog.2009.07.003]

71. Nadal, J.; Pèlachs, A.; Molina, D.; Soriano, J.M. Soil fertility evolution and landscape dynamics in a Mediterranean area: A case study in the Sant Llorenç Natural Park (Barcelona, NE Spain). Area; 2009; 41, pp. 129-138. [DOI: https://dx.doi.org/10.1111/j.1475-4762.2008.00858.x]

72. Ferrara, E.L.; Chong, A.; Duryea, S. Soap Operas and Fertility: Evidence from Brazil. Am. Econ. J. Appl. Econ.; 2012; 4, pp. 1-31. [DOI: https://dx.doi.org/10.1257/app.4.4.1]

73. Liu, Q.; Lu, Y. Firm investment and exporting: Evidence from China’s value-added tax reform. J. Int. Econ.; 2015; 97, pp. 392-403. [DOI: https://dx.doi.org/10.1016/j.jinteco.2015.07.003]

74. Campos, J.; de Rus, G. Some stylized facts about high-speed rail: A review of HSR experiences around the world. Transp. Policy; 2009; 16, pp. 19-28. [DOI: https://dx.doi.org/10.1016/j.tranpol.2009.02.008]

75. van der Zanden, E.H.; Verburg, P.H.; Schulp, C.J.; Verkerk, P.J. Trade-offs of European agricultural abandonment. Land Use Policy; 2017; 62, pp. 290-301. [DOI: https://dx.doi.org/10.1016/j.landusepol.2017.01.003]

76. Dobruszkes, F.; Chen, C.L.; Moyano, A.; Pagliara, F.; Endemann, P. Is high-speed rail socially exclusive? An evidence-based worldwide analysis. Travel. Behav. Soc.; 2022; 26, pp. 96-107. [DOI: https://dx.doi.org/10.1016/j.tbs.2021.09.009]

77. Garmendia, M.; Ureña, J.; Coronado, J. Long-distance trips in a sparsely populated region: The impact of high-speed infrastructures. J. Transp. Geogr.; 2011; 19, pp. 537-551. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2010.06.002]