Spatio-Temporal Pattern of Interprovincial Migration of Rural Population in China and Its Influencing Factors

Abstract

In the era of significant migration, the rural population has been a major component of China’s internal migration. Utilizing data from the last three population censuses, this study examines the evolution and network characteristics of interprovincial rural migration in China from 2000 to 2020. A panel model is employed to analyze the influencing factors. The findings reveal that the interprovincial migration of China’s rural population has increased rapidly, though the growth rate has decelerated. Provincial rural emigration scales have shown a differentiated trend, with the Heihe-Tengchong Line serving as a boundary, while the overall pattern exhibits a “center-periphery” structure. The primary migration destinations are the eastern coastal provinces. Between 2000 and 2020, China’s interprovincial rural migration network demonstrated strong centralization, but the agglomeration core of the network became more diversified. Spatially, interprovincial rural migration shows a clear preference for proximate destinations, and the interprovincial migration network has gradually shifted from an “umbrella-shaped” pattern to a “funnel-shaped” pattern. Key factors influencing the evolution of the interprovincial rural migration network include the per capita GDP, natural disasters, topographic relief, spatial distance, rural hukou population, and internet penetration rate. To facilitate orderly rural migration and support the major national strategies, it is essential to promote regional coordinated development, deepen rural hukou system reforms, and streamline factor flow mechanisms.

1. Introduction

Since the dawn of mankind, migration has been a crucial strategy for survival and development [1]. From the earliest days of settling near water and fertile land to the current global population migration, the depth and breadth of human migration have expanded with the development of productive forces, and the frequency of migration has also increased [2]. The rapid development of transportation and information technology, particularly since the industrial revolution, has significantly reduced temporal and spatial distances worldwide, facilitating urban–rural, regional and even long-distance transnational migration [3,4]. In recent decades, although public health emergencies such as SARS and COVID-19 have temporarily halted large-scale population migration, migration has quickly resumed as the epidemics have subsided [5,6,7].

Population migration is a “double-edged sword” that profoundly influences the status and trends of regional development. For destination regions, migration enhances labor conditions and promotes the concentration of production factors, such as talent and capital, thereby boosting regional development [8,9]. However, it can also increase the pressure on transportation, housing, and employment while introducing challenges such as racial tensions and social conflicts [10,11]. For source regions, migration alleviates population pressure and human–environment conflicts and strengthens their economic, social, scientific, and cultural ties with the outside world [12]. However, it can also lead to problems such as brain drain, labor shortages, and the erosion of local culture [13,14], especially in rural areas, where large-scale population exodus exacerbates rural decline [15]. Given the important role of population migration in regional development, scholars in demography, economics, statistics, and sociology have conducted extensive research on this issue and developed far-reaching theories and models, including the push-pull theory [16], the gravity model [17], and the labor migration model [18].

While migration is a global phenomenon, its patterns and driving forces exhibit significant variations across countries, particularly in rapidly transforming economies such as China. In China, population migration is closely tied to the hukou system (the household registration system in China) [19]. In the early years of the PRC, the hukou system was established to stabilize domestic public order, among other objectives. Between 1959 and 1978, the hukou system was consolidated by categorizing households as “agricultural” or “non-agricultural”, dividing the population into rural and urban residents. Strict transfer procedures and legal restrictions controlled population migration, confining rural residents to agricultural work and restricting their access to urban areas [20,21,22]. As China’s reform and opening up in 1978 progressed, the huge labor demand generated by the economic development necessitated the reforms to the hukou system to achieve the optimal allocation of production factors such as labor on a larger scale [23,24]. In this context, farmers going out to work or do business were allowed to provide their own food and settle in towns, marking the beginning of population migration. Subsequently, the introduction of a series of policies and measures, combined with rapid industrialization and urbanization, deepened the hukou system reform and triggered a wave of domestic population migration, and the scale of population migration has continued to rise rapidly [25]. The 2020 national population census confirmed China’s historic transition from a low-mobility “rural China” to a high-mobility “migrant China” [26].

China’s massive population migration has reshaped the country’s regional development pattern and drawn widespread attention from academia. Existing research has focused on the historical evolution and spatial patterns of population migration [27,28,29], the factors influencing migration and its driving mechanisms [30,31], socio-economic impacts, and resource-environmental effects [32,33,34]. Additionally, studies have reviewed the evolution of China’s hukou system reform and population migration policies [23,35]. As aging intensifies, an increasing number of studies have begun to focus on the migration of the elderly population to actively address population aging [31,36,37,38]. Overall, guided by classical population migration theories, these studies have explored the spatial and temporal characteristics of population migration and its influencing factors at both macro- and meso-levels, revealed the impact of socio-economic transformations on population migration and proposed countermeasures, and analyzed the role of families and family members as decision-makers in migration at the micro-level.

While migration in China has been widely studied, rural migration remains a critical yet underexplored dimension, particularly in the context of the hukou system and urban–rural disparities. Occupying the lion’s share of population movement [39], rural migration, in the context of the urban–rural dual system, profoundly reflects the urban–rural gap and interaction resulting from the hukou system. It not only promotes cross-regional resource flows and regional system changes but also provides a key perspective for understanding China’s unique social structure. The study of rural migration in China holds unique and significant research importance as it further highlights the complexity of urban–rural interactions and the far-reaching impacts of the institutional context on the basis of traditional population migration studies that do not differentiate between household categorizations.

Existing research on rural migration can be categorized into four areas: (1) the interaction between rural migration and rural areas, focusing on challenges to village governance, rural poverty, rural pensions, and impacts on the land system and economic development [40,41,42]; (2) the interaction between rural migration and urban areas, examining issues such as social integration, pressure on public services, and social stratification [43,44,45]; (3) people-centered rural migrants, encompassing topics such as demographic characteristics and well-being security [46,47]; and (4) rural migration activities. Relevant studies have primarily relied on sample surveys and questionnaires [48,49] and have usually adopted theories such as push and pull factors and proximity effects to analyze the scale, patterns, and determinants of rural migration in specific regions. However, these studies are often limited by their reliance on rural-level data, which constrains their ability to capture broader spatio-temporal patterns and regional variations, thereby limiting the understanding of the macro dynamics of rural migration.

The innovation of this study lies in its dual focus: it not only addresses the long-neglected issue of rural migration in China but also provides a macro-level analysis of its spatio-temporal patterns and network characteristics. Utilizing national census data from 2000, 2010, and 2020, this study explores the dynamics of interprovincial rural migration over the past two decades, filling a critical gap in the literature. Unlike existing studies that rely on limited sample surveys, this research leverages comprehensive national data to reveal the scale, patterns, and driving mechanisms of rural migration. Furthermore, by employing panel models to diagnose influencing factors, this study offers valuable insights for policymakers seeking to promote balanced regional development and sustainable urbanization, thereby supporting the implementation of new-type urbanization and rural revitalization strategies.

2. Materials and Methods

2.1. Social Network Analysis

Social network analysis (SNA), developed by sociologists, is a quantitative method for studying the structure and processes of human society using mathematical models and graph theory [50]. This method helps people understand social phenomena by analyzing the nodes and edges in social networks. With the advent of the big data era, SNA has been widely applied in social science research, including areas such as population migration [51,52], food trade [53,54], and urban structure [55]. These applications demonstrate its effectiveness in analyzing complex relational data, which aligns with the objectives of this study. At its core, relationships form the foundation of social network analysis, and there are various types of relationships between provinces. This study explores the characteristics of the interprovincial rural migration network in China from 2000 to 2020 using measures of network density and centralization. Centralization is typically examined through metrics such as degree centralization, betweenness centralization, and closeness centralization [56,57]. Based on the research design, this study focuses on degree centralization and betweenness centralization, which were calculated using Ucinet 6 [58]. Accordingly, this study analyzes four specific indicators: network density, outdegree centralization, indegree centralization, and betweenness centralization (

Table 1. Definitions and formulas of SNA indicators.

Indicator	Definition	Formula
Network density	Network density characterizes the closeness of links between nodes in a population migration network.	$D={\displaystyle \frac{l}{n\left(n-1\right)}}$	(1)
Outdegree centralization	Degree centralization measures the tendency of a single node to be more central than all other nodes in the network. Outdegree centralization quantifies the extent to which a network’s nodes are dominated by a single node or a small group of nodes in terms of outgoing connections. Similarly, indegree centralization measures the extent to which a network’s nodes are dominated by a single node or a small group of nodes in terms of incoming connections. The higher the outdegree centralization, the lower is the variability in node outdegree centralities; the higher the indegree centralization, the lower is the variability in node indegree centralities.	$C_{i,out}={\displaystyle \sum _{j=1}{a}_{ij,out}}$	(2)
		$C_{out}={\displaystyle \frac{{\displaystyle \sum _{i=1}^n\left(C_{i,out,\max }-C_i\right)}}{n^2-3n+2}}$	(3)
Indegree centralization		$C_{i,in}={\displaystyle \sum _{j=1}{a}_{ij,in}}$	(4)
		$C_{in}={\displaystyle \frac{{\displaystyle \sum _{i=1}^n\left(C_{i,in,\max }-C_i\right)}}{n^2-3n+2}}$	(5)
Betweenness centralization	Betweenness centralization denotes the difference between the node with the highest betweenness centrality and the average in a network. The higher this differential, the more important the intermediary becomes, especially as the network evolves into several smaller sub-networks.	$C_{ABi}={\displaystyle \sum _j^n{\displaystyle \sum _k^n\frac{g_{jk}\left(i\right)}{g_{jk}}}}$	(6)
		$C_B={\displaystyle \frac{{\displaystyle \sum _{i=1}^n\left(C_{AB\max }-C_{ABi}\right)}}{n^3-4n^2+5n-2}}$	(7)

Note: l and n are the number of flows and nodes in the population migration network, respectively. C_i,out and C_i,in denote the outdegree and indegree centrality of node i, respectively, and a_ij,out and a_ij,in refer to the size of the population migrating from node i to node j, and from node j to node i, respectively. C_ABi is the absolute betweenness centrality of node i, g_jk denotes the number of shortest paths between nodes j and k, g_jk(i) indicates that node i is precisely on this shortcut, and the ratio between the two values indicates that node i exerts control over the link between nodes j and k. C_out and C_in denote the outdegree and indegree centralization of the network, respectively. C_i,out,max and C_i,in,max refer to the maximum value of C_i,out and C_i,in, respectively. C_B refers to the betweenness centralization of the network, and C_ABmax is the maximum value of C_ABi.

).

2.2. Model Construction

Push-pull theory, originating from Ravenstein’s ‘Laws of Migration’ [59], is one of the classic theories of population migration. Building on this framework, Lee [16], an American demographer, further developed and expanded the theory of population migration by examining the dimensions of “migration factors”, “migration scale”, “migration flow”, and “migrant characteristics”. He pointed out that population migration results from a combination of pull and push factors between the destination and the source, which are also influenced by the present obstacles. Specifically, the pull refers to factors that enhance production and living conditions at the destination, while the push refers to adverse conditions at the source. Obstacles mainly encompass spatial distance, physical barriers, cultural differences, and other factors between the source and the destination. As research on population migration advances, the concepts of push, pull, and obstacles have been expanded to include social, economic, physical, and institutional dimensions [60,61,62,63]. Moreover, De Haas [64] argues that population migration functions as a result of individuals’ capabilities and aspirations to migrate, where the former encompasses economic capital, social capital, personal skills, and more, while the latter reflects people’s life preferences and their views on opportunities and life in other places.

Building on the push-pull theory, rural migration is conceptualized as being driven by three dimensions: regional environment, spatial barriers, and individual characteristics (Figure 1). At the regional level, the shortcomings of the source villages generate pressures that drive population emigration, while their advantages in certain areas create attractions for population immigration. Accordingly, the shortcomings and advantages of destination villages also produce push and pull factors for migration, respectively. Spatial barriers include national policies, topography, transportation accessibility, and other factors that can inhibit or strengthen the push and pull. Individual characteristics are primarily related to the capabilities and aspirations of the migrating population.

Based on the rural migration model, this study selects the provincial migration scale of the rural population as the dependent variable. Eight variables are preselected from the three dimensions of regional environment, spatial barriers, and individual characteristics, reflecting the key factors identified in the push-pull theory, namely, the rural hukou population, educational attainment of the rural population, per capita GDP, the number of beds in health facilities per 1000 people, internet penetration rate, natural disasters, topographic relief, and spatial distance. These variables are then used to construct a diagnostic model of the factors influencing interprovincial migration among China’s rural population. The variables, along with their definitions and abbreviations, are summarized in

Table 2. Preselected factors influencing the interprovincial migration of China’s rural population.

Dimension	Variable	Definition	Abbreviation
Regional environment	Rural hukou population	Population according to the household registration system (Person)	S_rpopu/D_rpopu
Individual characteristics	Educational attainment of rural population	Years of schooling per capita according to resident population (Year)	S_eduatt/D_eduatt
Regional environment	Per capita GDP	GDP divided by mid-year population (CNY)	S_pgdp/D_pgdp
Regional environment	Number of beds in health facilities per 1000 people	Number of beds of health care institutions at year-end/population at year-end × 1000 (beds)	S_health/D_health
Regional environment	Internet penetration rate	Number of internet users as a percentage of total resident population (%)	S_internet/D_internet
Regional environment	Natural disasters	Total population affected by natural disasters (10,000 person-times)	S_disaster/D_disaster
Spatial barriers	Topographic relief	Standard deviation of regional DEM rasters (-)	S_topo/D_topo
Spatial barriers	Spatial distance	Distance between centers of gravity based on administrative maps (km)	distance

. The model expression is as follows:

$$\begin{array}{l}M{R}_{ijt}={\beta }_0+{\beta }_{1}S\_rpop{u}_{it}+{\beta }_{2}D\_rpop{u}_{jt}+{\beta }_{3}S\_eduat{t}_{it}+{\beta }_{4}D\_eduat{t}_{jt}\\ \begin{array}{cc}& \end{array}+{\beta }_{5}S\_pgd{p}_{it}+{\beta }_{6}D\_pgd{p}_{jt}+{\beta }_{7}S\_healt{h}_{it}+{\beta }_{8}D\_healt{h}_{jt}+{\beta }_{9}S\_interne{t}_{it}\\ \begin{array}{cc}& \end{array}+{\beta }_{10}D\_interne{t}_{jt}+{\beta }_{11}S\_disaste{r}_{it}+{\beta }_{12}D\_disaste{r}_{jt}+{\beta }_{13}S\_top{o}_{it}\\ \begin{array}{cc}& \end{array}+{\beta }_{14}D\_top{o}_{jt}+{\beta }_{15}distanc{e}_{ij}+({\displaystyle \sum id})+{\epsilon }_{ijt}\end{array}$$

where MR_ijt denotes the total rural population migrating from province i to j at time t; S_rpopu_it, S_eduatt_it, S_pgdp_it, S_health_it, S_internet_it, S_diaster_it, and S_topo_it respectively represent the rural hukou population, educational attainment of the rural population, per capita GDP, number of beds in health facilities per 1000 people, internet penetration rate, natural disasters, and topographic relief of the emigration province i at time t; D_rpopu_jt, D_eduatt_jt, D_pgdp_jt, D_health_jt, D_internet_jt, D_diaster_jt, and D_topo_jt respectively refer to the rural hukou population, educational attainment of the rural population, per capita GDP, number of beds in health facilities per 1000 people, internet penetration rate, natural disasters, and topographic relief of the immigrating province j at time t; distance_ij is the straight-line distance between emigration province i and immigrating province j; and id and ε are dummy variables and a random error term, respectively.

2.3. Data Sources and Processing

The decennial population census provides comprehensive data on the size, structure, and characteristics of China’s population. According to the research design, rural demographic data such as population size, educational attainment, and population migration are obtained from the Tabulation on the 2000 and 2010 Population Census of China and China Population Census Yearbook (2020). Among these, data on population size and educational attainment are sourced directly from the first part of the census table, while data on population migration are derived using the long-form data in the second part. Specifically, the long-form data table titled “Population with a hukou registration outside townships and subdistricts, by type of current resident and hukou registration place” reveals insights into population migration. This table classifies the current resident place into three categories, namely, cities, towns, and villages, while the hukou registration place is categorized into townships (including townships, neighborhood committee of towns, village committee of towns, and subdistricts). According to the National Bureau of Statistics of China (NBS), townships and village committee of towns are classified as rural areas, while neighborhood committee of towns and subdistricts are classified as urban areas. As a result, we obtained the specific direction and scale of interprovincial migration of China’s rural population by back-calculating the data on current resident and hukou registration place. Since the long-form data are based on a 10% sample, the migration figures are scaled up by a factor of ten to estimate the total provincial rural migration.

Socio-economic data, such as per capita GDP, internet penetration rate, and number of beds in health facilities per 1000 people, as well as data on natural disasters are collected from the China Statistical Yearbook of 2001, 2011, and 2021. Straight-line distances between provinces are determined by calculating the distance between the centers of gravity based on the geographic center of each province. The topographic relief data are from the China 1 km topographic relief dataset [65].

3. Results

3.1. Spatial Patterns of Interprovincial Rural Migration

According to population census data, the scale of interprovincial migration of China’s rural population in 2020 increased from 25.80 million in 2000 to 94.71 million, with an average annual growth rate of 6.70%. Over a 10-year period, the average annual growth rate of the scale of interprovincial migration of the rural population reached 9.59% in 2000–2010, while it was only 3.88% in 2010–2020. This implies that the interprovincial migration of China’s rural population continued to grow from 2000 to 2020 but showed a slowing trend.

Spatially, the scale of rural emigration in each province from 2000 to 2020 is characterized by differentiation along the Heihe-Tengchong Line, with a clear “center-periphery” pattern, i.e., the central provinces were the high-value area, while the border and eastern coastal provinces were mostly the low-value area (Figure 2). Specifically, the scale of provincial rural emigration in 2000 was generally small, with 17 provinces below 0.50 million, and even 8 provinces less than 0.10 million, including Tibet (8.75 thousand), Beijing (23.70 thousand), Tianjin (28.23 thousand), Shanghai (30.64 thousand), Qinghai (46.03 thousand), Ningxia (48.78 thousand), Xinjiang (54.58 thousand), and Hainan (56.25 thousand). There are 9 provinces exceeding 1 million, of which the largest is Sichuan (3.65 million), followed by Hunan (2.79 million) and Anhui (2.71 million). In 2010, the number of provinces in which rural emigration was greater than 1 million increased to 18, with Anhui, Sichuan, Henan, and Hunan exceeding 5 million, reaching 8.30 million, 7.23 million, 6.98 million, and 5.96 million, respectively; the only provinces with fewer than 0.10 million were Tibet (20.30 thousand), Shanghai (26.37 thousand), Beijing (43.84 thousand), and Tianjin (63.90 thousand). In 2020, the number of provinces with rural emigration greater than 1 million increased to 20, with Henan and Anhui even exceeding 10 million, reaching 11.07 million and 10.82 million, respectively; the provinces lower than 0.10 million were Tibet (55.83 thousand), Shanghai (66.28 thousand), and Beijing (80.97 thousand). This highlights the continued concentration of emigration from central provinces.

While rural emigration exhibited a clear spatial pattern, the destinations of these migrants also revealed distinct regional preferences. In terms of destination, the most popular provinces for rural immigration during the research period were primarily the eastern coastal provinces of Beijing, Shanghai, Zhejiang, Jiangsu, Fujian, and Guangdong, and the ratio of interprovincial rural immigration in these six provinces accounted for 70.31%, 77.03%, and 72.59% of the national total in 2000, 2010, and 2020, respectively. Some provinces in western China experienced significant growth in the scale of rural immigration, particularly Sichuan and Chongqing, where the provincial rural immigration increments reached 1.04 million and 1.07 million in the period of 2000–2020, respectively (Figure 3). Specifically, in 2000, most provinces recorded immigration between 0.10 and 0.60 million, with only Tibet (47.71 thousand) and Qinghai (59.30 thousand) falling below 0.10 million, which were at the end of the range in the following two decades. At the same time, 7 provinces exceeded 1 million, led by Guangdong (9.87 million) and Zhejiang (2.52 million). By 2010, the number of provinces exceeding 1 million rose to 10, with Guangdong (18.15 million), Zhejiang (11.13 million), Shanghai (7.04 million), and Jiangsu (5.89 million) surpassing 5 million. This trend continued in 2020, with 17 provinces exceeding 1 million and Guangdong, Zhejiang, Shanghai, and Jiangsu remaining the top destinations. Notably, Beijing and Fujian also saw significant growth, reaching 4.62 million and 4.53 million, respectively. This reflects the persistent attractiveness of eastern coastal provinces as migration destinations.

3.2. Network Characteristics of Interprovincial Rural Migration

This study defines migration flows with scales greater than or equal to the national average as degrees and employs an improved degree-based SNA method to analyze the interprovincial migration network of China’s rural population (

Table 3. SNA indicators of the interprovincial rural migration network in China.

Indicator	2000	2010	2020
Number of relationships	361	133	43
Network density	0.388	0.143	0.046
Outdegree centralization (%)	39.11	43.78	33.11
Indegree centralization (%)	56.33	36.89	22.78
Betweenness centralization (%)	10.54	4.01	0.96

). The results reveal a high degree of outdegree and indegree centralization in the interprovincial migration network, indicating strong network centrality. The fact that the indegree centralization is greater than the outdegree centralization indicates that rural immigration is more concentrated in specific provinces than emigration, and the immigration provinces have a stronger impact on the interprovincial migration network of China’s rural population. However, in 2020, the situation reversed. The decline in both outdegree and indegree centralization from 2000 to 2020 suggests a decentralization trend in rural migration flows, and the migration activities are becoming more dispersed and balanced. This change is closely related to balanced regional economic development, industrial transfers, and improved transportation infrastructure. For example, whereas in the past, the rural population was concentrated in the eastern coastal provinces, with the improvement of the transportation network and economic development in the central and western regions, more people are choosing to take up employment and settle in emerging regions such as Chengdu-Chongqing and Zhengzhou, which is contributing to the decentralization of migratory activities and the balanced development of the region. The declining betweenness centralization indicates the diversification of core nodes in the interprovincial migration network of the rural population. In general, from 2000 to 2020, the interprovincial migration network of China’s rural population became more balanced, with the role of the core nodes gradually weakening and the concentration of rural emigration and immigration declining.

In addition to the overall network characteristics, the economic division of migration sources and destinations provides further insights into the spatial patterns of rural migration. From the perspective of economic division, central and western China consistently account for approximately 81% of national interprovincial rural emigration, highlighting their role as primary migration sources. Eastern China is the main destination of interprovincial rural migration, with its share of the national total interprovincial rural immigration in 2000, 2010, and 2020 being 76.49%, 82.96%, and 79.06%, respectively (Figure 4). From 2000 to 2020, the patterns of interprovincial rural migration networks in the four regions remained relatively stable but with notable shifts in regional dynamics. In 2000, intraregional migration dominated in the eastern region, accounting for 67.01% of total emigration. Northeastern China showed a dual pattern, with 45.53% of migration directed to eastern China and 41.49% occurring within the region. Central and western China primarily sent migrants to eastern China, representing 86.38% and 70.95% of their total emigration, respectively. By 2010 and 2020, the trend of intraregional migration in eastern China and rural migration from central and western China to eastern China initially strengthened but later weakened. Notably, the intraregional share of migration in northeastern China declined steadily, dropping to 25.62% by 2020, while the share of interregional migration to eastern China rose to 60.97% during the same period.

In the year 2000, rural out-migrants from Hunan, Guangxi, Jiangxi, and Hubei predominantly relocated to Guangdong, while Henan, Anhui, and Jiangxi experienced substantial migration flows toward the Yangtze River Delta (YRD) region, which includes Jiangsu, Zhejiang, and Shanghai. Concurrently, a significant number of rural residents from Sichuan and Chongqing migrated to economically advanced provinces such as Guangdong, Zhejiang, Shanghai, and Jiangsu. Consequently, Guangdong and the YRD region emerged as the two primary destinations for interprovincial rural migration, forming an “umbrella” pattern alongside the traditional agricultural areas and the Sichuan-Chongqing region as key sources. By 2010, both Guangdong and the YRD region had further solidified their roles as pivotal nodes in China’s interprovincial rural migration network, with the YRD region surpassing Guangdong as the dominant hub. The substantial rural migration from Henan to both Guangdong and the YRD region highlighted Henan’s increasingly prominent role in the network. Additionally, the number of rural migrants from Hubei to Guangdong and from Anhui to the YRD region continued to grow, leading to the evolution of the interprovincial rural migration network into a “funnel-shaped” pattern. By 2020, this “funnel-shaped” pattern had stabilized, reflecting the enduring influence of the economic hubs such as Guangdong and the YRD region, alongside emerging migration trends in western China driven by regional development and improved infrastructure. Meanwhile, the trend of rural out-migration from Jilin, Liaoning, and Heilongjiang became increasingly pronounced, while immigration to Xinjiang saw a significant rise (Figure 5).

In terms of the evolution of emigration–immigration dynamics, the sources of rural migration in 2000 were concentrated in the southwestern and central regions, with destinations primarily located in the eastern region, indicating a clear divergence. By 2010, the scale of migration from both popular sources and destinations had further increased, with a notable divergence between immigration and emigration patterns in the central and western regions compared to the eastern region. However, while these source regions continued to produce a large number of emigrants, they also attracted significant numbers of rural immigrants from other provinces. Similarly, popular destinations generated larger numbers of rural emigrants. By 2020, although a significant division between emigration and immigration provinces persisted, there was a more evident convergence between destinations and sources.

3.3. Influencing Factors of Interprovincial Rural Migration

To enhance data stability and mitigate scale-related biases, all variables were logarithmically transformed. The result of the Lagrange Multiplier (LM) test (p = 0.000) indicates that the random effect (RE) model is superior to the mixed effects model, and the result of the Hausman test (p = 0.000) indicates that the fixed effect (FE) model is superior to the RE model. Further, a comparison of the individual FE model and the two-way fixed-effects (FE_TW) model was conducted, and the results strongly reject the original hypothesis of no time fixed effects, leading to the selection of the FE_TW model. Due to the presence of explanatory variables that do not change over time, this study utilized the Least Squares Dummy Variable (LSDV) method, which allows for both fixed effects and the observation of variables that do not change over time. As a result, the LSDV method is employed to diagnose the drivers of interprovincial migration of China’s rural population from 2000 to 2020 (

Table 4. Model estimation results.

	OLS		RE		FE		FE_TW		LSDV
	Coefficient	z	Coefficient	z	Coefficient	z	Coefficient	z	Coefficient	z
ln S_rpopu	0.44 ***	−7.16	0.58 ***	12.45	−0.06	−0.57	−0.15	−1.18	−0.15	−0.97
ln S_eduatt	−1.81 ***	−5.83	−1.84 ***	−5.83	−1.07 **	−2.25	−0.50	−0.81	−0.50	−0.66
ln S_pgdp	−0.40 ***	−6.47	−0.19 ***	−3.79	0.01	0.23	−0.21 ***	−2.92	−0.21 **	−2.38
ln S_health	0.59 ***	9.47	0.41 ***	9.49	0.04	−0.66	0.04	0.55	0.04	0.45
ln S_internet	−0.00	−0.15	−0.01 ***	−3.24	−0.01 ***	−4.43	−0.01 ***	−5.03	−0.01 ***	−4.10
ln S_disaster	0.15 ***	6.56	0.14 ***	11.03	0.11 ***	7.16	0.11 ***	7.72	0.11 ***	6.30
ln S_topo	0.01	1.18	0.01	1.34	0.00	-	0.00	-	0.75 **	2.44
ln D_rpopu	1.13 ***	18.59	0.70 ***	15.05	0.75 ***	7.02	0.66 ***	5.05	0.66 ***	4.11
ln D_eduatt	2.72 ***	8.78	1.61 ***	5.1	−0.51	−1.08	0.05	0.10	0.05	0.08
ln D_pgdp	1.18 ***	19.02	0.84 ***	16.86	0.64 ***	11.51	0.41 ***	5.18	0.41 ***	4.22
ln D_health	−0.44 ***	−6.99	−0.14 ***	−3.14	0.16 ***	2.8	0.16 *	1.89	0.16	1.54
ln D_internet	−0.00	−1.40	0.00 *	1.68	0.00	0.77	−0.00	−0.11	−0.00	−0.09
ln D_disaster	−0.26 ***	−10.98	−0.14 ***	−10.96	−0.10 ***	−6.30	−0.09 ***	−6.92	−0.09 ***	−5.64
ln D_topo	0.01 **	2.08	−0.02 *	−1.76	0.00	-	0.00	-	−0.31 ***	−4.07
ln distance	−1.20 ***	−35.00	−1.20 ***	−22.49	−0.08	−0.24	−0.07 **	−2.34	−0.07 *	−1.91
Cons	−26.74 ***	−33.13	−23.76 ***	−25.55	−15.16 ***	−4.75	−8.55 **	−2.13	−9.40 **	−2.07
N	2790		2790		2790		2790		2790

Note: *** p < 0.01, ** p < 0.05, and * p < 0.1.

). The results indicate that per capita GDP, natural disasters, topographic relief, and spatial distance are the primary drivers of interprovincial rural migration. Additionally, the rural hukou population and internet penetration rate also play significant roles.

The effect of rural hukou populations in the destination province on the scale of interprovincial rural migration is significantly positive, which is not in line with expectations. This may be due to the fact that the size of the rural hukou population reflects, to a certain extent, the general situation of rural natural and human systems of the destination province. The large size of the rural hukou population implies that the province’s villages have socio-economic and resource-environmental advantages, thus attracting population immigration. In the source regions, the effect of the total rural hukou population on interprovincial rural migration is not statistically significant.

Economic development level, as represented by per capita GDP, is an important factor influencing interprovincial rural migration. A high level of economic development means that the province can provide more employment opportunities and that the income level of the people is higher. Accordingly, the province can attract more rural migrants from other provinces while retaining its rural population. A 1% decrease in the source’s economic development level increases interprovincial rural migration by 0.21%, while a 1% increase in the destination’s economic development level boosts rural immigration by 0.41%.

Natural disasters have a significant impact on interprovincial rural migration. Human instincts to seek benefits and avoid harm dictate that people tend to avoid areas prone to natural disasters since their basic security needs cannot be met and their productive and living activities cannot be effectively conducted in such areas. This situation is significant not only in the source provinces but also in the destination provinces. Every 1% increase in natural disasters in the source regions is associated with a 0.11% increase in rural emigration, and every 1% increase in natural disasters in the destination is associated with a 0.09% decrease in rural immigration.

Topographic relief is an important indicator of regional physical geography. Large topographic relief often implies poor spatial connectivity and accessibility, which is not conducive to productive and living activities. Meanwhile, topographic relief has an important impact on natural conditions such as climate and hydrology and plays a role in the distribution of natural resources such as farmland. In the case of farmland, for example, a high degree of topographic relief often means that farmland is insufficient and difficult to utilize. Therefore, large topographic relief in the source implies a larger scale of interprovincial rural emigration, while the scale of rural immigration in the destination is significantly negatively correlated with topographic relief.

Information accessibility characterized by the internet penetration rate in the source is negatively correlated with the scale of interprovincial rural emigration. In the current internet era, information has become an important resource for regional development. A high information accessibility means more resources are available to promote regional development, which in turn creates conditions for the rural population to remain in the region for employment and entrepreneurship. Additionally, information accessibility can reflect the level of regional economic development to a certain extent, and a higher value suggests a higher economic development level and lower rural emigration. In the destination, the effect of the internet penetration rate on rural immigration is not significant.

Spatial distance has a strong negative effect on interprovincial rural migration. Every 1% increase in distance between provinces is associated with a 0.07% decrease in the size of rural migration between them. In recent years, the central government and local governments have made great efforts to establish a modernized transport system, which has greatly facilitated the flow of people, goods, and information between regions. However, the role of spatial resistance due to distance in constraining rural migration remains significant. This underscores the enduring role of geography in shaping migration patterns, even in an era of economic and technological advancement.

4. Discussion

4.1. Gradually Reversed Rural Migration

Since the reform and opening-up in 1978, the Chinese government has implemented an unbalanced development strategy tilted towards the eastern region, prioritizing the investment of limited resources in this area, which in turn led to the development of the whole country through the development of the eastern region [66,67]. This strategy has shaped China’s unbalanced pattern of economic development and dictated a steady flow of population and other factors of production to the eastern region, especially the Yangtze River Delta, Pearl River Delta, and Beijing–Tianjin–Hebei [68]. In this study, we confirm the large-scale interprovincial migration in China over the past two decades, and the main destination has been the eastern coastal region. On the other hand, the rural population mainly migrates to urban areas for better production and living conditions due to the urban–rural dual structure and its resulting urban–rural gap [69,70,71].

Although socio-economic development has driven interprovincial rural migration to continue to rise, its growth rate has shown a clear downward trend, and the proportion of interprovincial migration within total rural migration has been declining. In the case of migrant workers, who account for the vast majority of rural migration, data from the National Bureau of Statistics (NBS) show that the proportion of interprovincial migrant workers in the total number of migrant workers has declined from 51.2% in 2009 to 38.2% in 2023. In addition to being influenced by major national strategies, this change is also closely related to the law of industrial gradient transfer. The former mainly refers to China’s vigorously implemented regional coordinated development strategy, which aims to form a new pattern of mutual promotion, complementary advantages, and common development among the eastern, central, and western parts of the country through sound market mechanisms, cooperation mechanisms, and support mechanisms [72,73,74]. The latter focuses on the central and western regions actively undertaking industrial transfers by leveraging their abundant natural resources, low factor costs, and large market potential, effectively promoting their industrialization and urbanization [75,76]. Meanwhile, industrial transformation and upgrading in the eastern region have resulted in the reduction in a large number of labor-intensive jobs, leading to low-skilled migrant workers being continuously pushed out of the labor market [77,78,79,80]. Additionally, economic slowdown has reduced employment opportunities, causing a significant number of migrant workers to return to their hometowns [81]. Based on the above trends and influencing factors, it is necessary to further explore the profound impacts of rural migration on economic and social development and propose targeted policy recommendations to address the current challenges.

4.2. National Strategy-Oriented Suggestions

Rural migration has played an important role in China’s economic and social development; not only does it provide abundant cheap labor for rapid industrialization, but it also greatly contributes to urbanization [82,83]. In fact, rural migration connects the city on one end and the countryside on the other, serving as a link between urban and rural areas. Large-scale rural migration has exacerbated rural decline, causing serious problems such as the rapid non-agriculturalization of production factors, hollowing out of construction land, and aging and weakening of social entities, which has become an important challenge for China in promoting rural comprehensive revitalization [84,85,86]. On the other hand, due to the hukou system that separates urban and rural areas, it is difficult for the vast majority of rural migrants to integrate into urban society, resulting in a huge gap between the urbanization rates of the hukou holders and the resident population, leading to the low-quality development of urbanization [85,87].

Population migration is an inevitable trend in socio-economic development [88]. With social development and social progress, the trend of rural migration is expected to continue and to expand steadily. If the existing system is not reformed, rural migration will certainly bring severe challenges to the implementation of rural revitalization and new urbanization strategies [84,89]. In view of the current situation of “migrant China”, some targeted measures need to be adopted to guide the orderly migration of the rural population, which can thereby effectively alleviate its negative impact and promote socio-economic development.

The first is to promote regional coordinated development. Unbalanced regional development is the underlying logic of population migration [90]. It is necessary to optimize the national pattern of the industrial division of labor based on regional comparative advantages and accelerate the removal of interest and policy barriers between regions, forming a new mechanism for regional coordinated development with strong integration, orderly competition, green coordination, sharing, and win–win results. The second is to deepen rural hukou system reform. Population migration is neither good nor bad, but the urban–rural dual structure based on the hukou system causes population migration to have a significant negative impact on socio-economic development [91,92]. Therefore, there is an urgent need to deepen hukou system reform under the orientation of fairness and justice and accelerate the settlement of the non-hukou population. Meanwhile, more attention should be given to improving the provision of public services for the migrants to narrow the gap of public services between the hukou and the non-hukou populations, to achieve a balanced and inclusive provision of basic public services at the place of residence and improve the quality of urbanization. The third is to smooth the factor flow mechanisms. A large part of the negative impact of population migration on regional development stems from the fact that a mechanism for the rational and orderly flow of factors has not been established [93]. Thus, it is necessary to accelerate the removal of institutional barriers to factor flow and to use rural migration as a link to smooth the flow channels of regional and urban–rural factors such as land and capital, thereby driving economic cycles and promoting the rural comprehensive revitalization and the development of new-type urbanization.

4.3. Limitations and Future Prospects

In the broad discourse on population migration, issues related to rural migration have not received sufficient attention, despite the fact that the majority of China’s mi-grants originate from rural areas—a demographic with unique characteristics pivotal to the nation’s economic and social development [94,95]. This underscores the importance of conducting research on rural migration. In this study, we explore the spatio-temporal patterns of interprovincial rural migration in China and its influencing factors over the past two decades. Our goal is to provide a macroscopic understanding of rural migration dynamics while acknowledging the limitations of using the separation between residential and hukou locations—a concept widely accepted by scholars in China—to define population migration. This metric often fails to capture the full trajectory of population movements and frequently misaligns statistical timing with actual migration behavior, thereby highlighting the need for more accurate and effective indicators as technology advances. Policy interventions, particularly those related to land rights and social security, play a critical role in promoting or limiting population migration. However, selecting appropriate indicators to quantify these interventions remains a significant challenge. While contextual factors such as economic disparities and the physical environment are undoubtedly important, moderating factors such as individual capabilities and social networks also require further investigation. To deepen the understanding of the rural migration patterns in China, future research should focus on regions at different stages of development. It is essential to conduct micro-level analyses of migration motivations, path choices, and occupational characteristics, as well as to incorporate socio-cultural factors such as gender, age, and family structure. Addressing the social consequences of migration, such as family separation and cultural disruption, will provide a more comprehensive perspective. By integrating micro-level research data, socio-cultural factors, and comparative studies across regions, we can better summarize the patterns of rural migration in areas at different developmental stages. This approach will offer a more nuanced and detailed understanding of China’s rural migration patterns and mechanisms, ultimately revealing the complexity and diversity of rural migration through the combination of macro and micro perspectives, and will thus provide a scientific foundation for the formulation of targeted policies.

5. Conclusions

In the past few decades, waves of industrialization and urbanization have resulted in an unprecedented rural migration in China, which became one of the key features of China’s social transformation. Utilizing data from the 5th, 6th, and 7th population censuses, this study examines the characteristics of rural migration in China from 2000 to 2010 and establishes an econometric model based on the theoretical understanding of rural migration to diagnose the influencing factors. The main findings are as follows:

From 2000 to 2020, the interprovincial migration of China’s rural population grew rapidly, although the growth rate gradually slowed down. Spatially, the scale of provincial rural emigration is characterized by differentiation along the Heihe-Tengchong Line, exhibiting a clear “center-periphery” pattern. The destinations of interprovincial rural migration are mainly the eastern coastal provinces such as Jiangsu, Shanghai, Zhejiang, and Guangdong.

The results of SNA show that the centrality of the interprovincial migration network of China’s rural population from 2000 to 2020 is strong, while the role of the core nodes has gradually weakened. Although the concentration of interprovincial rural immigration in specific provinces is more significant than that of emigration, both are weakening in their concentration characteristics during the study period.

From 2000 to 2020, the interprovincial migration of China’s rural population mainly occurred in the provinces on the southeastern side of the Heihe-Tengchong Line, showing a clear tendency toward proximity in the choice of destinations. The interprovincial migration network of the rural population has evolved from an “umbrella” pattern to a “funnel-shaped” pattern.

The model estimation results based on the LSDV method show that per capita GDP, natural disasters, topographic relief, and spatial distance are the central influences on the evolution of the interprovincial migration network of China’s rural population from 2000 to 2020, while the rural hukou population and the internet penetration rate have significant but comparatively smaller effects.