The impact of urban integration on the intention of settlement and residence among migrants: The empirical evidence from China

Abstract

Promoting urban integration is the key to improving the efficiency of labor allocation in developing countries and promoting coordinated regional development. Using the data of the migrants' observation sample in 2012–2017 released by China Migrants Dynamic Survey, this paper draws on theories related to urban integration and labor migration to study the urban integration status, intention of settlement and residence in China's agricultural household labor force in depth, as well as the correlation between the two. The study finds a significant decline in the intention of settlement and a gradual increase in the intention of residence among the migrants. The empirical regressions suggest that good urban integration, especially psychological integration, mainly contributes to their intention of settlement and residence, while the intention of residence is also more influenced by the social and economic integration dimensions. The mediating effect tests reveal that income level and housing stability are key pathways to economic integration, public medical participation and accessibility to medical services are key pathways to health integration, social status and local attachment are key pathways to social integration, and identity affiliation and psychological assimilation are key pathways to psychological integration. Future urbanization in developing countries requires not only further reform of the household registration system but also government, business, and social organizations at all levels to reduce the difficulties of urban integration in economic, health, social, and psychological aspects, enhance the degree of urban integration of the migrants, increase the intention of the migrants to settle and stay in different places and avoid the "migratory bird" migration of labor between regions. The "migratory bird" migration between regions is avoided.

1. Introduction

Labor mobility is a key factor in activating the labor resources endowment of developing countries, balancing the labor supply and demand among regions, and consequently improving labor efficiency. According to the data from China's seventh national population census in 2020, the total number of the migrants reached 376 million, with 331 million moving to urban areas, accounting for a significant 88.12% of the overall migrants. Massive population movements accompany rapid urbanization, especially in eastern coastal regions such as Jiangsu, Zhejiang, and Guangdong, where the urban population already exceeds 70%. However, when examining the development trajectory of major developed countries, it is evident that the urbanization rate beyond 70% signals a shift towards urban-rural integration, causing a deceleration in the urbanization process. Currently, significant challenges persist in achieving labor force integration for China's migrants. Job seeking, healthcare, and housing exert notable influences on the intentions of this population to settle and reside in their work locations, thereby impacting their urban integration (Cao, 2015; [1,2]). China Migrants Dynamic Survey (CMDS) show that in 2016 and 2017, over 50% of the migrants was willing to reside in their work locations for more than five years, but only about one-third were willing to transfer their household registration to these locations. This indicates that a key aspect of China's urbanization lies in the degree of urban integration for the migrants.

The objective of this study is to construct a comprehensive index for urban integration based on economic integration, social integration, health integration, and psychological integration. We aim to analyze the level and changes in the intention of the migrants to settle and reside, further investigating the impact of urban integration on these intentions. Research on the settlement and residence intentions of China's migrants is of paramount importance, involving multiple issues such as urbanization, social stability, economic development, social security, resource allocation, and cultural integration. The scale and distribution of the migrants influence urban population structure and resource allocation, making an understanding of their residence intentions crucial for urban resource planning and employment policy formulation. Additionally, the integration of the migrants and cultural exchange are essential for social harmony. Accurate data is necessary for the government to develop appropriate policies, promote orderly and reasonable population movement, and ensure social stability and sustainable economic development.

Given these considerations, this study combines with macroeconomic development data on migration, to comprehensively assess the urban integration level of the migrants with rural household registration. We compare the differences in urban integration among various types of migrants and analyze the impact of urban integration levels in different dimensions on their intentions to settle and reside. This research contributes significantly in several aspects. Firstly, it helps to capture the dynamic trends of population movement, facilitating the prediction of future population movements. Secondly, it provides empirical evidence for government policy-making, promoting social stability and economic development. Furthermore, optimizing urban resource allocation to meet the residential needs of the mobile population can drive sustainable urban development. Analyzing group differences aids in fostering social harmony, meeting the needs of various groups, and enhancing integration. Lastly, this study enriches the existing knowledge in relevant research fields and makes marginal contributions to the study of urban development. In conclusion, this research holds extensive value for policy-making, resource allocation, and both social and academic domains.

This study is based on the analysis of China Migrants Dynamic Survey (2012–2017), a national population monitoring survey organized by the National Health Commission. The survey utilized a random sampling method in areas where the migrants was concentrated, excluding Tibet, Hong Kong, Macao, and Taiwan. The sample is representative of both the national and provincial populations. The survey targeted individuals aged 15 and above who had resided in the destination for more than one month and who did not hold local household registration. The sampling framework was based on the annual report data on the national migrants from the year before the survey, using a stratified, multistage, and proportionate-to-size (PPS) sampling method. In this study, we first analyze the changing trends in the intentions of the mobile population to settle based on 544,989 observation data from 2012 to 2017. Next, we focus on the integration status and decomposed settlement intentions for the year 2017, utilizing the data from a special survey on urban integration involving 169,989 samples. The large sample size and reasonable distribution make this data representative of the basic characteristics and regional distribution of China's migrants with rural household registration. Regional data sources are from China Urban Statistics Yearbook and China Coastal City Economic Statistics Database.

2. Literature review

The issue of social integration of immigrants and refugees has always been an important topic of study in academia. Gordon [3] proposed a method for measuring the degree of integration across seven dimensions: cultural, structural, marital, identificational, attitudinal reception, behavioral reception, and civic engagement. Harder and other scholars [4] constructed a six-dimensional index system that includes psychological, economic, political, social, linguistic, and activity scope to measure the degree of integration of immigrants. The impact of immigrants on local economic and social development is also a focus of research. For example, Egger and colleagues [5] found that the addition of immigrants has allowed transnational companies to form more stable raw material supply relationships, reducing the number of suppliers and increasing trade with a few suppliers. Research by Hall and others [6] found that immigrants not only change the demographic ethnic composition of the host countries but also have significant effects on the economic, social, cultural, and political ecology of those nations.

On the issue of urban integration of migrants, the degree of urban integration of migrants varies significantly across regions and groups. Zhang et al. [1] suggested that living with children positively impacted urban social inclusion, while living with daughters had a more significant impact. Migrant distance and family wealth negatively impacted urban integration among older adults. Zou and Deng [7] showed that the digital economy had a significant negative impact on urban migrant integration and further found that the digital economy was more conducive to the economic integration of female and new-generation migrants while hindering their psychological integration hindering the sociocultural integration of older migrants [8]. Chen and Liu's [9] study suggested that although immigrants with better human capital are more inclined to settle in cities, sociocultural attachment plays an equally important role in determining immigrants' intention of settlement. In addition, studies by many scholars have found that sociocultural factors Vroome and Tubergen, 2014 and psychological factors (e.g., social networks, interaction with locals, sense of belonging, perceived social attitudes, etc.) are also strongly associated with the level of urban integration of migrants [1,2,[10], [11], [12], [13]].

Research on the settlement intentions and willingness to stay of the floating population mainly focuses on exploring various potential factors that influence settlement decisions. Existing studies have found that individual characteristics and human capital, such as age, gender, education level, and health status, significantly affect the settlement intentions and willingness to stay of the migrant population [[14], [15], [16]]. Some scholars have also found that institutional guarantees, such as the household registration system [17], health insurance, labor relations and work environment [18,19], and type of employment [20], all have a significant impact on the willingness of the migrant population to settle and stay. In addition, good interpersonal relationships [21] and positive social evaluations from the urban resident community [22] both help to strengthen the willingness of migrant workers to settle and stay. Furthermore, cities with clear geographical advantages [23] are more likely to attract migrant workers to settle, while air pollution significantly increases the intention of local hukou labor to move across districts, counties, and even countries [24][25].

Finally, other scholars proposed new ideas and perspectives from migration theory in the context of developing countries. Haas et al. [26] tested these hypotheses by examining the main determinants of the intention to return of Moroccan migrants across Europe. The results suggest that structural integration through labor market participation, education, and maintaining economic and social ties with the receiving country do not significantly impact the intention to return. Tang [27] considered whether psychological integration at the local level may underlie the relationship between social integration and the intention of residence, considering both structural and cultural integration, conceptualizing social integration as ties to locals, and access to emotional and instrumental social capital in the destination country [28]. Toruńczyk-Ruiz and Brunarska [29] found in their study of Ukrainian migrants that emotional, social capital, rather than instrumental social capital, indirectly affects the intention of settlement through place attachment. Mazza and Punzo [30] argued for a relationship between chain migration and spatial attractiveness correlation. They found that different parts of the city show different adaptations to migrants of different ethnicities, with groups primarily engaged in domestic and caregiving being more dispersed than those specializing in hawking and retailing. Sri Lankans, Mauritians, Senegalese, and Chinese have a significant spatial attraction. In contrast, the settlement patterns of Tunisians and Moroccans are consistent with a random distribution.

Throughout the existing studies, this paper finds that: first, urban integration is related to the intention of settlement or residence, but there are significant differences among migrants, and most of the existing literature only focuses on one of them, but few studies investigate the relationship between the two. Second, there are significant differences in urban integration, the intention of settlement, and the intention of residence among different types of migrants. However, few relevant analyses exist in the existing literature and a few focus on intergenerational differences. Third, existing studies are primarily based on research data from a few regions. However, there are significant differences among different regions and levels of cities in China, and the conclusions based on a few regions may not necessarily have universal applicability.

3. Urban integration and group comparison of the migrants with agricultural household registration

3.1. Basic characteristics of migrants

Table 1 reports the basic characteristics of migrants. Overall, the proportion of migrants with agricultural household registration is as high as 77.98%. The education level is an important factor that cannot be ignored in comparing the proportion of migrants based on different characteristics. The proportion of migrants with lower education levels is often several times higher in the agricultural household, male, single, and self-employed than in the highly educated migrants. Compared to the apparent differences in education, there is not much difference between migrants moving within or outside the province, such as the proportion of agricultural household registration flow to other regions in the province or to flow outside the province is about 38%. Notably, only 7% of migrants have quality jobs, such as government employees or state-owned enterprises workers, which indicates that unstable working conditions and poor job treatment are important factors for most migrants to go out to work.

Table 1.

Basic information about the migrants.

Variables	Overall	Intra-provincial mobility	Cross-provincial mobility	Low education level	High level of education
Agricultural household registration (%)	77.98	38.16	39.82	69.33	8.65
Male (%)	51.69	25.41	26.28	43.16	8.53
Han nationality (%)	90.59	44.61	45.98	74.50	16.09
Single (%)	15.11	8.04	7.07	10.50	4.61
Cross-provincial mobility (%)	49.29	0	49.29	41.50	7.79
Agency/State Enterprise (%)	7.01	4.10	2.91	3.57	3.44
Individual business (%)	34.09	17.69	16.40	31.02	3.07
Age (N/%)	36.66 (37.52)	36.54 (17.81)	36.78 (19.71)	37.78 (34.17)	31.36 (0.68)
Education level (L./%)	3.44 (35.72)	3.50 (38.74)	3.39 (32.85)	3.03 (93.37)	5.43 (6.63)
Health Level (L./%)	1.21 (23.58)	1.23 (50.72)	1.19 (49.28)	1.22 (93.00)	1.12 (7.00)
Monthly salary (N/%)	4328.49 (38.44)	3892.46 (47.03)	4157.51 (39.01)	4022.08 (39.14)	5696.66 (27.75)

Note: Educational level includes 7 levels, in order of 1 = no school, 2 = elementary school, 3 = junior high school, 4 = high school/junior high school, 5 = college specialist, 6 = undergraduate, 7 = graduate; health status includes 4 levels, in order of 4 = unable to take care of themselves, 3 = unhealthy but able to take care of themselves, 2 = essential health, 1 = healthy; family size refers to the number of family members living together in the household (including Local, home and other places, excluding children who are married and separated); the number of mobile cities to the number of cities where the total working and living for 1 month or more; the number of institutions/state enterprises including state organs, party organizations, state-owned and state-controlled enterprises, collective enterprises; the monthly salary refers to the sum of the last month's salary or net income plus the discounted subsidies of the employment unit or employer for food and accommodation, or 0 if there is none. N refers to the number of cases, while L refers the grading or level. In the context of factors such as age and education level, the value inside the parentheses represents the proportion of individuals above the average.

Some other characteristics also show the complete picture of the sample data in this paper. For instance, the majority of migrants are around 36 years old; the number of family members of the migrants is around 3, probably because such small families are more convenient to go out to work; the average mobility regions for migrants are about two cities; the average years of education of the migrants is not high, around three years; the average monthly wage is about 4328 RMB, which is much lower than the national urban non-private sector in 2017. Besides, the average monthly wage of employed persons (RMB 6193) level and the migrants with agricultural household registration is still concentrated in low-skill and low-wage employment. The overall health status of the agricultural migrants is good, with more than 95% of the migrants considering themselves healthy.

3.2. Analysis of intention of settlement and residence

The intention of settlement is the subjective intention of the migrants to settle down in the city where they work, which is determined by asking, "If you meet the conditions for settling down in the local area, are you willing to move your household to the local area". Those who explicitly expressed "willing" were considered willing to settle down, while those who answered "unwilling" and "not thinking" were considered as not willing to settle down for the time being. This classification is consistent with the existing literature [18,19]. In addition, the data also asked migrants about their intention of residence in the city, which was determined by asking, "Do you plan to stay and reside in the city in the future", where an explicit "yes" were considered a willing to reside in the city, thus analyzing the intention of residence in the city of migrant agricultural households. In this way, we analyze the intention of migrants to reside in the city.

As Fig. 1 presents a comparison of the intention of settlement of migrants with different characteristics in 2012–2017 (the intention of settlement of migrants who were not surveyed in 2015). It reveals that the intention of settlement of migrants in 2016–2017 has significantly decreased compared to the early stage of the survey. From the comparison of characteristics, more female migrants have a solid intention of settlement, but compared to male groups, there is not a big difference in the proportion; migrants with characteristics such as higher than average income level, higher education, and non-farm household registration have a higher proportion of intention of settlement, and the gap is evident compared to the corresponding groups, which may indicate that the latter face more difficulties in the urban settlement; married groups have a higher proportion of intention of settlement compared to single groups. Regarding age, the intention of settlement is higher among those under 30 years old and those between 30 and 60 years old and lower among those above 60 years old.

Fig. 1.

Comparison of the intention of settlement in different groups in 2012–2017 (the intention of settlement of migrants who were not surveyed in 2015).

Fig. 2 depicts the intention of residence of migrants with different characteristics in 2012–2017. The results find that the intention to residence of migrants was stable in 2012–2016, while there was a significant increase in the intention of residence in 2017. From the comparison of characteristics, the ratio of the intention of residence does not show a gap at the gender level; the intention of residence of various groups such as above average income level, higher education, non-farm household registration, and age group less than 60 years old are significantly higher than the comparison groups, which is very similar to the characteristics of the intention of settlement.

Fig. 2.

Comparison of the intention of residence in different groups in 2012–2017.

The comparison of the above results shows that the overall trend of the high intention of residence and low intention of settlement is consistent with the current reality of frequent mobility of migrants. In addition, the migrants with agricultural registration may face more difficulties settling in cities. At the same time, for the group with lower education and poorer job offerings, more frequent cross-regional mobility can expand the scope of job search and thus increase the probability of entering higher-income positions.

3.3. Urban integration status of migrants

Urban integration consists of four dimensions: economic integration, health integration, social integration, and psychological integration, each measured by a cluster of variables. Economic integration is measured by expenditure structure (the proportion of real-estate expenditure, i.e., rent and mortgage to total household expenditure) and living conditions (whether they live in purchased commercial housing, subsidized housing, small property rights housing, or self-built housing). Health is measured by whether migrants have purchased local urban residents' medical insurance, urban labor, or public medical insurance and whether the residents' health records have been established. Social integration was measured by the migrants' local social network (whether they interacted most with other locals), participating in non-profit organizations (whether they participated in local activities of labor unions, volunteer associations, hometown associations, hometown associations, hometown chambers of commerce, and other organizations), and public affairs. Psychological integration was measured by migrants' sense of belonging to the city, trust, discrimination, psychological identity, and perceived social attitudes, and eight related questions were included.

Table 2 presents a comparison of the status of urban integration among migrants with different characteristics, such as different household registration and whether they are inter-provincial or get higher education. First, the overall urban integration level of migrants with agricultural household registration is lower and significantly lower than that of migrants with non-agricultural household registration. 22.28% of the migrant agricultural household members live in self-purchased or self-built houses, much lower than non-agricultural household members (43.69%). Regarding health integration, the participation rate of the non-agricultural household registration migrants is higher than that of the agricultural household registration migrants in all indicators, especially in comparing urban residents' medical insurance participation. The gap is vast (the latter is about six times higher than the former).

Table 2.

Comparative analysis of urban integration of migrants with various types of agricultural household registration (%).

Integration dimension	Specific variables	Agriculture	Non-agricultural	Provincial	cross-provincial	Low education level	High level of education
Economic integration	No current difficulties with too low income locally	14.59	15.60	13.25	16.55	14.72	16.97
	Living in a self-built/purchased home	22.28	43.69	30.93	21.51	24.75	46.88
	Currently working in a stable local job	33.12	30.95	31.93	33.93	32.94	32.57
Health integration	Established resident health records	26.09	30.95	31.81	22.63	27.15	29.17
	Participated in urban residents' medical insurance	3.59	23.06	7.47	6.42	6.86	8.13
	Applied for a personal social security card	45.53	69.16	50.66	49.36	47.80	79.80
Social inclusion	Like the city/place where you live now	97.19	97.80	97.59	97.08	97.32	97.57
	Concerned about the changes in the city/place of residence now	94.92	96.98	95.99	94.63	95.17	97.42
	Most interaction with local people in spare time	29.60	46.84	40.41	24.82	31.45	49.75
	Have participated in union activities locally	6.77	16.36	8.55	8.35	6.98	28.14
Psychological integration	No feeling that local people look down on foreigners	81.78	87.73	85.97	80.03	82.52	90.07
	Would love to integrate and be a part of the local community	92.38	96.34	94.85	91.38	92.84	97.13
	Thought the locals would accept me as one of them	92.10	96.03	94.82	90.91	92.60	96.84
	Hygiene habits do not differ much from those of local citizens	78.68	85.02	83.31	76.83	79.42	89.46

Note: Participation in public affairs is measured by participation in union activities. Regarding psychological integration, some of the secondary indicator questions investigate the level of agreement, with 1 = completely disagree, 2 = disagree, 3 = agree, and 4 = completely agree, with "agree" and "completely agree" considered as agree and others as disagree. The percentages of an agreement are reported in the table.

Regarding social integration, the migrants with agricultural household registration interact less with locals and rarely participate in local public affairs such as trade union activities. In terms of psychological integration, although more than 90% of the migrant agricultural household members like and care about the city where they work and are willing to integrate with the locals, about 19% of the migrants still feel discrimination from the locals and believe that their hygiene habits differ significantly from those of the locals. Secondly, there is little difference in the proportion of integration in each city for the migrants who choose to work in or outside the province. The in-province migrants are more socially connected to the local population, having a high percentage of 40% compared to the out-of-province migrants. Finally, education level has a significant differentiating effect on the urban integration rate of the migrants. It can be found that the urban integration rate of the migrants with higher education is higher than that of the lower education group to varying degrees, and the urbanization adaptation performance is better. Among them, 79.80% of the migrants have applied for personal social security cards, 28.14% have participated in trade union activities, and nearly 90% think their hygiene habits are not much different from those of local citizens.

4. Theoretical mechanism and research hypothesis

The degree of urban integration is a complex, multidimensional concept that involves multiple aspects including social, economic, cultural, and psychological dimensions. For the migrant population, the degree of urban integration has a significant impact on whether they choose to settle in the migration destination and their willingness to remain there. This article constructs a theoretical framework for the mechanism by which urban integration affects the settlement and willingness to stay of the migrant population, conducting an in-depth analysis from four dimensions: economic, health, social, and psychological.

4.1. Mechanism analysis of economic integration

Economic integration is a key indicator of the adaptability of the migrant population to living and working in a new environment. It significantly influences their intentions to settle and remain through two mediating variables: income level and housing stability. Firstly, the income level, as a direct representation of economic integration, greatly affects the quality of life and future expectations of the migrant population [18]. When individuals find stable employment and secure a relatively high income after relocating, they are better able to afford the local cost of living and accumulate wealth. This sense of economic security inclines them to settle in the new location and to be willing to make long-term investments, such as in real estate or education, which are key indicators of the desire to settle. Secondly, housing stability is another important aspect of the sense of economic integration and a crucial factor in determining whether the migrant population is willing to settle. Obtaining a stable residence, whether through purchase or rental, provides a sense of security and belonging in life. The stability of housing reduces the uncertainties of daily living and increases the individual's willingness to stay and invest in the community, thereby enhancing the motivation to remain. Long-term residence at the same location also promotes the establishment of community connections, further strengthening the individual's desire to settle and remain. In summary, economic integration, by raising the level of income and housing stability, provides the migrant population with more resources and motivation to consider long-term residence and participation in local community life. These factors work together to enhance their intentions to settle and remain, thereby promoting their stability and personal development in the new environment.

Hypothesis 1a

Economic integration has a positive impact on the settlement and retention intentions of the migrant population.

Hypothesis 1b

Economic integration impacts intentions through intermediary mechanisms such as income level and housing stability.

4.2. Mechanism analysis of health integration

Health integration is a key factor for the migrant population when deciding whether to settle and stay in a city, and it is greatly influenced by the degree of public medical participation and the accessibility of medical services. Firstly, public medical participation is demonstrated by the migrant population's ability to enjoy equal access to local medical resources and participate in public health insurance plans. This directly affects their economic burden and psychological stress when dealing with health issues. When the migrant population feels included and protected by the medical system, they are more likely to develop a sense of belonging, which in turn strengthens their intention to settle [31,32]. On the other hand, the accessibility of medical services, including the reachability of hospitals, the convenience of seeking medical treatment, and the timeliness and quality of medical services, are practical considerations for migrant populations looking at long-term residency. If a city can provide efficient and convenient medical services, this will significantly increase the living satisfaction of the migrant population, making them more inclined to reside and develop in the city long-term. In summary, by enhancing public medical participation and the convenience of medical services, the sense of health integration for the migrant population is improved, which positively influences their intentions to settle and remain.

Hypothesis 2a

Health integration has a positive impact on the settlement intentions and willingness to remain of the migrant population.

Hypothesis 2b

Health integration affects intentions through mediating mechanisms such as public medical participation and convenience of medical services.

4.3. Mechanism analysis of social integration

Social integration is a key social factor for the migrant population in their decision to settle and their willingness to remain, primarily functioning through the two mediating variables of social status and local attachment. Social status reflects the position of the migrant population within the urban social structure, including job stability, income level, career development opportunities, and the establishment of social networks. If migrants can achieve higher social status in the city, they typically gain better economic security and a wider range of social resources, which contribute to improving their quality of life, thereby strengthening their desire to stay in the city. For instance, stable employment and income can enhance their sense of well-being in the city, making the migrant population more inclined to settle there [10]. Local attachment refers to the emotional connection and sense of identification that the migrant population has with the city. This includes a sense of belonging to the community, identification with urban culture, and adaptation to local lifestyles. When migrants establish strong local attachments in the community, they are more likely to develop deep emotional ties, which inspire their settlement intentions and decisions to remain. The openness, inclusiveness, and cultural diversity of a city usually facilitate the formation of local attachments.

Hypothesis 3a

Social integration has a positive impact on the settlement intentions and willingness to remain of the migrant population.

Hypothesis 3b

Social integration influences intentions through mediating mechanisms such as social status and local attachment.

4.4. Mechanism analysis of psychological integration

Psychological integration is a key psychological process for the migrant population to adapt to a new environment and form intentions to settle or remain. Identity belonging and psychological assimilation, as mediating variables, play an essential role. Firstly, a sense of identity belonging refers to the social acceptance and recognition that individuals feel in a new environment. If migrants can establish a strong sense of identity belonging in the place of relocation—that is, feeling like they are part of the community or local society—this feeling significantly enhances their sense of security and satisfaction, thus strengthening their willingness to settle and remain. This process can be achieved through participating in community activities, building local social networks, and embracing local cultural traditions. An increased sense of belonging makes migrants more likely to see the new environment as their "home" and more likely to make long-term residential decisions [13]. Secondly, psychological assimilation refers to the extent to which migrants blend in with the host society in terms of cultural identity, values, and lifestyles. When migrants share more in common with local residents in these aspects, they feel psychologically closer to the local society, which makes it easier for them to develop the intention to stay. Psychological assimilation includes learning the local language, following local social norms and cultural customs, and accepting and understanding the local values and belief systems. Such assimilation helps to reduce cultural conflicts, promote social harmony, and make migrants feel more comfortable and at ease in a new environment.

Hypothesis 4a

Psychological integration has a positive impact on the settlement intentions and willingness to remain of the migrant population.

Hypothesis 4b

Psychological integration influences intentions through mediating mechanisms such as identity belonging and psychological assimilation.

4.5. Heterogeneous effects based on multi-factor analysis

Urban integration is an important indicator measuring how well the migrant population adapts to urban life and forms intentions to settle and remain. In this process, socio-demographic factors such as the hukou system, gender, education level, and marital status have significant moderating effects on the urban integration and settlement decisions of the migrant population. The hukou, as a social institution, is directly linked to an individual's ability to enjoy urban public services and welfare, including education, healthcare, and employment. Different hukou statuses may lead to varying degrees of institutional barriers during the process of urban integration. For example, a non-local hukou may limit an individual's access to urban resources, decreasing their willingness to remain. Gender is another important moderating variable. Due to differences in societal gender roles and expectations, male and female migrants may have different experiences and perceptions in terms of employment opportunities, social network formation, and family responsibility distribution, which in turn affect their sense of belonging to the city and willingness to stay. Education level is usually closely related to an individual's employment opportunities and socio-economic status. A higher level of education helps migrants secure better job opportunities and integrate more easily into the upper echelons of urban social structures, thereby strengthening their intention to settle in the city. Migrants with lower levels of education may encounter more difficulties during urban integration, affecting their settlement and retention decisions. Marital status can also moderate the urban integration and willingness to remain of the migrant population. Married individuals often need to consider factors such as their spouse's and children's education and employment, and these family responsibilities may prompt them to prefer long-term settlement in the city. Single or divorced migrants, when deciding whether to stay in the city, may be influenced more by personal life choices.

Hypothesis 5

In the relationship between urban integration and the settlement and retention intentions of the migrant population, various factors such as hukou, gender, education level, and marital status have a moderating effect that produces heterogeneous impacts.

In order to present the theoretical analysis framework more clearly, this paper draws the specific analysis framework as shown in Fig. 3. By combining existing literature and theories, the paper divides the urban integration into four juxtaposed levels: economic (job stability, income sustainability, residential stability), health (health care inclusion, personal social security inclusion), social (urban favorability, urban change concern, local social closeness), and psychological integration (local acceptability, subjective positivity, custom compatibility). In addition, existing literature suggests that other factors such as individual characteristics (e.g., age, family size, number of moves, income, health status, etc.) and regional characteristics (e.g., level of economic development, level of openness, industrial structure) have impacts on the intention of the migrants to settle and reside in the city. Hence, these factors are incorporated into the theoretical framework (Fig. 3).

Fig. 3.

Theoretical framework diagram.

5. The impact of urban integration on the intention of settlement and residence of the migrants with agricultural household registration

5.1. Empirical design

This paper constructs an econometric model of the effect of urban integration on the intention of settlement and the intention of residence of the migrants with agricultural registration and performs regression analysis using the Logistics method (Eq. (1) and Eq. (2)).

$$Settl{e}_i={\alpha }_{1}Integratio{n}_i+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (1)

$$Residenc{e}_i={\alpha }_{1}Integratio{n}_i+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (2)

Where the dependent variables $Settl{e}_i$ and $Residenc{e}_i$ denote the intention of settlement and the intention of residence among migrants, respectively. The core independent variable, $Integratio{n}_i$, indicates the degree of urban integration, and $IndiCon{t}_i$ indicates the individual-level control variables, which consist of age level ($Ag{e}_i$), the number of household member ($Househol{d}_i$), the number of cities migrated ($Flo{w}_i$), personal income level ($Incom{e}_i$) and health status ($Sicknes{s}_i$). $SociCon{t}_j$ denotes regional level control variables, consisting of the economic level of the region ($Pgd{p}_j$), the degree of economic openness ($Ocea{n}_j$) and the industrial structure ($Str{u}_j$). ${\eta }_i$ and ${\epsilon }_i$ denote the individual fixed effect and the stochastic error term, respectively. See Table 3 for details.

Table 3.

Variable measures and data sources.

Variable Symbols	Variable Meaning	Metrics	Data source
Dependent variable
Settle	Intention of settlement	Survey item "If you meet the local settlement conditions, are you willing to move your household to the local area", if yes, then 1, otherwise 0.	CMDS 2017
Residence	Intention of residence	The survey item "Do you plan to stay in the local area for some time to come", if yes then 1, otherwise 0.	CMDS 2017
Core independent variable
Integration	Degree of urban integration	The variables of economic integration, security integration, social integration, and psychological integration were combined and measured by the entropy method	CMDS 2017
Control variables (individual level)
Age	Age Level	Measured by subtracting the survey item "year of birth" from 2017.	CMDS 2017
Household	The number of household member	The survey item "Number of family members living together"	CMDS 2017
Flow	The number of cities migrated	The survey item "Total number of cities you have moved to (including survey sites)".	CMDS 2017
Income	Personal income	The survey item "Your personal wage income/net income for the last month (or last employment)".	CMDS 2017
Sickness	Health status	The survey item "Your health status".	CMDS 2017
Control variables (regional dimension)
Pgdp	Economic level	Ratio of GDP to population	2018 China Urban Statistical Yearbooka
Ocean	Level of economic openness	The inflow area belongs to the coastal province is 1, otherwise it is 0.	China Coastal Cities Economic Statistics Database
Stru	Industrial structure	Ratio of the sum of the output of the secondary and tertiary sectors to the total regional output.	2018 China Urban Statistical Yearbook

^aAccording to the practice of the Chinese statistics department, the 2018 statistical yearbook contains information from the 2017 data.

5.2. Measure of urban integration: entropy method

The entropy method is an objective assignment method that fully considers the numerical information provided by each indicator and can better resolve the inherent conflicts among embedded criteria in multi-attribute decision problems to balance the relationships among many evaluation criteria to calculate more objective relative weights of indicators and make the evaluation results more accurate and reasonable [33]. Compared with other objective weighting methods, such as principal component analysis, the entropy method can overcome the limitation that individual weights deviate from the regular interval due to the interference of outliers, so it is widely used to determine total evaluation index weights in engineering technology and social economy. Therefore, it is widely used to determine the weights of comprehensive evaluation indexes in engineering and social economy. Generally, the smaller the entropy value of an indicator, the greater the degree of variation of its observed value, and the more information it provides, the stronger its importance in the comprehensive evaluation and the larger its weight, and vice versa (Table 4).

Table 4.

The weights of each secondary index calculated by the entropy weight method.

First-level Indicators	Second-level indicators	Weights
Economic integration	No current difficulties with too low income locally	0.4378
	Living in a self-built/purchased home	0.3069
	Currently working in a stable local job	0.2553
Health integration	Established resident health records	0.2788
	Participated in urban residents' medical insurance	0.5724
	Applied for a personal social security card	0.1487
Social integration	Like the city/place where you live now	0.0074
	Concerned about the changes in the city/place of residence now	0.0131
	Most interaction with local people in spare time	0.3050
	Have participated in union activities locally	0.6746
Psychological integration	No feeling that local people look down on foreigners	0.3364
	Would love to integrate and be a part of the local community	0.1287
	Thought the locals would accept me as one of them	0.1335
	Hygiene habits do not differ much from those of local citizens	0.4014

Based on the index weights determined using the entropy value method, this paper further measures the level of urban integration level in China. The economic integration, Health integration, social integration, and psychological integration variables are first dimensionless processed, and then the processed variables are multiplied with the corresponding weights to calculate the urban integration degree indicators comprehensively (Eq. (3) and Eq. (4)).

$$Z_i=\frac{X_i-X_i^{\min }}{X_i^{\max }-X_i^{\min }}$$ (3)

$$I_i=\frac{{\sum }_{i=1}^n{Z}_i{\varpi }_i}{{\sum }_{i=1}^n{\varpi }_i}\times 100$$ (4)

5.3. Descriptive statistics

The essential descriptive statistical characteristics of each variable used in this study are presented in the table. The variables' observed values, mean, standard deviation, median, minimal, and maximal values are reported in order (Table 5).

Table 5.

Variable descriptive statistics.

Variable Type	Variable	Observation	Mean	SD	p50	Min	Max
Dependent variable	Settle	169989	0.390	0.490	0	0	1
	Residence	169989	0.830	0.380	1	0	1
Core independent variables	Integration	169989	7.410	1.790	7	0	14
Individual-level control variables	Age	169989	36.66	11.07	35	15	96
	Household	169989	3.140	1.200	3	1	10
	Flows	169989	1.970	1.900	1	1	92
	Income	169989	4328	3874	3500	−180000	200000
	Sickness	169989	1.210	0.470	1	1	4
Regional-level control variables	Pgdp	169989	6.330	2.840	5.100	2.910	13.62
	Ocean	169989	0.240	0.420	0	0	1
	Stru	169989	35.21	71.77	10.41	3.150	296.2

6. Empirical analysis and testing

6.1. Baseline empirical results

Columns (1) to (3) of Table 6 report the overall regression results of the effect of urban integration on the intention of settlement among the migrants, incorporating individual-level control variables, regional-level control variables, and all control variables in that order. The estimated coefficients of the core variables indicate that the degree of urban integration significantly enhances the intention of the migrants to settle in households. Factors such as age, family size, number of mobile cities, and illness negatively affect the intention of the migrants to settle in a household, and individuals with higher income levels, as well as inflow areas with higher levels of economic development or economic openness, significantly increase the intention of settlement in a household. The structure of secondary and tertiary industries with higher added value is also more attractive for the migrants.

Table 6.

Regression results of the impact of urban integration on the settlement and residence intention of the migrants.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Settle			Residence
Integration	0.156***	0.183***	0.183***	0.268***	0.293***	0.279***
	(0.003)	(0.003)	(0.003)	(0.004)	(0.004)	(0.004)
Constant	−1.416***	−2.211***	−2.040***	−0.661***	−0.922***	−1.051***
	(0.037)	(0.029)	(0.043)	(0.045)	(0.035)	(0.051)
Control Variables	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control
Chi-Square	4188	13823	12135	6129	7335	6896
Log Likelihood	−91020	−106775	−87046	−61802	−74767	−61418
Pseudo-R2	0.0225	0.0608	0.0652	0.0472	0.0468	0.0532
Observation	139842	169989	139842	139842	169989	139842

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

Columns (4) to (6) report the overall regression results of the effect of urban integration on the intention of residence of the migrants, and the model structure is similar to the previous paper. The estimated coefficients of the core variables indicate that a higher degree of urban integration is conducive to a higher intention of residence. Factors such as age and illness are not conducive to an increased intention of residence, and individuals with larger family sizes and higher income levels show a stronger intention of residence. Inflows with higher economic development, a more open economy, and a more advanced industrial structure are attractive to the migrants for residency.

6.2. Testing mediating effects: a multi-dimensional analysis based on urban integration

6.2.1. Model specification and the introduction of stepwise regression

Urban integration is reflected by four dimensions: economic, health, social, and psychological, hence the mechanisms influencing the migrant population's intentions to settle and remain are determined by factors across these varied dimensions. The article constructs an econometric model based on the framework of stepwise regression (Eq. (5), Eq. (6) and Eq. (7)).

$$Settl{e}_i/Residenc{e}_i={\alpha }_{1}Integration\_X_i+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (5)

$$M_i={\alpha }_{1}Integration\_X_i+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (6)

$$Settl{e}_i/Residenc{e}_i={\alpha }_{1}Integration\_X_i+M_i+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (7)

Where $Integration\_X_i$ represents the four dimensions of urban integration, namely economic integration, $Integration\_Ec{o}_i$, health integration, $Integration\_Heat{h}_i$, social integration, $Integration\_Socia{l}_i$, and psychological integration, $Integration\_Psyc{h}_i$. $M_i$ serves as the mediating variable, which includes income level, $Incom{e}_i$, and housing stability, $House\_s{t}_i$, under the economic integration dimension, public medical participation, $Me{d}_i$, and medical service accessibility, $Avai{l}_i$, under the health integration dimension, social status, $Socia{l}_i$, and local attachment, $Local{d}_i$, under the social integration dimension, identity belonging, $Identit{y}_i$, and psychological assimilation, $Assim{i}_i$, under the psychological integration dimension. The meanings of other variables are consistent with the baseline model.

6.2.2. Mediation effect regression results

Based on the analysis of the theoretical mechanisms mentioned earlier, this paper sequentially conducts tests of mediating effects for each of these dimensional factors.

Firstly, the test for the mediating effect of economic integration. Looking at the results from Table 7, the estimated coefficients for mediating variables such as income level and housing stability are statistically significant at the 1% level. Economic integration has a significant positive impact on the intentions of the migrant population to settle and remain. Comparing the results, after adding variables such as income level and housing stability, the estimated coefficients for economic integration variables drop to 0.019 and 0.117, respectively, confirming the presence of mediating effects. Hypotheses 1a and 1b have been confirmed.

Table 7.

The test results for the mediating effect of economic integration.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Settle	Income	House_st	Settle	Residence	Income	House_st	Residence
Integration_eco	0.064***	0.082***	1.064***	0.019**	0.283***	0.082***	1.064***	0.117***
	(0.01)	(0.00)	(0.01)	(0.01)	(0.01)	(0.00)	(0.01)	(0.01)
Income				0.126***				0.257***
				(0.01)				(0.01)
House_st				0.379***				0.881***
				(0.01)				(0.02)
Constant	−0.642***	7.865***	−1.801***	−1.686***	0.943***	7.865***	−1.801***	−1.116***
	(0.03)	(0.01)	(0.04)	(0.10)	(0.04)	(0.01)	(0.04)	(0.11)
Control Variables	Control	Control	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control	Control	Control
Chi-Square	9977	1626	844.8	10024	3013	1626	844.8	3240
Log Likelihood	−88125	−13458	−68871	−88102	−63360	−13458	−68871	−63246
Pseudo-R2	0.0536	0.0570	0.00610	0.0538	0.0232	0.0570	0.00610	0.0250
Observation	139842	139773	139842	139773	139842	139773	139842	139773

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

Secondly, the test for the mediating effect of health integration. According to Table 8, the estimated coefficients for mediating variables such as Med and Avail are statistically significant at the 1% level, with Avail being significantly negative, indicating that the stronger the accessibility to medical services, the stronger the intention to settle and remain. Health integration has a significant positive impact on the intentions of the migrant population to settle and remain. Looking at the results after integrating variables such as public medical participation and accessibility of medical services, the estimated coefficients for health integration variables dropped to 0.229 and 0.261, respectively, confirming the presence of mediating effects of health integration. Hypotheses 2a and 2b have been confirmed.

Table 8.

The test results for the mediating effect of health integration.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Settle	Med	Avail	Settle	Residence	Med	Avail	Residence
Integration_health	0.232***	0.772***	−0.074***	0.229***	0.271***	0.772***	−0.074***	0.261***
	(0.01)	(0.02)	(0.01)	(0.01)	(0.01)	(0.02)	(0.01)	(0.01)
Med				0.218***				0.458***
				(0.04)				(0.06)
Avail				−0.052***				−0.199***
				(0.01)				(0.02)
Constant	−0.798***	−2.686***		−0.872***	0.835***	−2.686***		1.070***
	(0.03)	(0.12)		(0.04)	(0.04)	(0.12)		(0.05)
Control Variables	Control	Control	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control	Control	Control
Chi-Square	9977	1626	844.8	10024	3013	1626	844.8	3240
Log Likelihood	−88125	−13458	−68871	−88102	−63360	−13458	−68871	−63246
Pseudo-R2	0.0536	0.0570	0.00610	0.0538	0.0232	0.0570	0.00610	0.0250
Observation	139842	139842	139842	139842	139842	139842	139842	139842

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

Thirdly, the test for the mediating effect of social integration. According to Table 9, the estimated coefficients for mediating variables such as Social and Locald are statistically significant at the 1% level. Social integration has a significant positive impact on the intentions of the migrant population to settle and remain. Comparing the results, after taking into account variables such as social status and local attachment, the estimated coefficients for the health integration variables respectively decreased to 0.336 and 0.442, demonstrating that the presence of mediating effects of social integration has been validated. Hypotheses 3a and 3b have been confirmed.

Table 9.

The test results for the mediating effect of social integration.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Settle	Social	Locald	Settle	Residence	Social	Locald	Residence
Integration_health	0.385***	0.848***	0.554***	0.336***	0.528***	0.848***	0.554***	0.442***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Social				0.249***				0.235***
				(0.02)				(0.03)
Locald				0.430***				0.230***
				(0.02)				(0.01)
Constant	−1.686***	−2.794***		−1.671***	−0.376***	−2.794***		−0.324***
	(0.04)	(0.06)		(0.04)	(0.05)	(0.06)		(0.05)
Control Variables	Control	Control	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control	Control	Control
Chi-Square	10900	9233	3328	11793	4488	9233	3328	4078
Log Likelihood	−87664	−45324	−55080	−87217	−62622	−45324	−55080	−58720
Pseudo-R2	0.0585	0.0924	0.0293	0.0633	0.0346	0.0924	0.0293	0.0336
Observation	139842	139842	139842	139842	139842	139842	139842	120202

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

Fourthly, the test for the mediating effect of psychological integration. According to Table 10, the estimated coefficients for mediating variables such as Identity and Assimi are statistically significant at the 1% level. Psychological integration has a significant positive impact on the intentions of the migrant population to settle and remain. From the comparison of results, after including variables such as identity affiliation and psychological assimilation, the estimated coefficients for psychological integration variables dropped to 0.117 and 0.278, respectively, indicating that the presence of mediating effects of psychological integration has been confirmed. Hypotheses 4a and 4b have been confirmed.

Table 10.

The test results for the mediating effect of psychological integration.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Settle	Identity	Assimi	Settle	Residence	Identity	Assimi	Residence
Integration_psych	0.376***	0.720***	2.007***	0.117***	0.428***	0.720***	2.007***	0.278***
	(0.01)	(0.01)	(0.02)	(0.01)	(0.01)	(0.01)	(0.02)	(0.01)
Identity				0.521***				0.448***
				(0.02)				(0.02)
Assimi				1.631***				0.454***
				(0.04)				(0.03)
Constant	−2.105***	−0.872***	−3.386***	−3.188***	−0.671***	−0.872***	−3.386***	−0.934***
	(0.05)	(0.05)	(0.09)	(0.05)	(0.05)	(0.05)	(0.09)	(0.05)
Control Variables	Control	Control	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control	Control	Control
Chi-Square	11488	17241	29611	15406	4974	17241	29611	6087
Log Likelihood	−87370	−70218	−21333	−85411	−62380	−70218	−21333	−61823
Pseudo-R2	0.0617	0.109	0.410	0.0827	0.0383	0.109	0.410	0.0469
Observation	139842	139842	139842	139842	139842	139842	139842	139842

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

6.3. The endogeneity problem treatment: the introduction of diversified measurement methods

In the existing literature, when studying the relationship between urban integration and the intention of migrant workers to settle and stay, the endogeneity problem is less considered, and there is a lack of solutions to explore the problem. In order to alleviate the biased estimation results caused by endogeneity problems, this paper has selected as many factors affecting migrant workers' urban psychological integration as possible. However, due to the difficulty of measurement and the lack of relevant data, there may still be some omitted variables, which bring about endogeneity problems. For example, at the individual level, factors such as an individual's ability level and local sentiment may affect their intention of settlement and residence in the city; at the policy level, the settlement policy and children's enrollment threshold in the inflow area may also affect their intention of settlement and residence in the city, but these factors are difficult to be captured effectively. At the same time, there may be a reverse causal relationship between urban integration and intention of settlement and residence in the city: on the one hand, a more vital intention of settlement and residence in the city will lead to the establishment of a stable social network, a fixed living area, and regular participation in activities, resulting in a higher sense of urban integration; on the other hand, a higher degree of urban integration means that individuals already have a certain amount of human and social capital.

On the other hand, a higher degree of urban integration means that individuals have specific human capital and social capital, have more robust adaptability, can better integrate into urban labor organizations, have more employment opportunities and higher living standards, and are more willing to settle and reside in the city. The presence of omitted variables and reverse causality may lead to endogeneity in the intention of settlement and residence in the city, thus biasing the estimated coefficients of the core variables. Therefore, this paper uses a combination of analytical methods to mitigate the bias in the estimation results caused by the endogeneity of the model.

First, this paper uses the instrumental variable Probit (IV Probit) method for regression. IV Probit method fits models for binary dependent variables where one or more covariates are endogenous, and errors are normally distributed. This paper uses Newey and West's [34] minimum Chi-Squared (two-step) estimator for great likelihood estimation. However, the IV Probit method assumes that the endogenous covariates are continuous and there are also discrete endogenous variables in the model, so they are inappropriate for use with discrete endogenous covariates.

Second, since the variables measuring the degree of urban integration are binary dummy variables, which are difficult to handle by the conventional two-stage instrumental variables approach [35], this study uses the conditional mixed process estimation of instrumental variables (CMP) proposed by Roodman [36] to overcome the potential endogeneity problem in the empirical model.

Applying the first two endogeneity treatments shares a standard premise of finding the appropriate instrumental variables. In this paper, two more ideal instrumental variables are selected. The first one is the urban unemployment rate ($unemrat{e}_j$). On the one hand, employment is a fundamental prerequisite for the urban integration of migrants, and the urban unemployment rate can directly reflect the degree of urban integration of the migrants. A higher unemployment rate in a city indicates that the economic development and employment security system of the city may be problematic to a certain extent, which in turn makes it more likely that the short duration of labor contracts, employees leaving on their initiative, and workplace dismissals will lead to job instability, which will occur not only among the resident population of the city but also to a certain extent among the migrants flowing into the city. It affects the latter's urban integration.

On the other hand, the urban unemployment rate may indirectly affect the intention of settlement and residence in the city by affecting the degree of urban integration. The variable of the urban unemployment rate is measured by the average unemployment rate in the last four years in the inflow area. In this paper, the unemployment rate of each city in 2014, 2015, 2016, and 2017 are averaged as the average unemployment rate of the inflowing cities. The unemployment rate is calculated by the ratio of the number of urban registered unemployed persons in the city of individual inflow to the number of urban employees at the end of the year.

The second one is the proportion of residence permit applicants ($residcar{d}_j$). Regarding correlation, migrants in the same inflow area tend to imitate and learn from each other, resulting in a behavioral "cohort effect". It has been found in the literature that migrants tend to exhibit "homogeneous" behavioral characteristics in cities. Obtaining a residence permit for other migrants in the same influx area would lead to the same strategy for a specific migrant. However, from the perspective of exogenousness, the application or holding of residence permits by other migrants in the same inflow area is not directly related to the settlement and residence intention of the migrant.

Finally, this paper uses the Heteroskedasticity-Based Instruments (HBI) method. This method can construct instrumental variables based on the data and applies to the case where the endogenous variables or dependent variables are binary and is considered an effective method to deal with endogeneity problems. The basic logic is as follows (Eq. (8) and Eq. (9)):

$$Y={\alpha }_{1}Z+{\alpha }_{2}X+\epsilon ,\epsilon =\eta D+I_1$$ (8)

$$X={\alpha }_{3}Z+\upsilon ,\upsilon =\phi D+I_2$$ (9)

Where $Y$ denotes the dependent variable, which is the intention of settlement and residence, respectively. $X$ denotes the core independent variable, which is the degree of urban integration. $Z$ denotes the exogenous variable, which is the control variable in this paper. $D$ denotes unobservable factors, and $I_1$ and $I_2$ denote error terms. When the stochastic error term of Eq. (6) satisfies heteroskedasticity, the product of the regression residuals of the endogenous variables on the other exogenous variables and the exogenous variables after centrality treatment can be used as the instrumental variable for the degree of urban integration.

The regression results based on the three endogeneity treatments mentioned above are shown in Table 11. In the first stage regression results of each model, the unemployment rate has a significant negative impact on urban integration, and the residence permit registration rate has a significant positive impact on urban integration. At a technical level, the endogeneity parameters Atanhrho-12 for the CMP approach are −0.001 and −0.005, respectively, both statistically insignificant, indicating no severe endogeneity problem in the baseline model of this paper. In addition, the statistics used to test for weak instrumental variables (Cragg-Donald Wald F-statistic) are all greater than the critical value of 16.38 at the 10% level, indicating rejection of the original hypothesis of weak instrumental variables. The above results indicate that the selected instrumental variables are reasonable and valid. From the second-stage regression results of each model, urban integration has a significant positive contribution to the intention of settlement and stay of the migrants, further confirming the robustness of the core findings of this paper. The significance and sign of the estimated coefficients of the control variables are also consistent with the benchmark results.

Table 11.

Regression results of endogeneity processing.

Methods	IV Probit Method	CMP method	HBI Method	IV Probit Method	CMP method	HBI Method
	(1)	(2)	(3)	(4)	(5)	(6)
	Phase II	Phase II	Phase II	Phase II	Phase II	Phase II
Variable	Settle			Residence
Integration	0.112***	0.039***	0.116***	0.157***	0.039***	0.267***
	(0.002)	(0.001)	(0.005)	(0.002)	(0.001)	(0.014)
	Phase I	Phase I	Phase I	Phase I	Phase I	Phase I
Dependent variable	Integration			Integration
Instrumental Variables
unemrate	−0.015***	−0.017**	−0.038***	−0.041***	−0.057***	−0.062**
	(0.004)	(0.021)	(0.112)	(0.101)	(0.013)	(0.024)
residcard	0.864**	0.850***	0.720***	0.791***	0.452***	0.445***
	(0.308)	(0.008)	(0.018)	(0.022)	(0.129)	(0.184)
Control variables	Control	Control	Control	Control	Control	Control
Atanhrho-12		−0.001			−0.005
Wald chi-squared value		30994.37			52319.20
Cragg-Donald Wald F-statistic	49.416	81.491	53.139	109.383	82.941	72.421
Observations	169989	169989	169989	169989	169989	169989

Note: Clustering robust standard errors are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

6.4. Heterogeneity effects: an examination based on multifactorial moderating effects

From the above benchmark analysis, it is concluded that the degree of urban integration significantly contributes to the increase of the migrants' intention of settlement and residence in a household, but some individual factors that cannot be ignored also play a role. In this paper, we select essential factors such as household registration, gender, education level, and marital status as entry points and use the intersection and multiplication method to capture and examine the possible moderating effects of urban integration on the relationship between the intention of settlement and residence of migrants. The basic structure of the interaction model is constructed as follows (Eq. (10) and Eq. (11)):

$$Settl{e}_i={\alpha }_1\left(Integratio{n}_i\times Adjus{t}_i\right)+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (10)

$$Residenc{e}_i={\alpha }_1\left(Integratio{n}_i\times Adjus{t}_i\right)+{\beta }_{1}IndiCon{t}_i+{\beta }_{2}SociCon{t}_j+{\eta }_i+{\epsilon }_i$$ (11)

Where $Adjus{t}_i$ denotes the moderating variable, which is composed of variables such as household registration, gender, education level, and marital status. $Integratio{n}_i\times Adjus{t}_i$ denotes the cross-product term, which is used as the core variable in Eq. (5) and Eq. (6). Other variables and symbols indicate the meanings consistent with the baseline model.

Table 12 reports the regression results of the moderating effects of the factors of household registration, gender, education level, and marital status, with control variables consisting of individual-level variables and region-level variables. The estimated coefficients from columns (1) to (4) show significant positive moderating effects, indicating that migrants who are male, have higher education levels, and are married, further enhance their intention of settlement; the moderating effect of the household registration factor is negative, indicating that the migrants with agricultural household registration have a more vital intention of settlement. From the results of the estimated coefficients in columns (5) to (8), all four factors show significant positive moderating effects, indicating that the migrants with non-agricultural household registration, male, higher education level, and marriage status show a stronger intention of residence. Hypotheses 5 has been confirmed.

Table 12.

Heterogeneous regression results of urban integration on the settlement and residence intention of migrants.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Aspects	Household	Gender	Education	Marriage	Household	Gender	Education	Marriage
Variables	Settle	Settle	Settle	Settle	Residence	Residence	Residence	Residence
Integration_agri	−0.044***				0.022***
	(0.002)				(0.002)
Integration_gender		0.014***				0.031***
		(0.002)				(0.002)
Integration_edu			0.081***				0.082***
			(0.002)				(0.003)
Integration_mari				0.061***				0.120***
				(0.002)				(0.003)
Constant	−0.393***	−0.654***	−1.011***	−0.578***	0.887***	0.952***	0.701***	1.228***
	(0.035)	(0.034)	(0.035)	(0.034)	(0.043)	(0.041)	(0.042)	(0.041)
Control Variables	Control	Control	Control	Control	Control	Control	Control	Control
Individual fixed effects	Control	Control	Control	Control	Control	Control	Control	Control
Chi-Square\	9655	9094	11083	9913	2306	2479	3243	4544
Log Likelihood	−88286	−88567	−87572	−88157	−63713	−63627	−63245	−62595
Pseudo-R2	0.0518	0.0488	0.0595	0.0532	0.0178	0.0191	0.0250	0.0350
Observation	139842	139842	139842	139842	139842	139842	139842	139842

Note: Clustering robust standard errors at the city level are reported in parentheses, ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, and the models all control for urban fixed effects.

6.5. Decomposition of the contribution of urban integration to the impact of migrants' intention of settlement and stay in the city

In the previous study, the level of urban integration significantly contributes to the intention of settlement and residence among the migrants. The variable of urban integration is composed of four dimensions, including economic, health, social, and psychological integration, so it is not difficult to infer that the latter also has different degrees of contributing effects on the intention of settlement and residence. Further, in order to compare the differences in the contribution of different dimensions of urban integration levels to the impact of the intention of settlement and residence of the migrants, the Modified Dissimilarity Index (MDI) provided by Juárez and Soloaga [37] was introduced to calculate the value of inequality of opportunity (IOP) for the intention of settlement of the migrants and further decomposed it using the Shapley value decomposition method, which in turn allows the contribution of each dimension of urban integration to the differences in the intention of settlement and residence.

Fig. 4 presents the differences in the contribution of the four dimensions of urban integration level to the intention of the migrants to settle. Overall, psychological integration has a much more significant impact on the intention of settlement of migrants (50.18%) than social integration (27.46%), health integration (20.49%), and economic integration (1.87%). Positive psychological integration, including the sense of respect gained, a willingness to integrate locally, perceived acceptance, and similarity to locals regarding health habits, all contribute to the ultimate identity integration, i.e., the intention of settlement of migrants in all categories of agricultural households. In contrast, perceived discrimination, differences in hygiene habits, and conformity to the customs and practices of the old home by the migrants hinder mutual understanding and contact between migrants and natives, reinforce the migrants' hometown complex, and weaken the intention of settlement for migrants with cross-provincial and agricultural household. In comparing different characteristics, the psychological integration dimension has a significant impact; social and health integration have different contribution rates in household registration, gender, education level, and marital status. For instance, the impact of health integration on the intention of settlement for migrants with non-farm household registration is minimal. The migrants with higher education are more concerned about health integration and less sensitive to social integration, probably because more employment options make this group not too much pressure from society.

Fig. 4.

Contribution decomposition of the influence dimension of migrants' intention of settlement (%).

Fig. 5 presents the divergence of the contribution of the four dimensions of urban integration level on the intention of residence of the migrants. Overall, psychological integration has a much more significant impact on the intention of settlement (42.11%) than social integration (25.81%), economic integration (20.78%), and social security integration (11.29%). Similar to the contribution status of the intention of settlement, positive psychological integration plays an essential role in the intention of residence of the migrants. In addition to the psychological integration dimension, the intention of residence is more influenced by social and economic integration than the intention of settlement. On the one hand, economic integration significantly influences whether the migrants can live in the local area in a long-term and stable manner. The high cost of health care, education, and living in large cities also places higher demands on the earning capacity and job treatment of the migrants. In addition, purchasing or building a house is currently the most critical expenditure item for most migrants, so it can be considered the first step for a migrant to relocate and settle in a city.

Fig. 5.

Contribution decomposition of the influence dimension of migrants' intention of residence (%).

On the other hand, social integration has an important impact on the intention of residence. Social integration, such as socializing with locals, and participating in local union activities and local public affairs, also contributes to the intention of residence of migrants. Social interaction with locals is a crucial way to build a local social network, and a more robust social network helps migrants adapt to local habits and integrate into the mainstream culture while also helping them to access more and better resources for doing business or working.

7. Conclusion and discussion

7.1. Main conclusions

The study employed the CMDS from 2012 to 2017 on the migrants. Drawing on theories related to urban integration and labor migration, the research measured the degree of urban integration across four dimensions: economic, health security, social, and psychological. Furthermore, it analyzed the impact of urban integration on the settlement and residential intentions of the migrants. The findings are as follows: (1) There is an inverse trend in settlement and residential intentions among the migrants. Overall, the willingness to settle showed a significant decrease, while the intention to reside steadily increased. Across various demographic groups, those with higher income levels, higher educational backgrounds, non-agricultural household registration, married individuals, and those of lower to middle age exhibit a higher proportion of settlement and residential intentions. (2) Enhanced urban integration contributes to an increased willingness for both settlement and residential choices among the migrants. Factors such as gender, educational level, and marital status play significant positive moderating roles, whereas hukou factors exhibit a negative moderating effect. (3) The test results for the mediating effects indicate that income level and housing stability are important pathways for economic integration, public medical participation and accessibility of medical services are important pathways for health integration, social status and local attachment are important pathways for social integration, and identity affiliation and psychological assimilation are important pathways for psychological integration. (4) Psychological integration emerges as the most influential dimension in urban integration. Among the four dimensions, psychological integration holds the greatest impact on the settlement and residential intentions of the migrants, surpassing social, health security, and economic integration. Groups with higher educational levels display greater concerns for health security integration but are less sensitive to social integration. Residential intentions are more influenced by social and economic integration dimensions, with non-agricultural hukou and higher-educated groups being particularly influenced by the level of economic integration.

7.2. Discussion and implications

Labor mobility is key to activating labor resources in developing countries, balancing labor supply and demand between regions, and subsequently improving labor efficiency. However, the low level of urban integration among migrants is a significant factor leading to their relatively low willingness to settle and reside. This study found that the psychological integration level among China's migrants is generally low, and it significantly contributes to the weak willingness to settle and reside, aligning with findings in similar studies in developing countries [2,13]. Positive psychological integration, including feelings of being respected, willingness to integrate locally, perceived acceptance, and similarity in health habits with local residents, is key to improving the willingness to settle among rural-to-urban migrants. Psychological integration significantly affects various characteristics, and social integration and health security integration have different contributions concerning household registration, gender, education level, and marital status. For example, highly educated migrants focus more on health security integration, which is consistent with the findings of Lao and Gu [16]. Psychological integration plays a crucial role in the willingness of migrants to reside, similar to its contribution to the willingness to settle, reaffirming the viewpoints of most literature [1,2]. The importance of social integration and economic integration has been previously highlighted in the research by Vroome and Tubergen (2014); however, this study's focus includes how both economic and social integration affect the willingness to reside.

The findings of this study offer clear policy implications. Economically, the government should assist the migrant population in skill enhancement through special funds, provide housing subsidies and preferential loans, as well as offer entrepreneurship guidance and financial support to reduce barriers to economic participation. In terms of health, integrating into the public health system, establishing community medical centers, and organizing health education activities are essential to ensure that migrants have access to basic medical services and to improve their health management skills. Socially, fostering social exchange and integration through the operation of multicultural centers and community social activities, as well as urban planning that takes into account the needs of the migrant population, is encouraged. Psychologically, establishing counseling services, conducting social integration workshops, and formulating anti-discrimination regulations are necessary to support the psychological health and social identity of the migrant population. The meticulous implementation of these strategies requires an assessment feedback mechanism to ensure their effectiveness, a commitment to removing barriers, providing equal opportunities, and ensuring that migrants can successfully integrate into urban life economically, health-wise, socially, and psychologically.

This study focused on the Chinese case, and in comparison with cases in developed countries, it is evident that there are many aspects worth considering in developing countries [13,[38], [39], [40]]. First, the importance of inclusiveness and communication should be emphasized to promote cultural integration and reduce differences. Second, reducing the cost of urban living is crucial to enhancing the willingness of migrants to integrate economically. Third, health security contributes to the stability and willingness of migrants to settle. Finally, developing countries need to strengthen the social roles of labor unions, communities, and non-profit organizations, promoting interaction between migrants and local residents to increase the willingness to settle. These improvements will contribute to enhancing the level of urban integration among migrants in developing countries.

The limitations and prospects of this paper are as follows: 1. The empirical regression uses cross-sectional data, which lacks the verification that could be provided by relatively continuous panel data; 2. Besides the mediating variables identified in this article, there may be other significant social, educational, or economic factors that await exploration in subsequent research; 3. This article only utilizes the CMDS large-scale survey database, and there are other representative large-scale survey databases not used in this study for comparative analysis; 4. Research on issues related to immigration is gradually shifting focus towards new urbanization, new infrastructure, or the wealth gap, which have not been systematically discussed in this article.

Ethics declarations

• •Review and/or approval by an ethics committee was not needed for this study because this is an observational social science study that does not involve experiments with humans or animals.
• •Informed consent was not required for this study because the research does not involve experiments on humans or animals, and the data used has been downloaded and utilized in accordance with relevant regulations from the respective official websites.

CRediT authorship contribution statement

Yuanhong Hu: Writing – review & editing, Writing – original draft, Software, Formal analysis, Conceptualization. Yiran Hao: Validation, Project administration. Xianghu Li: Visualization, Supervision, Methodology. Jingdong Luan: Resources, Supervision. Pengling Liu: Writing – review & editing, Investigation, Data curation, Conceptualization.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article.