Early Urban Migration and Long-Term Settlement Intention: Evidence from Migrants in China

Abstract

In recent years, with the further advancement of China’s urbanization process and the continued expansion of its migrants, increasing attention has been directed toward migrants’ intentions to settle and their integration into host cities. However, the role of early urban experiences on migrants’ long-term residence intention remains unclear. Using data from the Chinese Migrant Dynamics Monitoring Survey, we examine the impact of age at first migration on the long-term settlement intention of migrants in China and the underlying influencing mechanisms. We find that migrants who migrate to cities at a younger age have higher intentions to settle in cities. The mechanism analysis suggests that migrating to cities at a younger age facilitates social integration, promotes the convergence of habits and improves dialect proficiency. Moreover, a heterogeneity analysis reveals that the positive effects of early city migration on long-term settlement intentions are more pronounced among migrants who migrate less frequently, are unmarried and have resided in the local area for a longer period. Further study shows that early migration to cities also positively impacts the settlement intention involving the transfer of one’s household registration (hukou) status.

1. Introduction

Urbanization plays a critical role in developing economies and societies [1], driving modernization and improved living standards. In 2023, the world’s urbanization rate reached around 56%. In other words, over half of the population now lives in cities. However, many of the people living in cities are part of the migrant population. In China, there are approximately 376 million migrants—making up 26.6% of the total population. In recent years, following the removal of unequal regulations and restrictions on the urban labor market and rural–urban migration in China, the size of the migrant population has been expected to grow further [2]. As a result, the intention of these migrants to settle permanently in host cities and their intention and ability to integrate into urban society have become crucial. The higher settlement intentions of migrants are vital for fostering social stability, economic development, and labor market sustainability in cities [3]. When migrants want to settle down in a city, they are more likely to invest in housing, education, and local relationships [4], which in turn strengthens their sense of belonging and accelerates their integration into urban life.

Considering the importance of the long-term settlement intentions of migrants for urban and social development, increasing attention has been directed toward migrants’ intentions to settle and their integration into host cities. Previous research has focused mainly on four aspects to explain the settlement intention of migrants. First, there are economic factors. Economic factors are generally considered important drivers influencing migrants’ settlement intentions, including income levels [5], employment opportunities [6], and cost of living [7]. These factors directly affect migrants’ quality of life and economic security in their destination areas. Second, there are social factors. Most studies focus on the key role of social networks [8] and social integration [9] in migrants’ settlement decisions. Third, language and cultural factors play a significant role. Cultural adaptability and language proficiency are believed to greatly influence migrants’ settlement intentions, as they help facilitate smoother integration into the local community [10,11]. Those who can proficiently use the local language and adapt to local culture are more likely to choose to settle long-term in their new environment [11]. Fourth, there is the policy environment. Factors such as migration policies and social welfare systems also significantly influence migrants’ settlement intentions [12,13]. The friendliness of policies and the level of support for migrants can determine their willingness to settle long-term in a particular area. For example, China’s household registration system is considered one of the important reasons hindering migrants’ willingness to reside long-term [12].

The literature review highlights several critical factors influencing the settlement intentions of rural migrants, including economic, social, language, and cultural aspects, as well as the policy environment. These dimensions offer valuable insights into how migrants perceive their quality of life in their new environment. However, a significant gap exists in the literature regarding the relationship between early urban-migration experiences and the long-term settlement intentions of rural migrants. Despite the increasing focus on the factors affecting the settlement intentions of rural migrants, few studies delve into how early migration experiences in urban areas may shape their long-term settlement intentions. This oversight limits our understanding of the long-term settlement decision. Early urban migration may profoundly influence a migrant’s social integration, cultural adaptability, and local dialect ability. Additionally, their initial urban experiences might help rural migrants navigate and adapt to urban lifestyles and habits, thereby increasing their integration potential and settlement intention.

Therefore, this paper emphasizes that early urban experiences, particularly the age at first migration to a city, may significantly impact migrants’ long-term settlement intention in cities. Specifically, we focus on the following two main questions: (1) Does the age at first migration to urban areas affect the settlement intention of migrants? And (2) What are the potential mechanisms underlying this relationship? With the help of the China Migrants Dynamic Survey dataset from 2017, we empirically analyze these questions.

Although numerous studies have investigated the factors influencing migrants’ settlement intentions, these intentions vary significantly in stability and persistence across different migrant groups, city sizes, and regions. Notably, research on the impact of early migration experiences on long-term settlement intentions remains relatively limited. Previous studies primarily focused on the early and middle stages of China’s urbanization, when circular migration between rural and urban areas was predominant, and migrants’ intentions to settle permanently in destination cities were relatively weak [14]. Scholars have examined multiple determinants of migrants’ settlement intentions, with a predominant view that the hukou (household registration) system is the primary, if not the sole, barrier to permanent urban settlement [15,16]. However, a growing body of research challenges this perspective, suggesting it overemphasizes the hukou system’s role [17]. In reality, settlement intentions are shaped by a broader set of factors, including labor market conditions in the destination city, migrants’ human and social capital, family livelihood strategies, and personal characteristics [18,19,20]. This study aims to address this gap by examining how early migration experiences influence migrants’ long-term settlement intentions and exploring the underlying mechanisms. Migrants entering urban areas gain opportunities to enhance their human capital by learning from and emulating high-skilled local workers, which can significantly boost their performance in the labor market [21]. Research shows that the longer migrants accumulate work experience in large cities, the more substantial their wage growth is, underscoring the economic advantages of extended urban exposure [21,22]. Consequently, younger migrants not only experience greater integration, but also acquire the skills necessary for sustained success in the city, strengthening their intentions to settle permanently.

Our contribution to the literature is threefold. First, this paper enriches the existing body of research on migrants’ settlement intentions. While previous studies have extensively examined the factors influencing migrants’ settlement intention [5,6,10,23], they have largely overlooked the role of early urban experiences. We address this gap by focusing on how early urban experiences shape migrants’ settlement intentions, which contributes to the literature on studying migrants’ settlement intentions. Second, much of the literature in labor economics has emphasized the influence of early urban experiences on labor market outcomes for migrants. However, few studies have investigated how early urban experiences affect settlement intention [21,24]. Our study deepens the understanding of how early urban migration impacts migrants’ settlement intention, and contributes to this body of literature. Third, the role of language and culture in migrants’ urban life is an emerging area of interest within the academic community [25,26,27]. Our research demonstrates that migrants who migrate to urban areas at a younger age develop stronger dialect skills, which, in turn, facilitate their settlement intention in urban cities. This study contributes to this growing body of literature by underscoring the critical importance of language proficiency and cultural integration in fostering urbanization.

2. Theoretical Analysis and Research Hypothesis

Referencing existing theories and literature, we explore the mechanisms through which the age at first migration to urban areas impacts the settlement intention of migrants. The mechanisms mainly include three components: the social integration effect, the habit convergence effect, and the dialect ability effect (Figure 1). We conduct theoretical analysis from the above three aspects and propose research hypotheses.

Migrants who relocate to urban areas at a younger age tend to assimilate more easily into the cultural and social fabric of their new environment. Younger migrants are generally more adaptable to urban lifestyles, allowing them to understand local customs and build relationships with residents more effectively, which fosters smoother integration [28]. Theories of cognitive and social development suggest that younger individuals exhibit greater flexibility and openness to new experiences, making it easier for them to internalize social norms, behaviors, and cues [29]. This adaptability is reinforced by psychological and sociological research, which highlights the higher capacity for social learning and adjustment in younger people [30]. For example, studies indicate that younger migrants are more likely to actively engage in local cultural and social activities, promoting the formation of social networks and the adoption of urban cultural norms [31]. These early-established connections are crucial for the economic integration of migrants. Thus, those who migrate to cities at a younger age tend to integrate into urban life more rapidly.

Existing research consistently shows that social integration plays a pivotal role in shaping migrants’ settlement intentions, with numerous studies highlighting the multifaceted ways in which integration influences long-term residence decisions. Social integration, encompassing the development of social ties, participation in local activities, and the adoption of cultural norms, helps migrants foster a sense of belonging and attachment to their host cities. Niu and Zhao (2018) demonstrated that migrants who actively engage in local social networks are more likely to express a strong intention to settle, as these social connections provide emotional support, mitigate feelings of isolation, and strengthen their ties to the community [32]. In a similar vein, Zhang et al. (2023) emphasized that cultural adaptability, including language proficiency and participation in local customs, enhances a migrant’s sense of integration, which in turn significantly boosts their willingness to remain in urban areas long-term [11]. Chen (2023) further noted that social integration helps reduce the negative effects of discrimination and social exclusion, which are key deterrents to settling down in a new environment [10]. Wang and Shen (2022) confirmed that economic factors linked to social integration, such as improved labor market outcomes through social networks, also enhance the likelihood of long-term settlement [9]. Taken together, these studies show that social integration—by providing emotional, cultural, and economic support—acts as a critical determinant of migrants’ settlement intentions. However, while social integration is important for migrants, migrants may still face challenges in securing stable, high-income jobs. Low-wage employment or exploitation in the labor market is a risk, especially for undocumented or lower-skilled migrants. There is no doubt about the importance of social integration for migrants’ long-term residence intentions.

Based on the above analysis, we present Hypothesis 1, as follows.

Hypothesis 1.

Early migration to the city can promote migrants’ long-term residence intention by promoting social integration.

Early migration to cities is closely linked to the process of habit convergence, where migrants gradually adapt their behaviors and lifestyles to align with those of residents. Migrating at a younger age allows individuals to assimilate more quickly into the urban environment, as they are more flexible and open to adopting new habits. Studies have shown that younger migrants are more likely to adjust their daily routines, social behaviors, and consumption patterns to match those of the local population [33]. This adaptation process is crucial for their social and economic integration. For instance, younger migrants are often more adaptable in adopting urban hygiene practices, dietary preferences, and other social norms, which helps bridge the cultural gap between migrants and locals [34]. Additionally, habit convergence fosters a sense of belonging, as migrants who adopt local customs feel more integrated into the community [35]. Research by Zhu (2007) indicates that early exposure to urban life enhances migrants’ ability to develop habits that align with the urban lifestyle, such as efficient time management and professional work practices [14]. Moreover, migrants who adjust their habits to fit local standards are more likely to experience social acceptance, which can reduce the psychological distance between them and the native population [36].

Due to differences in daily living habits between migrants and residents, migrants often experience external pressures stemming from urban social health norms [37]. These differences in living habits can act as a significant marker distinguishing migrants from locals, and they may hinder migrants’ ability to integrate smoothly into local society. If migrants’ lifestyles diverge greatly from those of the local population, the psychological gap between the two groups can grow wider. As a result, migrants may perceive themselves as targets of discrimination or resistance from the local community [38]. Over time, this makes it increasingly difficult for migrants to fully integrate, potentially leading to a lack of belonging in the city [39]. Therefore, adjusting their hygiene habits to align more closely with local standards is key to narrowing the psychological distance. By adopting hygiene practices similar to those of residents, migrants are likely to strengthen their identification with the host city and be encouraged to engage more fully in its society.

Based on the above analysis, we present Hypothesis 2, as follows.

Hypothesis 2.

Early migration to cities can enhance migrants’ long-term residential intentions by facilitating habit adaptation and convergence.

The critical period hypothesis in linguistics suggests that younger individuals possess a greater capacity for language acquisition compared to adults, particularly during early childhood when the brain is especially receptive to learning languages [40,41]. During this phase, the brain’s heightened neural plasticity allows children and younger individuals to more easily absorb and retain linguistic structures. Cognitive development theories further reinforce this idea, indicating that younger individuals demonstrate increased cognitive flexibility and stronger learning abilities, which facilitate the process of acquiring new languages [42,43]. Consequently, migrants who move to urban areas at a younger age are more likely to be in this critical period, which enables them to learn the local dialect more efficiently. So, migrants who move to urban areas at a younger age are more likely to exhibit stronger dialect ability. In this study, dialect ability is primarily defined by the fluency with which migrants speak the local dialect of the city where they reside.

Proficiency in the local language allows migrants to communicate effectively with native residents, helping to build trust-based relationships. Additionally, dialect proficiency supports migrants in better integrating into the local community [44]. Research consistently demonstrates that proficiency in local dialects plays a significant role in influencing migrants’ settlement intentions. Migrants with higher levels of dialect proficiency tend to experience smoother social integration, as they can communicate more effectively with residents, thereby building stronger social networks. This communication not only fosters a sense of belonging, but also helps migrants navigate local institutions and access social services more easily, which enhances their willingness to settle long-term. According to Huang et al. (2023), dialect proficiency strengthens migrants’ social ties, which in turn reduces social isolation and increases their desire to remain in the city [45]. Xie et al. (2021) further noted that migrants who can speak the local dialect are perceived more positively by residents, facilitating better social acceptance and reducing barriers to integration [10]. The ability to communicate in the local dialect also enables migrants to participate more fully in community activities, which reinforces their emotional and cultural connection to the city. Wu and Logan (2016) emphasize that dialect proficiency is a key determinant of migrants’ ability to access employment opportunities and build economic stability, which is critical for settlement decisions [35]. Finally, Baker et al. (2021) suggest that dialect ability contributes to a migrant’s ability to integrate culturally, which, combined with economic and social factors, significantly increases the likelihood of their long-term settlement intentions [46].

Based on the above analysis, we present Hypothesis 3, as follows.

Hypothesis 3.

Early migration to the city can promote migrants’ long-term residence intention by increasing dialect ability.

3. Materials and Methods

3.1. Data Source

The data for this study are primarily drawn from the 2017 China Migrants Dynamic Survey (CMDS2017). This survey collects comprehensive information on migrants across all provinces in China, providing high credibility and strong national representation. CMDS2017 offers detailed data on migrants’ settlement intentions in cities, along with their age at first migration, making it particularly well-suited for analyzing the impact of age at first migration on settlement intentions. The study sample was selected based on the availability of relevant data, specifically regarding whether individuals have started their businesses and their age at first migration. Our analysis focuses exclusively on migrants who relocated from rural to urban areas, with the age at first migration restricted to those between 12 and 60 years old. After excluding cases with missing values, the final sample consists of 146,281 migrants.

3.2. Variables

3.2.1. Description of Variables

The dependent variable in this study is the long-term settlement intention of rural migrants [47,48]. It is constructed based on responses to the following question: “How long do you plan to stay if you intend to remain in this location?” The answer options are (1) 1–2 years, (2) 3–5 years, (3) 6–10 years, (4) more than 10 years, (5) permanent settlement, and (6) unsure. A dummy variable is created, equal to 1 if the migrant selects options (3), (4), or (5), indicating a longer-term settlement intention, and 0 otherwise.

3.2.2. Independent Variable

The independent variable in this study is the age at first migration. We obtain detailed data on this from the CMDS 2017, which provides specific information about the timing of migrants’ initial move to urban areas. Using this data, we can calculate the age at which each individual first migrated to an urban location.

3.2.3. Control Variables

Referring to existing studies (Huang and Chen, 2022; Liu et al., 2024), we incorporate the following control variables [7,14]: gender (male = 1, female = 0), age, ethnicity (Han = 1, others = 0), education level (no formal education = 1, primary school =6, middle school = 9, high school = 12, college = 14, undergraduate = 16, graduate = 1), marital status (married = 1, others = 0), health status (very poor = 4, poor = 3, good = 2, excellent = 1), entrepreneurship (self-employed or employer = 1), and migration frequency (the total number of cities the migrant has moved to). We also account for whether the migration is trans-province (yes = 1) or trans-city (yes = 1). Additionally, we include city fixed effects, such as destination-city fixed effects, origin-city fixed effects, and the combined fixed effects of both origin and destination cities.

These control variables were selected to account for key demographic, socioeconomic, and migration-specific factors that influence migrants’ settlement intentions, as highlighted in previous studies. For instance, gender, age, and marital status capture essential demographic influences, with previous research indicating that these factors often affect integration levels and settlement stability. Ethnicity is included to address potential cultural and social integration challenges faced by non-Han migrants, which may impact their willingness to settle long-term. Education level, entrepreneurship, and health status are socioeconomic indicators that reflect a migrant’s ability to adapt and thrive in an urban environment, affecting both economic security and social integration. Migration frequency, along with trans-province and trans-city migration status, provides insight into mobility patterns and commitment to a specific location, which are critical for assessing settlement intentions.

Table 1. Descriptive statistics.

Variables	Abbreviation	Mean	SD	Min	Max
Key variables
Long-term residence intention	Residence	0.415	0.493	0	1
Age at first migration	Age_mig	24.99	8.834	12	59.25
Age	Age	36.31	9.474	16	60
Gender	Gender	0.516	0.500	0	1
Ethnicity	Ethn	0.903	0.296	0	1
Marital status	Marriage	0.823	0.382	0	1
Education	Edu	9.987	3.189	0	19
Health	Health	1.188	0.445	1	4
Entrepreneurship	Entre	0.336	0.472	0	1
Migration frequency	Mig_num	1.997	1.925	1	92
Trans-province migration	Mig_prov	0.509	0.500	0	1
Trans-city migration	Mig_city	0.318	0.466	0	1

presents the descriptive statistics of the variables used in this study. As shown, only 41.5% of migrants express long-term residence intentions, indicating significant room for improvement in enhancing the long-term settlement intentions of the migrant population. The mean age at first migration is 24.99, suggesting that many Chinese migrants relocate to cities relatively late. The descriptive statistics also reveal key characteristics of Chinese migrants. First, the average age of migrants is 36.31, suggesting that the majority are relatively young. Second, the average number of migrations per person is approximately two, indicating that most migrants do not move frequently. This pattern indirectly suggests that Chinese migrants tend to settle in one city rather than migrate frequently. Lastly, over half of the migrants are engaged in inter-provincial migration.

In addition to the statistical analysis of the migrant population’s characteristics and their long-term residence intentions, we conducted a simple analysis of the relationship between long-term residence intention and age at first migration, using statistical graphs. The results are presented in Figure 2. The graph reveals a significant negative correlation between migrants’ long-term residence intention and their age at first migration. In other words, the older the individuals are when they first migrate to cities, the lower their willingness to settle long-term. However, it is important to note that this represents a simple correlation analysis, and a rigorous empirical study is required to establish a causal relationship between the two variables.

3.3. Econometric Models

We used OLS regression to estimate the impact of the age at first migration on the entrepreneurship activities of migrant workers. We consider the following regression model:

$$Residenc{e}_{ijk}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_{1}Age\_mi{g}_{ijk}+{\mathsf{\alpha }}_{2}Contro{l}_{ijk}+{\mathsf{\lambda }}_k+{\mathsf{\gamma }}_j+{\mathsf{\lambda }}_k\times {\mathsf{\gamma }}_j+{ϵ}_{ijk}$$

(1)

where $i$ represents rural-to-urban migrants, $j$ represents the cities of destination, and $k$ represents cities of origin. ${\mathsf{\lambda }}_k$ and ${\mathsf{\gamma }}_{\mathrm{j}}$ represent the cities-of-destination fixed effects and cities-of-origin fixed effects, respectively. ${\mathsf{\lambda }}_k\times {\mathsf{\gamma }}_j$ represents the combined fixed effects of both destination and origin cities. $Residence$ represents whether the migrant worker chooses to start a business or not. $Age\_mig$ represents the age at first migration to urban areas. $Control$ is a vector of control variables for the migrant. ${ϵ}_{ijk}$ is the random error term. The effect of the age at first migration on migrants’ entrepreneurship is captured by ${\alpha }_1$.

4. Results

4.1. Basic Results

Table 2. Basic regression results.

	(1)	(2)	(3)	(4)	(5)
Dependent Variable	Residence	Residence	Residence	Residence	Residence
	OLS	OLS	OLS	OLS	OLS
Age_mig	−0.0038 ***	−0.0123 ***	−0.0119 ***	−0.0122 ***	−0.0120 ***
	(0.0005)	(0.0006)	(0.0005)	(0.0005)	(0.0005)
Age		0.0119 ***	0.0110 ***	0.0112 ***	0.0111 ***
		(0.0008)	(0.0007)	(0.0007)	(0.0007)
Gender		−0.0364 ***	−0.0313 ***	−0.0322 ***	−0.0302 ***
		(0.0053)	(0.0047)	(0.0046)	(0.0046)
Ethn		−0.0554 ***	−0.0430 ***	−0.0142 **	−0.0107
		(0.0187)	(0.0128)	(0.0066)	(0.0069)
Marriage		0.1732 ***	0.1448 ***	0.1457 ***	0.1465 ***
		(0.0092)	(0.0072)	(0.0070)	(0.0073)
Edu		0.0306 ***	0.0269 ***	0.0263 ***	0.0261 ***
		(0.0026)	(0.0022)	(0.0018)	(0.0017)
Health		0.0002	−0.0089 **	−0.0100 ***	−0.0143 ***
		(0.0059)	(0.0039)	(0.0038)	(0.0040)
Entre		−0.0141 *	0.0051	0.0094 *	0.0113 **
		(0.0073)	(0.0059)	(0.0053)	(0.0054)
Mig_num		−0.0243 ***	−0.0183 ***	−0.0176 ***	−0.0163 ***
		(0.0020)	(0.0013)	(0.0013)	(0.0013)
Mig_prov		−0.1156 ***	−0.1373 ***	−0.1127 ***	0.0188
		(0.0185)	(0.0097)	(0.0087)	(0.1932)
Mig_city		−0.0153	−0.0346 ***	−0.0398 ***	−0.0546
		(0.0114)	(0.0087)	(0.0069)	(0.1033)
Cities-of-destination fixed effects	NO	NO	YES	YES	YES
Cities-of-origin fixed effects	NO	NO	NO	YES	YES
Combined fixed effects	NO	NO	NO	NO	YES
N	146,281	146,281	146,280	146,273	137,478
R2	0.0046	0.0941	0.1500	0.1612	0.2311

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1%, 5%, and 10% are denoted by ***, **, and *, respectively.

presents the key regression results on the effect of age at first migration on migrants’ settlement intentions. The estimation results from Column (1) through (5) show that age at first migration has a statistically significant negative impact on migrants’ settlement intentions, even after accounting for various control variables and including fixed effects. These findings suggest that migrating to cities at an earlier age significantly increases the likelihood of long-term settlement intentions. Additionally, the regression results indicate that migrants who relocate more frequently are less likely to settle long-term. Conversely, married migrants show a stronger intention to settle long-term, and those with higher education levels are more likely to express long-term settlement intentions.

4.2. From Correlation to Causality

The initial regression results show only a statistical association between the age of first migration and migrants’ long-term settlement intention, without providing evidence of a causal relationship. To establish causality, it is essential to address potential endogeneity issues. These concerns primarily arise from two sources: first, the omission of important control variables that may be correlated both with migrants’ age at first migration and their settlement intentions; and second, the possibility of reverse causality, where migrants who move to cities at a younger age might inherently have a stronger intention to settle long-term. As a result of these endogeneity concerns, the baseline estimates may lack precision.

To tackle potential endogeneity, we employ the instrumental variable (IV) approach. The instrument we use is the amount of spring rainfall in the migrants’ county of origin during the year they first migrated to urban areas. Based on fundamental agricultural principles, spring rainfall is critical for rural grain production. Both excessive rainfall, which can lead to flooding, and insufficient rainfall, which can cause drought, negatively affect grain yields in the region. When grain production and income decline, farmers are more likely to migrate to cities in search of alternative livelihoods. Conversely, in regions with normal rainfall, farmers are less inclined to migrate, resulting in later migration. We propose that spring rainfall in a migrant’s county of origin during the year of their first migration is correlated with their age at first migration. Moreover, as a natural climate variable, spring rainfall in their counties of origin before their first migration is largely exogenous to their current settlement intentions in cities. Therefore, we argue that the spring rainfall variable meets the requirements for a valid instrumental variable in our study.

To determine whether spring rainfall is normal, we calculate annual spring rainfall data for each county in China. The original data comes from the “China Ground Climate Data Daily Dataset”. Following [49], we calculated the “percentage precipitation anomaly” variable to classify the drought and flood levels in a region. We define “normal rainfall” as a percentage precipitation anomaly between −25% and 25%.

The results of the instrumental variable (IV) estimation are presented in columns (1) to (2) of

Table 3. Results of instrumental variable (IV) regression.

	(1)	(2)
Dependent Variable	Age_mig	Residence
	IVProbit	IVProbit
	First stage	Second stage
Age_mig		−0.0119 ***
		(0.0037)
Spring rainfall (normal)	0.8670 ***
	(0.1066)
Control variables	YES	YES
Cities-of-destination fixed effects	YES	YES
Wald test	814.644
p-value	0.000
N	106,085	106,085

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% are denoted by ***. Control variables include all variables listed in Table 2.

. Column (1) displays the first-stage results of the IV estimation, indicating that normal rainfall is not a weak instrument, as evidenced by its statistically significant coefficient. Column (2) reveals that age at first migration continues to have a statistically significant negative effect on the settlement intention of migrants under the IV estimation. Overall, the regression results using instrumental variables demonstrate that age at first migration retains a significant impact on migrants’ settlement intentions after addressing the endogeneity issues. It is worth noting that the coefficients from the instrumental variable regression results are similar to those from our earlier baseline regression. This similarity underscores the reliability of the empirical findings presented in our study.

4.3. Robustness Analysis

In the previous section, we used the instrumental variable method to identify the causal relationship between migrants’ settlement intentions and their age at first migration. Additionally, we conducted several robustness checks to ensure the reliability of our baseline results.

Table 4. Robustness analysis.

	(1)	(2)	(3)	(4)
Dependent Variable	Residence	Residence	Residence	Residence
	Delete “not sure” about settlement intention	First migration was for working purposes	Completed their education before first migration	Adding control variables at the family level
Age_mig	−0.0138 ***	−0.0119 ***	−0.0113 ***	−0.0114 ***
	(0.0006)	(0.0006)	(0.0005)	(0.0005)
Control variables	YES	YES	YES	YES
Cities-of-destination fixed effects	YES	YES	YES	YES
Cities-of-origin fixed effects	YES	YES	YES	YES
Combined fixed effects	YES	YES	YES	YES
Control variables at the family level				YES
N	107,767	111,790	111,166	137,478
R2	0.2813	0.2257	0.2222	0.2386

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2.

presents the findings from these tests.

First, regarding the question on settlement intention—specifically, how long rural migrants plan to stay—respondents were given the following answer options: (1) 1–2 years; (2) 3–5 years; (3) 6–10 years; (4) more than 10 years; (5) permanent settlement; and (6) not sure. We excluded migrants who selected option (6) “not sure,” as including these uncertain responses could distort the empirical results. Column (1) presents the results, showing that the coefficient remains significantly negative.

Second, recognizing that some migrants may have initially migrated for reasons other than work, such as visiting family or pursuing education, we restricted the sample to only those whose first migration was specifically work-related. The results in Column (2) show that the coefficient remains significantly negative, reducing concerns that the motivation for migration might affect the results.

Third, we further refined the sample to include only individuals who completed their education before their first migration to urban areas. The results in Column (3) again show that the coefficient remains significantly negative.

Last, based on the baseline regression, we further introduced control variables at the family level, primarily including the natural logarithm of family income and the number of family members living together in the city. The results in Column (4) show that the coefficient remains significantly negative after adding control variables at the family level. These findings indicate that our baseline regression results are robust across different specifications.

4.4. Mechanism Analysis for the Effect of Early Migration to Cities

In this section, we aim to explore the potential influence mechanisms of early migration to cities on migrants’ long-term settlement intention.

4.4.1. Social Integration Effect

According to the Theoretical Analysis section above, early migration to cities can promote migrants’ settlement intention by promoting social integration. We first look at the relationship between migrants’ age at first migration and social integration. The existing literature suggests that psychological integration can best reflect the social integration of migrants [50]. To measure the social integration of migrants, we posed the following question in CMDS2017: “Do you agree with the statement: I am very willing to integrate into the local community and become one of them?” Respondents were provided with the following answer options: (1) Strongly Disagree; (2) Disagree; (3) Agree; and (4) Strongly Agree. These response options are scored from 1 to 4, with higher scores indicating better social integration. Column (1) of

Table 5. Mediating role of social integration.

	(1)	(2)	(3)
Dependent Variable	Social Integration	Residence	Residence
	OLS	OLS	OLS
Age_mig	−0.0049 ***
	(0.0004)
Social integration		0.1542 ***	0.1464 ***
		(0.0057)	(0.0051)
Control variables	YES	NO	YES
Cities-of-destination fixed effects	YES	YES	YES
Cities-of-origin fixed effects	YES	YES	YES
Combined fixed effects	YES	YES	YES
N	137,478	137,478	137,478
R2	0.1501	0.2070	0.2367

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2.

shows the impact of age at first migration on migrants’ social integration. The coefficients of “Age_Mig” in Column (1) are negative and statistically significant, which indicates that early migration to cities can indeed help to promote migrants’ social integration.

We further test the impact of migrants’ social integration on their long-term settlement intention. The empirical results presented in Columns (2) and (3) demonstrate that the coefficients for social integration are significantly positive. These findings indicate that early migration to cities can increase long-term settlement intentions by promoting the social integration of migrants. The above results confirm Hypothesis 1.

4.4.2. Habit Convergence Effect

According to Hypothesis 2, early migration to cities can enhance the settlement intention of migrants by promoting the convergence of their habits with those of residents. Due to data limitations, we use hygiene habits here as a proxy for habits. We constructed the habit convergence variable based on the question “Do you agree that my hygienic habits are quite different from local citizens?”. Responses such as “completely agree”, “basically agree”, “disagree” and “completely disagree” are coded as 4, 3, 2, and 1, respectively. Results are shown in

Table 6. The mediating role of habit convergence.

	(1)	(2)	(3)
Dependent Variable	Habit Convergence	Residence	Residence
	OLS	OLS	OLS
Age_mig	−0.0061 ***
	(0.0004)
Habit convergence		0.1450 ***	0.1315 ***
		(0.0064)	(0.0053)
Control variables	YES	NO	YES
Cities-of-destination fixed effects	YES	YES	YES
Cities-of-origin fixed effects	YES	YES	YES
Combined fixed effects	YES	YES	YES
N	137,478	137,478	137,478
R2	0.1546	0.2064	0.2339

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2.

. In Column (1), the coefficient of public health services is significantly positive at the level of 1%. It implies that public health services play a crucial role in fostering the convergence of hygienic habits between migrants and residents. In Column (1), the coefficient of “Age_mig” is significantly negative at the level of 1%, suggesting that habit convergence can enhance migrants’ settlement intention.

We further test the impact of habit convergence on their long-term settlement intention. The empirical results presented in Columns (2) and (3) demonstrate that the coefficients for habit convergence are significantly positive. These findings indicate that early migration to cities can increase long-term settlement intentions by facilitating habit adaptation and convergence of migrants. The above results confirm Hypothesis 2.

4.4.3. Dialect Ability Effect

According to Hypothesis 3, early migration to cities can promote the settlement intention of migrants by enhancing migrants’ dialect ability. Regarding dialect ability, the question in the CMDS2017 dataset is, “How well do you master the local language in the city?” If the local language refers to Mandarin, it indicates proficiency in Mandarin. If the local language refers to a specific regional dialect, it indicates proficiency in that particular dialect.

Table 7. Mediating role of dialect ability.

	(1)	(2)	(3)	(4)
Dependent Variable	Residence	Dialect Ability	Dialect Ability	Residence
	OLS	OLS	OLS	OLS
Age_mig	−0.0127 ***	−0.0134 ***
	(0.0017)	(0.0030)
Dialect ability			0.0702 ***	0.0559 ***
			(0.0058)	(0.0054)
Control variables	YES	YES	NO	YES
Cities-of-destination fixed effects	YES	YES	YES	YES
Cities-of-origin fixed effects	YES	YES	YES	YES
Combined fixed effects	YES	YES	YES	YES
N	13,535	13,535	13,535	13,535
R2	0.1547	0.5304	0.0699	0.1443

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2. The data on dialect ability come from CMDS 2014.

presents the mediating role of dialect ability in the relationship between age at first migration and settlement intention. Column (1) examines the impact of age at first migration on settlement intention by using 2014 CMDS data. The negative coefficient indicates that the older the migrants are when they first migrate to the city, the less likely they are to engage in entrepreneurial activities. Column (2) investigates the relationship between age at first migration and dialect ability by using 2014 CMDS data (Since the CMDS 2017 data do not include information on migrants’ dialects, we used the CMDS 2014 data for analysis.). The negative coefficient suggests that the younger the migrants are when they first migrate to the city, the more fluent their dialect ability.

We further test the impact of migrants’ dialect ability on their long-term settlement intention. The empirical results presented in Columns (3) and (4) demonstrate that the coefficients for social integration are significantly positive. These results indicate that early migration to cities can promote the settlement intention of migrants by improving their dialect ability. The above results confirm Hypothesis 3.

4.5. Heterogeneous Effect

This section examines the heterogeneous impact of early migration to cities on migrants’ entrepreneurship. The analysis focuses on three main aspects: migration frequency, education level, and city type.

Table 8. Heterogeneous effects.

	(1)	(2)	(3)	(4)	(5)	(6)
Dependent Variable	Residence	Residence	Residence	Residence	Residence	Residence
	Mig_num		Marriage		Lifetime
	Mig_num = 1	Mig_num > 1	Marriage = 1	Marriage = 0	Lifetime ≥ 5	Lifetime < 5
Age_mig	−0.0158 ***	−0.0105 ***	−0.0117 ***	−0.0148 ***	−0.0115 ***	−0.0080 ***
	(0.0006)	(0.0005)	(0.0005)	(0.0011)	(0.0006)	(0.0005)
Control variables	YES	YES	YES	YES	YES	YES
Cities-of-destination fixed effects	YES	YES	YES	YES	YES	YES
Cities-of-origin fixed effects	YES	YES	YES	YES	YES	YES
Combined fixed effects	YES	YES	YES	YES	YES	YES
N	69,337	63,439	112,155	21,803	60,951	71,833
R2	0.2537	0.2573	0.2267	0.2702	0.2339	0.2409

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2.

presents the empirical results. Column (1) analyzes migrants who have migrated only once, while Column (2) examines those who have migrated multiple times. The larger negative coefficient in Column (1) compared to Column (2) suggests that for migrants with lower migration frequency, the positive impact of early migration to cities on settlement intention is more pronounced.

Column (3) analyzes migrants who are married, while Column (4) examines those who are not married. The larger negative coefficient in Column (4) compared to Column (3) suggests that for migrants who are not married, the positive impact of early migration to cities on settlement intention is more pronounced.

Column (5) analyzes migrants who have resided in the local area for more than five years, while Column (6) examines those who have resided in the local area for less than five years. The larger negative coefficient in Column (5) compared to Column (6) suggests that for migrants who have resided in the local area longer, the positive impact of early migration to cities on settlement intention is more pronounced.

4.6. Further Analysis: The Impact of Early Migration to Cities on Settlement Intention

To gain deeper insights into how the age at first migration influences settlement intention, in this part, we focus on the settlement intention involving the transfer of one’s household registration (hukou) status. The settlement intention involving the transfer of one’s household registration (hukou) status is constructed based on the answers to the following question. The question is “If you meet the local settlement requirements, would you be willing to transfer your household registration (hukou) to this location?”, with answer options (1) willing; (2) unwilling; and (3) undecided. The dummy variable is equal to 1 if rural migrants answered the question with options (1), and 0 otherwise.

Table 9. The impact of early city migration on settlement intention involving hukou transfer.

	(1)	(2)
Dependent Variable	Residence	Residence
	OLS	OLS
Age_mig	−0.0014 ***	−0.0028 ***
	(0.0003)	(0.0004)
Control variables	NO	YES
Cities-of-destination fixed effects	YES	YES
Cities-of-origin fixed effects	YES	YES
Combined fixed effects	YES	YES
R2	137,478	137,478
N	0.2112	0.2207

Note: Standard errors in parentheses are clustered at the city level. The significance levels of 1% by ***. Control variables include all variables listed in Table 2, but for mig_num.

presents the findings of this analysis. Column (1) and Column (2) all show a negative coefficient, suggesting that early migration to cities, similarly, can also promote settlement intention involving the transfer of one’s household registration (hukou) status.

5. Conclusions and Discussions

As urbanization advances toward high-quality development, the social integration of migrants and their long-term intention to settle in cities has gained importance. Migrants who relocate to cities at a younger age are generally more inclined to remain in urban areas. Using data from the Chinese Migrant Dynamics Monitoring Survey (CMDS), this study examines how the age at first migration affects migrants’ long-term settlement intentions and explores the underlying mechanisms. Our findings indicate that migrants who move to urban areas at an earlier age are more likely to have strong intentions for long-term settlement. Mechanism analysis reveals that early migration fosters social integration, supports habit convergence, and improves dialect proficiency. Additionally, heterogeneity analysis shows that the positive influence of early migration is especially significant among migrants who have moved only once, those who are unmarried, and those who have resided locally for an extended duration. The analysis also suggests that early migration positively affects migrants’ willingness to transfer their household registration (hukou) status.

This study contributes to the literature by highlighting the often-overlooked role of early life experiences in shaping long-term settlement decisions. By bridging the gap between migration studies and urban integration, this analysis provides new insights into how early urban experiences influence social capital formation and settlement decisions. While much of the existing research has focused on early experiences and labor market outcomes, our study extends these findings to settlement intentions, thus broadening the understanding of early experiences in migration dynamics. Our results also underscore the role of language and culture, showing that migrants who move to cities earlier in life develop stronger language skills, significantly enhancing social integration and the likelihood of permanent settlement.

This study has certain limitations. The CMDS dataset, while comprehensive, primarily captures urban migration patterns up to 2017, which may not fully reflect recent trends influenced by evolving migration policies. Additionally, our study focuses on specific aspects of social integration, potentially overlooking other cultural or economic factors that contribute to settlement intentions. Future research could provide a more thorough understanding by incorporating more recent data and broader measures of integration. Additionally, due to data limitations, this study relies solely on cross-sectional data for empirical analysis. We acknowledge that establishing causal relationships is challenging when using cross-sectional data, as they impact the strength of any conclusions drawn. To address potential issues in the empirical analysis, we applied the instrumental variable method.

Our findings are consistent with recent domestic studies that highlight the importance of social and linguistic integration in shaping settlement intentions among China’s migrant population. International studies similarly underscore the role of early migration and integration experiences in fostering long-term settlement across various urban contexts. Comparative research on the impact of early migration in different national settings would further clarify the universality and contextual limitations of our findings.

The findings hold significant implications beyond China, especially as migrant integration and settlement have become global challenges. Encouraging early migration is essential for fostering long-term settlement intentions, a critical concern for policymakers both nationally and internationally. Migrants often face more obstacles than residents in securing resources such as housing, healthcare, and social services, which can hinder their integration and discourage permanent settlement. Policies that support younger migrants’ transitions into urban areas by creating environments conducive to social and cultural integration are thus crucial. A key recommendation is to design targeted programs that improve access to affordable housing, quality healthcare, and educational opportunities. Such programs not only address immediate needs, but also strengthen migrants’ connections to the local community, increasing the likelihood of long-term settlement. Expanding language training programs, particularly those focusing on local dialect proficiency, can further accelerate social integration and bridge cultural gaps between migrants and residents, especially in regions where linguistic differences may create social distance.