Land Attachment, Intergenerational Differences and Land Transfer: Evidence from Sichuan Province, China

: It is of great signiﬁcance to explore the inﬂuencing factors of land ﬂow to promote moderate-scale agricultural operation. However, few studies have explored the quantitative inﬂuences of land attachment and intergenerational difference on land transfer. Based on the survey data of 540 rural households in Sichuan Province, this study uses factor analysis method to divide land attachment into land satisfaction, land rootedness, and land dependence, and further empirically tests the impact mechanism of land attachment and intergenerational difference on land ﬂow by using Probit model and Tobit model. The results are as follow: (1) land attachment is signiﬁcantly correlated with land ﬂow-out, but not with land ﬂow-in. (2) Different dimensions of land attachment have different impacts on land ﬂow-out. Among them, land rootedness and land dependence have signiﬁcant negative impacts on farmers’ land ﬂow-out behavior and land ﬂow-out area, while land satisfaction has a signiﬁcant positive impact on farmers’ land ﬂow-out behavior and has no signiﬁcant impact on the land ﬂow-out area. (3) Different generations of land attachment have different impacts on land ﬂow-out. Among them, the land attachment of the new-generation farmers has no signiﬁcant impact on land ﬂow-out. Among middle-aged farmers, land dependence had a signiﬁcant negative impact on land ﬂow-out behavior and area, and land rootedness had a signiﬁcant negative impact on land ﬂow-out behavior; however, land satisfaction had a signiﬁcant positive impact on land ﬂow-out behavior and area. Among the older generation of farmers, land dependence has a signiﬁcant negative impact on land ﬂow-out behavior and area, while land satisfaction and land rootedness have no signiﬁcant impact on land ﬂow-out behavior and area. Therefore, in promoting the practice of land ﬂow, we should pay attention to the differences of farmers’ emotional demands, improve the supporting policies of land ﬂow by classiﬁcation, reduce farmers’ dependence on “land security”, solve farmers’ concerns on land ﬂow, and promote the rational ﬂow of land factors.


Introduction
The development of modern agriculture has attracted much attention.Activating land management rights and promoting land circulation are important means to achieve moderate-scale agricultural operation and are an inevitable trend of China's agricultural modernization [1][2][3].In recent years, several Central Documents NO.1 (as the first policy statement released by Chinese central authorities each year, the document is seen as an indicator of policy priorities.)have placed a special emphasis on promoting land flow and supporting moderate-scale operation [4].At the same time, China has successively issued documents such as the Opinions on Improving the Measures for The Separation of Contracted Management Rights of Rural Land (an opinion document issued by the general office of the CPC Central Committee and the general office of the State Council) and The Operating Standards for the Flow and Exchange Market of Rural Land Management Rights (Trial) (a management standard formulated in accordance with relevant laws, regulations, and policies) to promote the orderly flow of land-management rights and realize large-scale agricultural operations [5][6][7].Despite the continuous attention and active promotion of such policies by the government, the development of large-scale agricultural operation based on land transfer has not been as smooth as expected [8].According to the statistics of the Ministry of Agriculture and Rural Affairs, the land circulation area in China reached 555 million mu (1 mu ≈ 0.067 ha) in 2020, accounting for about 40.1% of the contracted farmland area in China [9].However, a large amount of cultivated land is still managed by small farmers in a decentralized manner, and the land transfer has not completely reversed the agricultural economic pattern dominated by small-scale farmers in China [10].For a long time, small farmers, as the main body of agricultural production and management in China, and small-scale production of farmers have produced a series of problems, such as land fragmentation and decentralization [11], low agricultural production efficiency [12], and inability to use economies of scale [13], which may pose challenges to economic and environmental sustainability and restrict the further development of modern agriculture in China.Researches show that land transfer can promote the improvement of economies of scale and efficiency [14] and improve agricultural benefits and farmers' income through the improvement of economies of scale [15].Therefore, it is of great practical significance to explore how to promote farmers' land circulation to realize large-scale management.
In the studies of land transfer factors, existing literature mainly focuses on individual characteristics [16], family characteristics [17], and policy characteristics [18,19], and provides guidance and reference for this paper.However, most of these studies are based on the hypothesis of "rational man" and pay more attention to farmers' interests and land policies, ignoring the attachment of "man" to land in the man-land relationship.Almost every aspect of human activities is driven by emotion [20], and farmers' land transfer behaviors are no exception.Land attachment is the product of farmers' psychological response to land [21,22], which plays a guiding and regulating role in farmers' land transfer behaviors.In addition, with the change of times and economic development, the trend of intergenerational differentiation of farmers in China has become increasingly obvious [23].Intergenerational differences will lead to gradually differentiated cognition, emotion, and behaviors of farmers of different generations [24], which will inevitably affect their decisionmaking regarding land transfer behaviors.Thus, it is necessary to analyze the influence mechanism of land attachment and intergenerational difference on land transfer.Currently, few studies [21,25] focus on the impact of land attachment and intergenerational differences on land transfer.Although they provide some ideas and basis for this paper, they still have some problems.First, they lack a systematic analysis combining qualitative and quantitative analysis.Second, the attachment to land is complex and diversified, but they are limited to a certain emotion, which may not fully reflect the impact of land attachment on land transfer decision-making.
Therefore, the study uses the factor analysis method to divide land attachment into land satisfaction, land rootedness, and land dependence; analyzes the influencing mechanism of farmers' land attachment on land transfer; and further explores intergenerational differences from the perspective of land attached to the impact on the land circulation differences, utilizing the survey data of 540 rural households in Sichuan Province.

Overview of the Study Area
Sichuan Province is located in Southwest China.The terrain is high in the West and low in the East, with rich and diverse geomorphic types, mainly including plains, hills, and mountains.With a total area of 486,000 km 2 , the province governs 21 prefecture level administrative regions and 183 county-level divisions.By the end of 2020, registered residence had a registered population of 90.816 million in Sichuan Province, including 56.061 million rural residents.The regional GDP reached 4859.876 billion yuan (1 yuan ≈ 0.14 euro or 0.15 dollar) and the agricultural output value was 470.188 billion yuan [26].Sichuan had 100.842 million mu of cultivated land, and the per-capita cultivated land in the province was about 1.11 mu [27].Among the sample counties, Yuechi County had the largest permanent population of more than 700,000, while Gaoxian county and Jiajiang County only had a permanent population of more than 300,000.In terms of the total regional output value, Jiajiang County and Yuechi County were more than 20 billion, but Gaoxian county was less than 17 billion.
Jiajiang County, Yuechi County, and Gao County are representative counties in plain, hilly, and mountainous areas of Sichuan Province, respectively.Their common characteristics are the small area of cultivated land per capita, the serious aging of agricultural labor force, and the widespread phenomenon of land transfer.Based on these common characteristics, the above counties were selected as the location of this study.

Data Source
The data used in this study mainly come from the questionnaire survey conducted by the research group in Sichuan Province in July 2021.The survey methods are as follows [28]: Firstly, in order to compare the impact of land attachment on land transfer under different agricultural production conditions.Jiajiang County (plain area), Yuechi County (hilly area), and Gao County (mountainous area) were selected as the research sites according to the level of economic development and landform.Secondly, according to the high, medium and low levels of township economic development, three sample townships were selected from each sample county.Thirdly, three sample villages were selected from each sample township in the same way.Finally, we randomly selected 20 farmers from each sample village and interviewed them or their families.The content of the questionnaire mainly includes the basic information of farmers, capital and land use, and so on.The survey involved 3 counties, 9 townships, and 27 villages.The response rate of farmers was 100%, and 540 valid questionnaires were obtained.Among the 540 respondents, 310 farmers transferred land, accounting for 57.4%.There were 230 farmers who had not transferred land, accounting for 42.6% of the total.The average contracted cultivated land scale of 540 respondents is 3.67 mu.Specially, there are 17 farmers under 1 mu, 51 farmers under 1-2 mu, 121 farmers under 2-3 mu, 132 farmers under 3-4 mu, 86 farmers under 4-5 mu, 93 farmers under 5-6 mu, and 68 farmers above 6 mu.The distribution map of the sample area is shown in Figure 1.Land attachment originates from the theory of place attachment.Place attachment refers to the positive emotional connection or relationship between people and their living environment [29].Researchers have deeply explored place attachment from the perspectives of anthropology, geography, sociology, and environmental psychology, and emphasized the connection between material and culture in terms of identity [30,31]; place; community and social linkages [32]; the symbolic meaning of place attachment [33]; and the cognitive, emotional, and behavioral components of place attachment [34].Place was defined as a unique geographical space, including the biophysical properties of farm property and surrounding landscape, as well as the social connections and relationships cultivated in this space [35].Farmers can form a separate attachment to each of these elements (farms) in a special way.At the same time, some people believed that for farmers, land is not only a space to provide them with life, work, entertainment, and social communication, but also a place with symbolic significance and repositories of emotion [22,36,37].Therefore, land attachment can be said to be place attachment in a narrow sense, which refers to the positive emotional relationship between people and land [38].Researchers held that farmers are rooted in the land and have a deep attachment to the land.For instance, Gray [39] proposed that farmers' attachment to land can be attributed to the innate "genetic metaphor".Carr [40] found that Chaga women on Mount Kilimanjaro in Tanzania have a deep attachment to their land and home.Quinn and Halfacre [22] also found that both farmers living on their own land and farmers renting land have a strong attachment to land.
Although researchers have conducted rich discussions on the phenomenon of land attachment and deeply conveyed land attachment in the study of the emotional relationship between farmers and land, the academic definition of land attachment is still vague [41].This paper studies the attachment of Chinese farmers to cultivated land.Therefore, in the definition of land attachment, the article not only refers to the definition of local attachment, but also draws lessons from the views of Chinese scholars.Chinese academic circles often express farmers' emotional attachment to land as land complex or land dependence.For example, Chen [42] defined the land complex as a deep mysterious emotion of possession, love, and attachment to the land deeply hidden in the hearts of farmers.Zhang et al. [43] believed that land complex refers to the unique attitude of farmers to give land an emotional and mysterious value.Combined with the reality and the author's perception, this paper considers that land attachment is a kind of affection, possession (or nostalgia), and dependence of farmers on the land they own (or used to own).In terms of the dimension division of land attachment, the study also draws lessons from the dimensions of place attachment: sense of place, place identity, place dependence, rootedness, and place satisfaction [44,45].Then the paper also refers to the division of land complex or land dependence in Chinese Studies: traditional beliefs, land dependence, and off-farmland business [43]; farmers' economic and emotional dependence on land [43]; and satisfaction, rootedness, and land dependence [46].Therefore, under comprehensive consideration, this study divides land attachment into three dimensions: land satisfaction, land rootedness, and land dependence.Among them, land satisfaction refers to farmers' satisfaction with land production and utilization; land rootedness refers to farmers' sense of identity, attention, and sense of return to land; and land dependence refers to the economic and emotional dependence of farmers on land.

The Impact of Land Attachment on Land Transfer
Why do farmers get attached to their land?Because land brings not only a living environment and material output [22,40], but also social relations and social identity [47].What impact does land attachment have on farmers' behavior?Ingram and Kirwan [48] mentioned in their research that due to farmers' emotional attachment to the land, elderly farmers are still unwilling to give up their shares in the farm even if they intend to leave the farm, and they also worry that the people who take over their farm would not seriously manage it.Ma [49] investigated the depth of farmers' attachment to land and found that it affected the transformation of farmers' land use: from multiple crop planting to single crop planting, cash crop planting, or abandonment.
The influence of farmers' cultivated land attachment on cultivated land transfer behavior is the focus of this section.We can explain the influence mechanism between land attachment and land transfer from the perspective of emotional sociology.Most emotional sociologists believed that emotion is the key mechanism to guide and stimulate micro actions [50].Emotion is the power of action, affecting people's attitude and cognition, and guiding people's decisions and behaviors.Studies have shown that if a person has a strong emotional attachment to an item, they are more likely to keep the item instead of selling it [51], and the emotional attachment to the item will affect the individual's mentality of "all his geese are swans", which in turn can affect the way people deal with things [52].In the context of rural China, farmers develop emotional attachment to the land in the long-term cultivation process, and the intensity of this attachment depends on individual needs.The individual needs of farmers are complex, and the influence of land attachment on land circulation is also complicated.In general, the stronger farmers' land attachment is, the more inclined they are to grow their own land or flow into land rather than flow out of land.Specifically, the higher the satisfaction of farmers on land use, the stronger the rootedness, the heavier the economic dependence and emotional dependence, and the more inclined they are to grow the land themselves or flow into land rather than flow out of land.However, at present, the main force of rural farming in China is middle-aged farmers, who are the sandwich generation that have to look after their elderly and their little children [53,54].Their economic rationality is stronger than their survival rationality, and they are more willing to adopt a part-time livelihood strategy, so they are more cautious in the aspect of land flow-in.Thus, the following hypotheses are proposed: H1a: Land attachment has a significant negative impact on land flow-out, but no significant impact on land flow-in.
Under the interaction of the urbanization process and land transfer policy, the relationship between man and land is constantly changing and gradually becoming more unstable.Therefore, the dimension of land attachment is unstable, and different dimensions of land attachment may have different impacts on land transfer.Thus, the following hypotheses are proposed: H1b: There may be differences in the effects of land satisfaction, land rootedness, and land dependence on land transfer-out.

The Impact of Land Attachment on Land Transfer from the Perspective of Intergenerational Differences
Intergenerational difference was proposed by Karl Mannheim, which represents the group characteristics with differences in values, preferences, attitudes, and behaviors among generations due to different birth years and growth backgrounds [55].Subsequently, Lyons and Kuron [56] defined "generation" as a group with the same birth age and experiencing the same major social events in the key growth stage.There are differences in values and preferences among individuals of different generations [57], which can lead to different value judgments and behavior patterns [58].Some scholars have concluded that the essence of intergenerational differences is the result of the combined effect of time effect, generation effect, and age effect [59], while the sample farmers in this study are less affected by time effect.Therefore, the intergenerational differences in this paper are mainly reflected in the generation effect and age effect.Generation effect refers to the differentiation of behavior and cognition between generations [58,60].For example, the older generation of farmers in the sample experienced the establishment of the household contract responsibility system in their middle-age and youth, and regaining the land made them have a deeper emotion for the land [21].While the new generation of farmers experienced the wave of market economic development and the continuous improvement of market consciousness, they formed a behavior attitude completely different from that of the middle-age and older generations of farmers.The age effect refers to the difference in psychological characteristics caused by age differences [59].For instance, the older generations of farmers are relatively older, have a long time to deal with land, and their sense of land "real right" and land attachment are stronger than those of middle-age and new-generation farmers.Therefore, with the changes of rural society, the depth of land attachment of several generations has produced intergenerational differences [21,25].This difference is largely affected by multiple factors, including gender, age, educational level, income, ideology, and social economy.
Specifically, the older generations of farmers have a very deep attachment to land [61].Many older generation farmers rely on land for their livelihood all their life [62,63].Land is the foundation of their survival, which gives them a sense of security [22], joy, and belonging, and forms strong land attachment.Under the dual effects of land attachment and distrust of the land-transfer market, the older generation of farmers are unwilling to transfer out of the land.Meanwhile, the restrictive effect of aging on labor capacity is stronger than the promoting influence of aging on land attachment [43], which can lead to the inability of elderly farmers to transfer in land.For the middle-aged generation of farmers, the intensity of their land attachment is between the new generation and the old generation.Some middle-aged generation farmers are in the prime of their life and are vigorous, and the deep emotion accumulated from long-term contact with the land made them more enthusiastic about agricultural production, so they were not willing to flow out of the land.Meanwhile, other middle-aged farmers consider life more rationally.No matter how much land they cultivate, they pay more attention to the economic value and living security brought by land [21].Such middle-aged farmers may flow out the land due to the low comparative benefits of agriculture.However, like the old generation of farmers, their attachment to the land makes it difficult to give up the land, and they tend to transfer out of the land for a short time or often pay attention to the utilization of the transfer-out land.The younger generation of farmers gradually abandoned their land attachment with the development of their rational and economic awareness [61].So, compared with the middleaged and the old-generation farmers, the land attachment of the new generation of farmers is obviously lighter.Johnsen [64] showed that when household income depends more on non-agricultural income, participation in agricultural activities and land income decreases, resulting in the separation of people and land, and reducing individual identification with the land.The new generation of farmers have higher education and employment ability [65], and the non-agricultural income is higher than the agricultural income.They are not only no longer limited to land production but also have gotten rid of the shackles of "land is the lifeblood".Besides, the new generation of farmers have relatively short contact with the land and do not pay attention to land production.In addition, under the influence of the thought of "despising peasant" [41], they have a fading land attachment and are more willing to transfer out of the land than in.In a word, farmers' land attachments are various among the different generations, which leads to different land transfer decisions.Thus, the following hypothesis is proposed: H2a: The land attachment of the new generation of farmers has a significant positive effect on land flow-out, but no significant impact on land flow-in.
The land attachment of middle-aged and older generation farmers has a significant inhibitory effect on land flow-out, but has no significant influence on land flow-in.
H2b: There may be intergenerational differences in the impact of land satisfaction, land rootedness, and land dependence on land transfer-out.
Based on the above analysis, this study brings land attachment, intergenerational differences, and land transfer decisions into the same analysis framework (Figure 2) and discusses the impact of land attachment and intergenerational differences on land transfer, in order to provide a new perspective for the study of land transfer.
H2b: There may be intergenerational differences in the impact of land satisfa land rootedness, and land dependence on land transfer-out.
Based on the above analysis, this study brings land attachment, intergeneration ferences, and land transfer decisions into the same analysis framework (Figure 2) an cusses the impact of land attachment and intergenerational differences on land tra in order to provide a new perspective for the study of land transfer.

Dependent Variables
The two dependent variables of this study are land transfer behaviors and area transfer behaviors specifically include "transfer-out behavior" and "transfer-in beha which mainly take "does your family transfer out land?" and "does your family tra in land?" as the basis of the farmers' transfer behaviors.If the answer is yes, we rec as 1, otherwise as 0. The transfer area divided "transfer-out area" and "transfer-in a which was measured by "how many mu of land does your family transfer out?" and many mu land does your family transfer in?".Then we wrote down the farmers' an directly.

Core Explanatory Variables
The one core independent variable of this study is land attachment.This stud vides the measurement items of land attachment into three dimensions: land satisfa land rootedness, and land dependence.The first dimension is land satisfaction, w mainly measures whether farmers are satisfied with the status of land use by a whether they agree with the statements of "I feel very satisfied with the current lan purpose" and "I feel very satisfied with the current land use condition".The second land rootedness dimension, corresponding to the questionnaire questions "I alway attention to my own land" and "I hope my future generations can understand the

Variable Definitions 2.3.1. Dependent Variables
The two dependent variables of this study are land transfer behaviors and area.The transfer behaviors specifically include "transfer-out behavior" and "transfer-in behavior", which mainly take "does your family transfer out land?" and "does your family transfer in land?" as the basis of the farmers' transfer behaviors.If the answer is yes, we record it as 1, otherwise as 0. The transfer area divided "transfer-out area" and "transfer-in area", which was measured by "how many mu of land does your family transfer out?" and "how many mu land does your family transfer in?".Then we wrote down the farmers' answers directly.

Core Explanatory Variables
The one core independent variable of this study is land attachment.This study divides the measurement items of land attachment into three dimensions: land satisfaction, land rootedness, and land dependence.The first dimension is land satisfaction, which mainly measures whether farmers are satisfied with the status of land use by asking whether they agree with the statements of "I feel very satisfied with the current land use purpose" and "I feel very satisfied with the current land use condition".The second is the land rootedness dimension, corresponding to the questionnaire questions "I always pay attention to my own land" and "I hope my future generations can understand the rural land", which mainly measures whether farmers still have deep embeddedness to the land.The third is the dimension of land dependence.Corresponding to the questionnaire questions "land is not only the source of living of a family, but also the spiritual pillar of an individual" and "land is the basic living security even when I work in cities or cannot engage in agricultural production", it mainly measures whether farmers still have economic and emotional dependence on land.There are 6 questions in total in the three dimensions, and the variables are assigned 1-5 according to the answers of farmers.In the reliability test, Cronbach 'α coefficient of the 6 questions is 0.71, greater than 0.7, indicating good reliability.In the validity test, the calculated KMO value is 0.68; manifesting the 6 questions is suitable for factor analysis on account of their KMO value, which is greater than 0.6 and the significance level is high.In principal component analysis, three common factors are extracted according to the preset and rotated by the maximum variance method.The cumulative variance contribution rate reaches 76.07%, which can effectively explain the problem information.The results of factor analysis are shown in Table 1.Another core independent variable of the study is intergenerational differences.In the measurement of intergenerational differences, considering the serious aging of agricultural labor force in the study area, and drawing on the research of Xie and Huang [65], this research divides generations by 1955 and 1970.Then the study generates dummy variables to signify intergenerational differences.If the head of a household was born after 1970, that person is named "new generation of farmer", and the value is 1.If he or she was born between 1955 and 1970, he or she is called "middle-aged generation of farmer", with a value of 2. If born before 1955, he or she is recorded "older generation of farmer", and we assign a value of 3.

Control Variables
On the basis of the relevant studies [3,[65][66][67][68], the study summarizes the control variables into four categories: the first is the characteristics of the head of household (gender and years of education), the second is the characteristics of the family (the proportion of non-agricultural income, the value of agricultural fixed assets, etc.), the third is the characteristics of social security, and the fourth is the characteristics of the village.Variable definitions and descriptive statistical analysis are shown in Table 2.

Research Models 2.4.1. Probit Model
When the explained variables are "transfer-out behavior" and "transfer-in behavior", they belong to two categories of variables [8,9], so the Probit model is constructed for parameter estimation.The basic regression formula is set as follows: In Formula (1), Y 1i represents whether farmer i transfers out or in land; sat i represents the i's satisfaction with land use; roo i represents the land rootedness of farmer i; dep i represents the land dependence of farmer i; Σ con i is a set of control variables; β 0 is a constant term; β 1 , β 2 , β 3, and β 4 respectively represent the corresponding regression coefficients, and ε i is the random error term.

Tobit Model
When the explained variables are "transfer-out area" and "transfer-in area", their values are similar to continuous variables and have more zero values [65,69].Therefore, the Tobit model is used to test the impact of land attachment on land transfer scale.The basic regression equation is set as follows: In Formula (2), Y 2i represents the land transfer-out or transfer-in area of farmer i, and the setting of other variables is consistent with Formula (1).

Basic Regression Results of the Impact of Land Attachment on Land Transfer
The basic regression estimation results are shown in Table 3. Table 3 shows that although the impact coefficient of land attachment on land transfer-in is mostly positive, it is not significant, because land transfer-in is more affected by agricultural production conditions, non-agricultural income, and other factors.From the perspective of land transfer out, land rootedness and land dependence in land attachment have a negative impact on farmers' land transfer-out behavior and area at a significant level of 5%, indicating that land rootedness and land dependence can inhibit land transfer-out.With other conditions unchanged, if the land rootedness is improved by one unit, the land transfer out probability will be reduced by 3.2% and the land transfer-out area will be reduced by 39.9%.Similarly, if the land dependence increases by one level, the land transfer-out probability will be reduced by 3.8% and the land transfer-out area will be reduced by 45.2%.Different from land rootedness and land dependence, land satisfaction significantly and positively promoted farmers' land transfer-out behavior at the level of 10%, but had no significant impact on the area of land transfer out.From the above results, we can see that the impact of land attachment on land transfer out is complex.Although the overall effect is inhibition, it has both inhibition and promotion in detail, which verifies H1a and H1b.

Model Robustness Test of the Impact of Land Attachment on Land Transfer
Because the explained variables "transfer-out behavior" and "transfer-in behavior" are assigned with "0 or 1", they belong to binary variables.The Probit model is used for estimation in basic regression.In order to test the robustness of the regression results, this study uses the model replacement method to replace the Probit model with the Logit model for estimation.In addition to the model substitution method, the common robustness check methods also include variable substitution method.This study replaces the explanatory variables, replaces the measurement items of "transfer-out area" and "transfer-in area" with "transfer-out rent" and "transfer-in rent", and carries out regression analysis again.The robustness test results (Table 4) show that although the impact of land satisfaction on land transfer-out behavior fails to pass the 10% significance test, the regression coefficient is still positive, and the land rootedness and land dependence in land attachment significantly inhibit land transfer out at the level of 5%, respectively.This is basically consistent with the basic regression estimation results, indicating that the results are relatively robust and reliable.

Intergenerational Differences in the Impact of Land Attachment on Land Transfer
According to the previous analysis, the land attachment of farmers in different generations is different, so how can the intergenerational differences of land attachment be reflected in land transfer?From this aspect, we answer this question by grouping regression according to the three generations of new, middle-aged, and older generation.Considering the length of the paper, only the regression results with the explanatory variables of "transfer-out behavior" and "transfer-out area" are reported.The results (Table 5) show that the land attachment of the new generation of farmers has no significant impact on land transfer out, which may be because the new generation of farmers have higher market awareness and pay more attention to interests in land transfer out.Among the middle-aged generation of farmers and the older generation of farmers, land dependence has an inhibitory effect on land transfer-out behavior and land transfer-out area, but land satisfaction and land rootedness only have a significant impact on the middle-aged farmers, and the effect is completely opposite.Thus, H2a and H2b are partially verified.

Conclusions and Discussions
Based on the above analysis, this study mainly draws the following conclusions: (1) There is a significant correlation between land attachment and land transfer out, but not with land transfer in.On the whole, farmers' land attachment can negatively affect their transfer-out behavior and area.(2) Different dimensions of land attachment have different effects on land transfer out.Among them, land rootedness and land dependence have a significant negative impact on farmers' land transfer-out behavior and land transfer-out area, while land satisfaction has a significant positive impact on farmers' land transferout behavior, but has no significant impact on land transfer-out area.Why does land satisfaction promote land transfer out?The main reason may be that with the acceleration of urbanization and the continuous promotion of land policies, farmers' awareness of land use has gradually deepened.Even if they are unable to participate or invest in farming, they will transfer out of land rather than abandon land because of their inner satisfaction and love for the land.At the same time, it is precisely because of this satisfaction and love that many farmers are more cautious in considering the area of land transfer out.
(3) Different generations of land attachment have different effects on land transfer out.Among them, the land attachment of the new generation of farmers has no significant impact on land transfer out.In middle-aged farmers, land satisfaction and land dependence have a significant impact on land transfer-out behavior and land transfer-out area, but the effect is opposite.Land satisfaction is a positive impact, and land dependence is a negative impact.Besides, land rootedness only has a significant negative impact on land transfer-out behavior.Among the older generation of farmers, land dependence has a significant negative impact on land transfer-out behavior and land transfer-out area, while land satisfaction and land rootedness have no significant impact on land transfer out.Compared with the results of basic regression, the inhibitory effect of land rootedness on the transfer out of land of the older generation of farmers is no longer significant in group regression.It may be due to the decline of their labor ability and the popularization of old-age insurance, which reduce their concern and expectation for land.

Policy Recommendations
According to the above conclusions, the following suggestions are put forward: (1) The government should respect the subject status of farmers and pay attention to the differences of farmers' emotional demands.The Chinese government has repeatedly stressed that promoting land transfer and developing modern agriculture should be closely linked with China's national conditions.Since ancient times, Chinese farmers have had deep attachment to land, which should be paid attention to.On the one hand, the formula-tion and implementation of land policy should fully consider the "emotional" factors of farmers, follow the "people-oriented" principle, respect the wishes of farmers, and care about the emotional demands of different groups of farmers.On the other hand, in the practice of promoting land transfer, the executor should safeguard the legitimate rights and interests of farmers, absorb the opinions of different groups of farmers, and adopt the strategy of gradual guidance and promotion.(2) The urban and rural social security system should be improved to solve farmers' "security concerns" by classification.Most farmers believe that land transfer means the loss of "land security", which can easily result in resistance to it.Therefore, the government and society should actively promote the construction of a social security system and give full play to the alternative role of social security.On the one hand, they should strengthen and improve the rural social old-age security system and do a good job in the coordination and connection between the new rural old-age insurance and other security systems, which has a positive impact on the transfer of land by elderly farmers; on the other hand, they should speed up the establishment of a national mobile social security system and improve the level of social security, which can effectively solve the social security problem of non-agricultural employment of young and middle-aged farmers and reduce the resistance of land transfer.(3) Local governments should carry out employment skills training to improve farmers' employability.For farmers engaged in agriculture, the government should regularly conduct agricultural skill training, vigorously introduce advanced agricultural production technology, and raise their agricultural production and management skills.For non-agricultural employment farmers, the government should perfect the non-agricultural employment market, improve the employment security mechanism, and launch non-agricultural employment skills training in order to improve their non-agricultural employment ability and non-agricultural income.These may stimulate the transfer-in willingness of farmers who engaged in agriculture and the transfer-out willingness of farmers who engaged in non-agricultural employment.(4) The government should establish a sound agricultural land transfer trading market and improve the transfer supervision mechanism.The incompleteness of the property rights trading market and the loopholes in the supervision mechanism hinder the circulation of agricultural land to a certain extent.Therefore, on the one hand, they should establish a comprehensive and professional agricultural land transfer trading market, ensuring it has an open, fair, and standardized operation; improve farmers' sense of trust in the transfer market; and stimulate farmers' willingness to transfer.On the other hand, they should improve the land transfer supervision mechanism, strengthening supervision measures, standardizing land transfer behaviors, strengthening services, properly solving contradictions and disputes, and improving farmers' satisfaction.

Figure 1 .
Figure 1.Distribution map of sample area.

Figure 2 .
Figure 2. Theoretical analysis framework of this study.

Table 1 .
Factor analysis results of land attachment.

Table 2 .
Description of variable and descriptive statistics.

Table 3 .
The regression results of the impact of land attachment on land transfer., and * respectively represent significance at the statistical level of 1%, 5%, and 10%.The numbers in parentheses below the coefficient are robust standard errors.The "Dy/dx" is the average marginal effect, and the number in parentheses are Z statistics.a the reference group is 3 = Yuechi County.

Table 4 .
The robust tests of the impact of land attachment on land transfer.

Table 5 .
The intergenerational difference of the impact of land attachment on land transfer.