The Way toward Sustainability: Policy Attention Evolution of Chinese Local Governments to Promote Entrepreneurship of Returnees Based on Grounded Theory and Social Network Analysis

Abstract

As an important livelihood strategy choice to promote the sustainable development of themselves and regions, returnees’ entrepreneurship has played a vital role in the governance of all-level Chinese local governments since the 21st century. The policy attention reflects the importance that the local governments attach to this issue, and the evolution process also reflects relevant policies’ internal rules and characteristics. Based on this, this paper, taking 242 policy texts related to returnees’ entrepreneurship issued by local governments from 2000 to 2021 as the research objects, uses grounded theory and social network analysis in order to explore the policy attention evolution from the aspects of time, space, specific policy domains and cooperative relationship among policy subjects. Finally, this research finds the main conclusions as follows: (1) The policy attention evolution on time dimension has the characteristic of synchronism, different levels and “reciprocal U”. (2) The policy attention evolution on space dimension is closely linked to “pull” and “push” forces of local population flow and has neighborhood effects. (3) The policy attention evolution on specific policy areas will gradually spread and diversify with the change in local governance environment and stage goals. (4) The evolution trend of collaborative social network among policy objects has the characteristic of “loose-central-loose”.

1. Introduction

Entrepreneurship plays a pivotal role in a country’s innovation and long-term economic growth [1] and also helps solve the problem of social employment [2]. This is because it may contribute to the formation of many self-employed ventures as well as small and micro enterprises, which can help create significant numbers of jobs. In recent years, China has joined the ranks of upper-middle-income countries, although economic growth has gradually slowed down, entering a “new normal” stage of economic development. In this critical period of the country’s development, to achieve sustainable economic growth and successfully avoid the middle-income trap, a national strategy of promoting various groups’ entrepreneurship will play a crucial role [3].

Following reforms and opening up of the country’s economy over the past 40 years, China’s economic development is a history of industrialization, urbanization, and labor force migration [4]. The latter is one of the most important factors in the economic development of the “new” China. Among the labor force are “Returnees”, both a special group and a unique phenomenon emanating from the China’s population flow and labor transfer. Since this group was created in China’s special governance and political practice context and the global research is insufficient, the definition is still not clear. In addition, there is also a lack of a uniform theoretical viewpoint. In 2022, the National Health Commission defined this group as “people returning to the countryside from other places.” Thus, it can be seen, in China’s governance context, that the “entrepreneurship of returnees” means “people returning to hometown and starting a business in here”, with “hometown” in this context usually referring to economically backward areas, often in the countryside.

The livelihood strategy choice of starting a business in one’s hometown is an effective way to improve the development of returnees as well as their regions [5]. However, due to the factors, such as venture capital shortage, lack of policy support and entrepreneurial infrastructure, the enterprises built by returnees have to face some adverse conditions: low resist-risk ability, high dependency on natural resources, and damage to the local ecological environment. The possibility of entrepreneurship failure is greatly increased, and the income of them will be greatly reduced, even resulting in individual bankruptcy and family breakdown. As a result, providing more comprehensive policy support as well as optimizing the policy attention distribution pattern in order to increase the entrepreneurship sustainability of returnees has become imperative [6].

This paper creatively conducts a systematic qualitative and quantitative analysis of the policy texts issued by Chinese local governments from the perspective of policy attention as well as through the combination of programmatic grounded theory and social network analysis. The aim is to explore the policy attention evolution rules of Chinese local governments in dimensions of time, space, and specific policy domains, and the various tendencies of policy subjects participating in the cooperation network as well as how these rules help promote returnees’ entrepreneurship. It is found that the process of policy attention evolution of Chinese local governments is essentially a process of promoting the sustainability of returnees’ entrepreneurship. This paper also puts forward relevant policy suggestions to provide theoretical support for the political practice of Chinese local governments to promote the ability, quality, and especially the sustainability of returnees’ entrepreneurship. Finally, the overall research aim is to provide insights from China’s experience to help other countries explore ways to promote the sustainable development of returnees and their regions.

2. Literature Review

The concept of “sustainable entrepreneurship” was not clear until 2010, when JBV published a special issue emphasizing that “sustainable development” has become an important concept to reduce environmental damage and social problems and discussed the role of entrepreneurship in it [7]. The concept of “sustainable entrepreneurship” gradually received validity. Before sustainable entrepreneurship, academic discussion focused more on “green entrepreneurship” or “social entrepreneurship”. Some researchers included both in the category of sustainable entrepreneurship, but an increasing number of research studies emphasize the difference among them: green entrepreneurship emphasizes the ecological innovation [8] to carry out entrepreneurial behaviors based on green technology to realize the economic and environmental sustainable [9]; and social entrepreneurship emphasizes the innovation of product or process innovation in order to achieve social goals and solve social problems [10]. With the emergence of sustainable entrepreneurship, researchers have gradually paid attention to the possibility of the integration of economy benefits, society benefits and ecology benefits in entrepreneurship [11]. The studies note that there is much academic research on entrepreneurship, mainly focusing on three areas: the individual influencing factors of sustainable entrepreneurship, the institutional influencing factors of sustainable entrepreneurship, and business models for sustainable entrepreneurship and types of sustainable entrepreneurship.

First, the individual influencing factors of sustainable entrepreneurship include the identification and evaluation of sustainable opportunities, the willingness for sustainable entrepreneurship and the behavior of sustainable entrepreneurship. In the identification and evaluation of sustainable opportunities, it is found that the perception of entrepreneurs’ threat to natural or public environment, altruism and entrepreneurial knowledge [12] all affect the identification and evaluation of sustainable opportunities. In addition, factors such as entrepreneurial self-efficacy, pro-environmental values [13], and personal initiative [14] also have a significant effect on it. In the study of the willingness of sustainable entrepreneurship, sustainable attitudes and desire, such as altruism and external reward [15], sustainable entrepreneurial orientation [16], and business experience [17] will have dual influences on the willingness to start a sustainable business. In the research on the influencing factors of sustainable entrepreneurial behavior, sustainable orientation, beliefs that entrepreneurship can affect the world [18], the obtaining and maintenance of legality of the sustainable business [19], strategic management ability, operation ability, the ability to embrace diversity and interdisciplinary ability, systematic thinking, specification ability, forward thinking ability and interpersonal factors [20] all can influence the behavior for sustainable entrepreneurs.

Second, previous studies have shown that sustainable entrepreneurial activities are affected by macro, meso and micro factors. Among them, mandatory, normative and simulated isostructuralism pressures jointly promote sustainable entrepreneurial activities [21]. These pressures are also influenced by macro (e.g., environmental regulations in export markets), meso (intermediary organizations), and micro (owners’ sustainable values) factors. In addition, factors such as resistant users (micro influencing factor), ineffective government policies (meso influencing factor), economic and financing barriers (meso influencing factor), and linkages between R&D and users (between micro and meso influencing factor) [22].

Among them, the influencing factors of macro, meso and micro levels do not exist in isolation. It is found that both new and existing start-ups in sustainable entrepreneurship contribute to the formation of sustainable industries [23]. In addition, collective organizations of enterprises can provide a coordination mechanism for achieving goals, thereby minimizing cognitive and practical barriers to sustainable entrepreneurship [24]. The interaction between new ventures and public authorities, such as regional political coordination capacity [25] and political participation of sustainable entrepreneurs [26], can also have an obvious significance of sustainable entrepreneurial activities. Finally, in a conservative institutional environment (where there is no or fewer institutional supports), the activities of sustainable entrepreneurs will also have an impact on the sustainable market supervision, normative and cultural cognitive institutions [27].

Third, related research also discusses the business model of sustainable entrepreneurship. Based on the perspective of evolutionary economics, the evolution process of sustainable entrepreneurial business model can be included as “variation to election to reservation” [28]. In addition, the construction of sustainable business model should include at least the following five aspects: sustainable value should include the value form of economic, social and environmental benefits [29]; needs a sustainable value flow system among multiple stakeholders including natural environment and society [30]; requires a value network with new purpose, new design and new governance [31]; and the innovation of sustainable business model can be realized by internalizing the externality through product-service system (PSS) [32]. Finally, the existing studies holds that the sustainable entrepreneurial model should at least follow these principles: resource perpetuation principle, benefit combination principle, strategy satisfaction principle, quality management principle and contribution principle [33].

Finally, research on the types of sustainable entrepreneurship is abundant. Based on the dimensions of entrepreneurial sustainability and entrepreneurial performance, types of sustainable entrepreneurship can be divided into “static type”, “transformation type”, “responsible type” and “constructional type” [34]; based on the motivation of enterprise sustainable development, it can be divided into the types of “business driven”, “legal driven”, “lifestyle driven” and “value driven” [35].

In conclusion, existing research on sustainable entrepreneurship has accumulated relatively rich results, mainly focusing on the individual influencing factors of sustainable entrepreneurship, the institutional influencing factors of sustainable entrepreneurship, business models for sustainable entrepreneurship, and types of sustainable entrepreneurship.

However, research on sustainable entrepreneurship has three deficiencies: First, local governments’ policy attention determines policy formulation in promoting sustainable entrepreneurship, while also reflecting policy development direction from the local government perspective. Although existing studies on this topic conclude that public policies affect all aspects of entrepreneurial behaviors [36], few scholars discuss relevant policies of local governments to promote sustainable entrepreneurship from the perspective of policy attention. Second, returnees, as an important subject of sustainable entrepreneurship, optimizing the policy attention distribution pattern in order to promote the entrepreneurship sustainability of this group is of great significance for their own livelihoods and local development and should be taken more seriously; however, studies on this issue are relatively scarce. Third, most existing studies provide an interpretation of local governments’ policy on promoting sustainable entrepreneurship of returnees from the perspective of “what it should be” [37] and follow an “idealized” path [38]. This focuses on qualitative research from a logical reasoning standpoint yet lacks quantitative analysis of the reality of local government behaviors.

3. Research Design

3.1. Research Framework

This research adopts a method of combining programmatic grounded theory and social network analysis to analyze 242 policy texts issued by local governments in China between 2004 and 2021. Sources include official letters, official notices, official meeting summaries, laws and regulations, and development plans. Specifically, this study first conducted a programmatic grounded theory on the policy texts through the method of open coding, spindle coding and selective coding. Thus, the main categories that Chinese local governments focus on include improving infrastructure, innovating the service system, improving the policy guarantee system, and improving organization and implementation, then taking the “push–pull theory” as the core category of the theoretical model which influences factors to promote migrant workers starting a business in their hometown, to reflect a logical relationship between the above main categories, corresponding categories and concepts. This provides a focus for studying the policy attention evolution of local governments in dimensions of time, space, and specific policy domains. In addition, this study also conducts social network analysis on policy texts. First, Python and Jieba’s word segmentation tools are used to refine keywords of relevant policy subjects involved in policy texts. Second, the visualization tool Net Draw in Ucinet software is used to draw the collaboration social network graph of policy subjects participating in Chinese local governments’ promotion of new businesses in hometown in different periods. Based on this, the study explores the variation tendency of policy subjects participating in the cooperation network. Finally, according to the research conclusions, the policy recommendations are advanced, deficiencies of the research are summarized, and future research directions are proposed. The study creates a research framework to reflect the research process intuitively (Figure 1) and also raises four research questions:

• Q1: How does Chinese local governments’ policy attention evolve over time and help promote returnees’ entrepreneurship?
• Q2: How does Chinese local governments’ policy attention evolve in the space dimension and help promote returnees’ entrepreneurship?
• Q3: How does Chinese local governments’ policy attention evolve in specific policy domains and help promote returnees’ entrepreneurship?
• Q4: How does the cooperative relationship evolve among policy subjects and help promote returnees’ entrepreneurship?

3.2. Research Methods

3.2.1. Programmatic Grounded Method

Programmatic grounded theory is one of the three schools of grounded theory, which is derived from classical grounded theory. Grounded Theory was first proposed by Glaser and Strauss in The Discovery of Grounded Theory: Strategies for Qualitative Research. Glaser and Strauss published three books together, and then diverged in their scholarly focus due to their different understandings of grounded theory, which led to the founding of separate grounded theory schools. In 1978, Glaser published Theoretical Sensitivity, a prominent example of classical grounded theory. Later, in 1987, Strass published Qualitative Data Analysis, marking the birth of programmatic grounded theory.

Strauss and Corbin [39] co-authored Basics of Qualitative Research Analysis, leading to programmatic grounded theory spreading widely. In addition, Strauss’s students developed the programmatic grounded theory. For example, Clarke adopted the method of “situational analysis” to collect data [40], a method used to identify and innovate existing knowledge, as well as the non-artificial element scene and the actor element scene.

In the programmatic grounded theory, the core of the coding process is to make explicit the direction of the concept, while a change of concept direction will produce new concept connotations. A change of concept direction continues until theoretical saturation, a result of the interaction of different concept directions. With the 6C model (causes, contexts, contingencies, consequences, covariances and conditions), it is relatively simple to divide category aggregation into concepts. Coding is the discovery and naming of categories with various directions. Programmatic grounded theory divides the coding process into open coding, spindle coding and selective coding as shown in Figure 2.

The applicability and relevance of the programmatic grounded theory in this study are as follows: First, at present, there are few studies on the evolution of policy consequent to returnees starting a business in their hometown, and research dimensions and foci are not clear. As a qualitative method, programmatic grounded theory seeks continuously to compare, sort and refine both concepts and categories [41]. By sorting, summarizing and analyzing original data, new concepts and relations can be found and new theories constructed, which is conducive to the dimension confirmation of the research on policy attention evolution. Second, the programmatic grounded theory is a “bottom-up” stylized theoretical analysis method, obtained via in-depth analysis and comparison, producing relevant information helping to determine the classification of influence factors as well as to build the relationship between them [42]. This is used to comprehensively identify influence factors of attention and deep reasons of Chinese local governments’ policy attention evolution promoting returnees to start a business in their hometown. Third, there is now substantial research indicating that programmatic grounded theory is a scientific approach in the field of social science; as a result, both methodology and research procedures are relatively developed. In addition, the coding process of this method also has the characteristics of a step by step approach, strictness and a higher level of procedure, helping provide feasible guidance for the new qualitative research [43].

3.2.2. Social Network Analysis Approach

In 1954, John A. Barnes explicitly proposed the concept of “social network”. He pointed out that the social structure is the expression of the relationship model between social actors, and social network analysis is the basic approach of social structure research. Barnes used a social network analysis approach to inform pioneering research on the social structure of Norwegian fishing villages. This helped make the concept of “social network” more widely known, including in relation to innovative explorations of communities, people, occupations and social capital [44]. J. Clide Mitchell later emphasized that the application of the concept of “social network” should pay attention to the analysis of micro-interpersonal relationships in the context of social structural research [45]. In the 1960s, he organized an anthropological field survey with a structural perspective in Zimbabwe, using collected data in the context of the sociometric method of graph theory [46].

Social network analysis is a quantitative research method from the perspective of “relationships”. It studies social phenomena and social structure by analyzing and discussing network structures and attributes. In political practice, there is a great need for cooperation and close economic and technical relations among policy subjects, manpower, material resources, technology and other elements, which together form a closely cooperative network. Each policy subject is regarded as the “node” in the network. Compared with orthodox research methods, such as questionnaires and interviews, social network analysis can transform the complex interactive relationship of various policy subjects into simple and convincing image data through software as well as focusing on the structure and closeness of the relationship between actors in the network, and present them in matrix and schema intuitively, flexibly and quantitatively.

In conclusion, this paper conducts programmatic grounded theory in studying policy attention evolution of local governments in dimensions of time, space, and specific policy, such as external characteristics, by using the social network analysis method on exploring internal cooperation relationships and the closeness among policy subjects in different periods. Thus, the research purpose can be realized more comprehensively and systematically because of the two methods complementing each other.

3.3. Data Collection

In total, 419 policy texts come from all levels of Chinese local government portal websites and the “Laws and Regulations of Peking University Library” by taking “local government” “returnees” and “starting a business in hometown” as keywords. Moving to artificial screening and the removal of invalid policy texts, the focus was on whether policy contexts reflect the consistency of Chinese local governments’ governance behavior toward encouraging returnees to start a business. Specifically, the research identified and removed repeat policies, workshops and other meetings which did not satisfy the requirements of this research. Finally, 242 valid policy documents were collected relating to Chinese local governments’ encouragement of returnees to start a business in their hometown, between 2004 and 2021.

4. Data Analysis

4.1. Analysis of Programmatic Grounded Theory

4.1.1. Open Coding

Opening coding is the first coding program of programmatic grounded theory. It presents original data as a series of important concepts and categories through intensive analysis and summary. In the process of open coding, researchers do not hold any presupposition; the extraction of relevant concepts and categories is based on the analysis and comparison of original data word by word and sentence by sentence. In order to identify the evolution of policy to encourage returnees to start a business in their hometowns by Chinese local governments, in relation to time, space and specific policy domains, the paper first sorted out the original data (policy texts) and removed similar and repeated sentences. Through the policy text analysis, this research extracted 657 basic concepts, and determined their meaning and nature and categorized related concepts under the same categories based on logical relationships. The categories were named. Finally, 354 corresponding categories were extracted through categorization.

Table 1. Examples of open coding categorization.

Original Statements	Concepts	Corresponding Concepts
The municipal government sets up special support funds of CNY 15,000,000 for people returning home to start businesses to provide financing support for home-returning entrepreneurs, as well as training and typical awards for entrepreneurship.	Strengthening governmental guidance and utilizing venture capital funds	Effective financial service
The government can provide secured loans for startups of utmost CNY 150,000 for eligible home-returning entrepreneurs who engage in individual operations and loans of up to CNY 450,000 for starting businesses in partnership or organization.	Providing social security subsidies for people with difficulties in starting a business	More financial support
Municipal and county administrative examination and approval service halls open comprehensive service windows for home-returning entrepreneurs, achieving the goal of handing at one time for returning home to start businesses.	Abolishing and delegating the power of administrative licenses	Lower standard
Home-returning business operators with the potential for sustainable development and leading will be selected to conduct training activities. Municipal startup training camps provide subsidies for entrepreneurs for CNY 15,000 each year, which will be disbursed from municipal support funds for entrepreneurship and employment.	Cultivating outstanding individuals as home-returning entrepreneurs	Publicizing and guide returnees to start a business
Strengthen internet training in all walks of life and set up e-commerce training courses for people returning home to start businesses. Additionally, it encourages home-returning people with master’s degrees or above to operate e-commerce businesses in Heze.	Encouraging people returning home to start businesses relying on the platform and business network	Enhancing online and offline infrastructures for internet startups
Home-returning business operators with the potential for sustainable development and leading will be selected to conduct training activities. Municipal startup training camps provide subsidies for entrepreneurs for CNY 15,000 each year, which will be disbursed from municipal support funds for entrepreneurship and employment.	Cultivating outstanding individuals as home-returning entrepreneurs	Making great efforts to publicize and guide returnees to start a business

As the space is limited, only the first emerged concepts and corresponding concepts are displayed in Table 1.

shows the process.

4.1.2. Spindle Coding

Spindle coding is the second stage of implementing the coding program of programmatic grounded theory. It is a process involving further research into many corresponding categories formed by open coding in order to obtain the main categories with more generality and a higher level of abstraction. This process aims to discover and construct all kinds of relations among concepts, corresponding categories and main categories, including the causality relationship, temporal relationship, precedence relationship, situational relationship and similarity relationship, so as to organically connect the corresponding categories obtained in the previous stage. This research summarizes the 657 concepts obtained in the open coding stage, and 15 corresponding categories were obtained. Then, on the basis of corresponding categories, four main categories were extracted and obtained: completing infrastructure, innovating service system, improving policy guarantee system, practicing organization and implementation.

Table 2. Examples of spindle coding categorization.

Main Categories	Corresponding Categories	Concepts
Completing infrastructure	Developing business incubators for home-returning people relying on the integration of stock resources	Offering support in relation to returning entrepreneurs’ basic information and needs.
	Enhancing the development of online and offline infrastructure for internet startups	Encouraging telecommunication enterprises to increase investment in developing internet and mobile internet.
	Enhancing the infrastructure construction of grassroots service platform	Promoting employment and social insurance service platforms in both counties and countryside.
Innovating service system	Reducing barriers to start business in the hometowns of returnees	Completing registration for people returning home to start businesses
	Strengthening enterprise training for people returning home	Implementation of special training projects.
	Completing public service for people returning home to start businesses	Taking into consideration overall public service system reforms, such as social insurance, housing, education, and medicine.
	Improving intermediary services for people returning home to start businesses	Supplying service-related information to market analysis, management and guidance.
	Guiding the connection of “mass innovation” and returning home to start businesses	Providing science and technology service for them.
Improving policy guarantee system	Implementing policies to cut taxes and fees	Corporate income tax reduction.
	Providing more financial support	Offering people returning home to start businesses asocial insurance subsidy.
	Effective financial service for people returning home to start businesses	Strengthening governmental guidance and utilizing venture capital fund.
	Completing support system of business incubators for home-returning people	Completing infrastructure of business incubators, for instance, water and electricity, transportation, communication and broadband network.
Practicing organization and implementation	Strengthening the demonstration–promotion ability	Playing a demonstration role by developing a demonstration base of counties’ internet startups.
	Making great efforts to publicize and guide	Strengthening publicity and interpretation in a popular form for people returning home to start businesses.
	Improving organization and coordination	Making sure the efficiency of work through specific schedules and detailed rules of implementation.

As the space is limited, only the representative concepts and corresponding concepts are displayed in Table 2.

shows the main categories, corresponding categories and their concepts.

4.1.3. Selective Coding

Selective coding is the third stage of implementing the coding program of programmatic grounded theory. The main task of selective coding is to make further analysis of the main categories obtained by the spindle coding and discover the core category with stronger generalization to connect all the other categories so as to form a “storyline” that can run through the whole process. This research, through the investigation of the above 4 main categories and 15 corresponding categories, identifies “the pull force of immigratory place” in “push and pull theory” as the core category to summarize all concepts, corresponding categories and main categories.

Table 3. Typical relationship structure of main categories.

Typical Relational Structure	Connotation of Relational Structure
Completing infrastructure → Consolidating the entrepreneurial foundation for returnees → Increasing the pull force in immigratory place	The construction of entrepreneurial infrastructure is the material foundation for migrant workers to start their businesses in their hometown. Measures such as developing business incubators for home-returning works relying on the integration of the stock resources, enhancing the development of online and offline infrastructure for internet startups as well as grassroots service platform provide higher-quality platforms for migrant workers starting business in their hometown, thus improving the “pull force” of the immigratory place directly.
Innovating service system → Optimizing the entrepreneurial environment of returnees in their hometown → Increasing the pull force in immigratory place	The support and guarantee of public policies are conducive to solving the problem of lack of entrepreneurial resources of returnees and improving their entrepreneurial risk-resistance capacity. Measures such as reducing barriers to start business in their hometowns, implementing policies to cut taxes and fees and providing more financial support are conducive to improving policy for returnees, to improve the “pull force” of the immigratory place directly.
Improving policy guarantee system → Solving the problem of lack of entrepreneurial resources of returnees → Improving their entrepreneurial risk-resistance capacity → Increasing the pull force in immigratory place	The support and guarantee of public policies are conducive to solving the problem of lack of entrepreneurial resources of returnees and improving their entrepreneurial risk-resistance capacity. Measures such as reducing barriers to start business in their hometowns, implementing policies to cut taxes and fees and providing more financial support are conducive to improving policy for returnees, to improve the “pull force” of the immigratory place directly.
Practicing organization and implementation → Stimulating endogenous motivation for returnees → Improve their entrepreneurial quality → Increasing the pull force in immigratory place	Improving and expanding the quality and scope of public policy implementation for migrant workers starting business in their hometown is the fundamental influence factor of the success of their entrepreneurship, involves implementing policy and organizing accurately, necessary to stimulate endogenous motivation for returnees and improve their entrepreneurial quality. Measures such as improving organization and coordination, strengthening the demonstration–promotion ability and making great efforts to publicize and guide are conducive to further increasing the quality of public policies, thus improving the “pull force” of the immigratory place directly.

shows the typical relationship structure among them.

“The pull force of immigratory place” is a concept proposed by Lee [47] based on the “Law of population migration” originally stated by Ravenstein [48] and improved by introducing personal factors, as Lee believes that there are both push and pull forces in the immigratory and migratory places, and there are intervening factors between them. Lee forms an interpretative framework of population migration based on the push and pull theory, comprising “two dimensions and four factors”. “Two dimensions” refers to the push and pull forces both in the immigratory place and migratory place. “Four factors” refers to “immigratory place factors” “migratory place factors” “personal factors” and “intervention factors”. The driving force of population migration is the result of the difference between the push force and pull force in the interaction of “four factors”. This research found that the policy objective of four main categories obtained through programmed grounded theory is to increase pull force of the immigratory place. Therefore, this article builds a theoretical framework of influencing factors of returnees starting a business in their hometown on the basis of the “two dimensions and four factors” framework so as to provide a reference for subsequent research (Figure 3).

4.1.4. Theoretical Saturation Test

Theoretical saturation is the evaluation criteria for stopping coding, that is, no additional data can be obtained to enable researchers to discover new categories, or even if new variables can be discovered, further data collection and analysis would make relatively little new contribution to the conceptualization. This research’s analysis of reserved policy text in the same coding method found no new concepts and categories, and the relationship between the categories has not changed significantly. Therefore, the typical relationship structure among concepts, corresponding categories and main categories built in this paper is theoretically saturated.

4.2. Analysis of Social Network Approach

4.2.1. The Statistical Analysis of Keywords Frequency

The frequency of keywords is the intuitive reflection of evolution in the cooperative relationship among various policy subjects in the process of policy formulation and implementation of Chinese local governments in promoting returnees to start businesses in their hometowns. To facilitate subsequent analysis as well as making the results more accurate, this research first obtained the frequency boundary between high frequency words and low frequency words based on Zipf’s law [49], which can provide a sound basis for determining their point. Then, using Python and Jieba Segmentation extracts, keywords of policy subjects were used in the process of policy formulation and implementation between 2003 and 2021. Finally, due to differences in the setting and names of departments among Chinese local governments, this research unified 29 keywords according to their main functions and constructed the keyword list on the grounds of keyword frequency (

Table 4. The Frequency and Proportion Table of Keywords.

Keywords	Frequency	Proportion
City people’s government	201	31.7%
Ministry of Human Resources and Social Security	156	24.6%
Provincial people’s government	57	9.0%
Country people’s government	28	4.4%
Administration committees of the development zone	27	4.3%
Administration committees of the high-tech zone	23	3.6%
Administration committees of industrial park	23	3.6%
Ministry of Finance	20	3.1%
Public institutions	18	2.8%
National Development and Reform Commission	18	2.8%
Banks and financial institutions	10	1.6%
Colleges, universities and vocational schools	7	1.1%
Enterprises	6	0.95%
Tax Bureau	6	0.95%
Demonstration areas of urban-rural integration	5	0.79%
Administration committees of the enterprise zone	5	0.79%
Administration committees of the resorts	3	0.47%
Administration committees of the medical high-tech zone	3	0.47%
People’s organizations	3	0.47%
Administration committees of the economic cooperation zone	3	0.47%
Economic and Information Technology Commission	2	0.31%
Department of Land and Resources	2	0.31%
Department of Housing and Urban-Rural	2	0.31%
Agriculture Department	2	0.31%
Department of Commerce	1	0.16%
Civil Affairs Bureau	1	0.16%
Administration committees of overseas Chinese economic zone	1	0.16%
Administration Committees of international inland port zone	1	0.16%
Department of Transportation	1	0.16%

).

4.2.2. The Social Network Analysis of Keywords

On the basis of the above research, this paper uses Net Draw, a visualization tool in the Ucinet, to collect the cooperation number of policy subjects in order to analyze the social network of cooperation among them in three different periods and to draw visual network diagrams. The dimensions of breadth and strength of the cooperation social network in different periods were used to investigate the relationship changes of different policy subjects in the cooperation social network.

More precisely, the breadth index is a very important reference of the number of subjects and reflects the formation process of “multi-subjects coordination” mode and the precision of policy implementation. The network cohesion index reflects the cooperation frequency and closeness among subjects: the closer it is to 1, the higher the coordination degree among subjects.

5. Conclusions

5.1. The Policy Attention Evolution on Time Dimension Has the Characteristic of Synchronism, Different Levels and “Reciprocal U”

On the basis of the main categories obtained from the programmatic grounded theory, namely, the main domains of Chinese local governments to encourage returnees to start a business in their hometown, this study identified in each year the number of policy texts related to the above four main domains. This reflects the allocation and evolution of the policy attention of Chinese local governments to promote the entrepreneurship of returnees in the dimension of time (Figure 4). This mainly reflects the characteristics of synchronism, different levels and “reciprocal U” of the above-mentioned four main domains on the time dimension.

Before 2015, due to a lack of guidance from central government, local governments explored and formed their own policies to encourage returnees to start their own business, so the number of policy texts is small. In 2015, the General Office of the State Council issued a national public policy to encourage returnees to start a business in their hometown. After 2015, the local governments’ policy guidance and decision-making basis on this issue was gradually made clear. They established policies within the scope of the administration to make a superior policy, with less pressure of synchronous coordinates in policy attention to encourage returnees to start a business in their hometown. How Chinese local governments contact each other also relies on superior policy requirements and administrative structures and is somewhat fragmented. This avoids the problem of governance failure caused by policy attention imbalance between different local governments and their departments [50] and forms a synergy between policies and actions in order to promote the returnees’ entrepreneurship as well as the synchronization of main domains in promoting the entrepreneurship of returnees.

In 2015, The General Office of the State Council issued The Opinions on Supporting returnees and Others to Start Businesses in Hometown. In 2017, the National Development and Reform Commission, the General Office of the Ministry of Agriculture and Rural Affairs, and the Administrative and Human Resources Department of the Poverty Alleviation Office of The State Council jointly issued The Notice on Implementing the Five-Year Action Plan for Training returnees and Others to Start Businesses in hometown. As a result, the policy attention of Chinese local governments, which was originally relatively fragmented, gradually focused on completing the infrastructures, innovating the service system, improving the policy guarantee system, and improving organization and implementation. Under the sync actions of the state and local governments, the policy attention distribution of Chinese local governments to promote the entrepreneurship of returnees changed from strong to weak in relation to improving the policy guarantee system, innovating the service system, completing infrastructures, and improving organization and implementation, with variable levels of success in different policy domains. The focus of local governments’ policy attention is conducive to establishing and consolidating the policy support, service, and system foundation of returnees’ entrepreneurship in order to create favorable external and internal environments.

With the development of the governance of Chinese local governments to promote returnees’ entrepreneur, such as the construction of a business park, and the improvement of financial and policy supports, a better public service system and other infrastructure improvements made progress. After 2017, public policies related to completing infrastructure, innovating the service system, improving the policy guarantee system, improving organization and implementation gradually declined from this peak. At the same time, with the improvement of the governance basis of local governments in promoting the entrepreneurship of returnees, policy attention turned to improving the stability of entrepreneurship and the ability to resist market risks step by step. This helped promote the high-quality and sustainable development of enterprises, and also encouraged the allocation of the policy attention of Chinese local governments to move to a new state of competition and diversification, while the original distribution structure of policy attention was both dispersed and adjusted. As a result, the evolution of policy of Chinese local governments to promote the entrepreneurship of returnees shows an obvious “reciprocal U” structure on the time dimension.

It should be noted that the evolution characteristics of policy attention of Chinese local governments to promote the entrepreneurship of returnees are not static and unchangeable, but a dynamic process. The “reciprocal U” structure does not mean the decline of policy attention but indicates the diversification and richness of policy domains; it is a “relative reciprocal U” structure in the whole development cycle and policy domain. At the same time, the domain of improving policy guarantee shows a minor peak after 2019, due to COVID-19 outbreak, with a great impact on the entrepreneurship of returnees. Therefore, Chinese local governments issued relevant policies to ensure the orderly entrepreneurship activities of returnees in the background of normalized epidemic prevention and control. However, considering the sudden impact of this event, this study does not include an explanation and analysis of the policy evolution of Chinese local governments in the time dimension.

5.2. The Policy Attention Evolution on Space Dimension Is Closely Linked to “Pull” and “Push” Force of Local Population Flow, and Has Neighborhood Effects

The evolution of the policy attention of Chinese local governments to encourage migrants and returnees to start a business also requires investigation of the influence of regional attributes. This study adopts the homogenous division method based on Strategies and Policies for Coordinated Regional Development released by the Development Research Center of The State Council in 2005, which divides China’s local governments (excluding Hong Kong, Macao and Taiwan) into eight economic zones according to their spatial proximity, similar natural conditions, natural endowments and social structures. This method can better control the economic differences within different regions, on the basis of internal homogeneity.

Table 5. Distribution of policies to promote entrepreneurship of returnees in different regions.

Regions	Provinces	The Number of Policy Releases	The Proportion of Policy Releases (%)
Northeast Comprehensive Economic Region	Heilongjiang Province, Jilin Province, Liaoning Province	22	10.95
Northern Coastal Comprehensive Economic Region	Beijing, Tianjin, Hebei Province, Shandong Province	13	5.39
Southern Coastal Comprehensive Economic Region	Fujian Province, Guangdong Province, Hainan Province	15	6.22
Eastern Coastal Comprehensive Economic Region	Shanghai, Jiangsu Province, Zhejiang Province	7	2.91
Middle Reaches of Changjiang River Comprehensive Economic Region	Hunan Province, Jiangxi Province, Anhui Province, Hubei Province	42	17.43
Northwest Comprehensive Economic Region	Gansu Province, Qinghai Province, the Ningxia Hui Autonomous Region, Tibet Autonomous Region, the Xinjiang Uygur Autonomous Region	10	4.15
Southwest Comprehensive Economic Region	Yunnan Province, Guizhou Province, Sichuan Province, Chongqing, Guangxi Zhuang Autonomous Region	63	24.32
Middle Reaches of Yellow River Comprehensive Economic Region	Shaanxi Province, Shanxi Province, Henan Province, the Nei Monggol Autonomous Region	69	28.63

shows the number of policies releases in the following eight economic zones from 2004 to 2021, and indicates the space dimension of the distribution and evolution of policy attention and the space dimension of Chinese local governments to promote the entrepreneurship of returnees, showing the following two characteristics.

First, the policy attention evolution of Chinese local governments to encourage returnees to start a business in their hometown on the space dimension is closely linked to “pull” and “push” forces of local population flow. According to Table 5, local governments in the Middle Reaches of Yellow River, Comprehensive Economic Region, Southwest Comprehensive Economic Region and Middle Reaches of Changjiang River Comprehensive Economic Region have the highest proportion of policies to promote the entrepreneurship of returnees, indicating that the above regions are the most active and significant in the policy development. This is based on the desire of the above regions to increase the “pull force” and reduce the “push force” of local population flow into their regions. As already mentioned, according to the “push–pull theory” of population mobility, population mobility involves both migratory and immigratory places. As a result, differences and similarities of attributes between different regions or cities play an important role in the decision of whether returnees will choose to return to their hometown to start a business. Compared to the Eastern Coastal Comprehensive Economic Region and Northern Coastal Comprehensive Economic Region, the Middle Reaches of Yellow River Comprehensive Economic Region and the Southwest Comprehensive Economic Region are far behind in terms of economic development level, infrastructure construction and public service resources. In addition, employment opportunities, living conditions and other aspects also have no advantages.

Therefore, the “push force” of population mobility in these regions is relatively large, which also explains why Henan Province, Guizhou Province, Sichuan Province, Guangxi Zhuang Autonomous Region and Anhui Province are the provinces with the largest number of labor service exports in China. As a result, these regions have a stronger will to promote the “push force” through policy to develop the local infrastructure construction, service system and policy guarantee as well as improving employment and living conditions in order to reduce the “push force” to attract returnees back to their hometown to start a business. This is the reason why these regions are the most active in the development of relevant policies. In contrast, regions with strong “pull force” and weak “push force”, such as Northern Coastal Comprehensive Economic Region, Southern Coastal Comprehensive Economic Region and Eastern Coastal Comprehensive Economic Region are considered migratory places. Large-scale population migration will make these regions face the reality and challenges of a complicated social structure, frequent population flows, diversified interest demands and varying values in urban governance. Therefore, they will have relatively little desire to attract returnees to return to their hometowns and start a business, thus, the proportion of policies release is low. Therefore, the policy attention distribution pattern of local governments should be further optimized and improved, and in particular, the attraction of economically backward areas for returnees to start businesses should be promoted gradually via reducing the “push force” and increasing the “pull force”.

Second, neighborhood effects also exist in the policy attention evolution of Chinese local governments to encourage returnees to start a business in their hometown on the space dimension. Through the study of Southwest Comprehensive Economic Region and Middle Reaches of Yellow River Comprehensive Economic Region, which have the most active policies to promote the entrepreneurship of returnees, this paper indicates that with Sichuan Province and Henan Province as the center, it has the strongest neighborhood effects among local governments in the regions. Furthermore, compared to the policies’ distribution in different local governments to enourage migrant workers starting a business in their hometown, it can be grasped quite clearly that the Southwest Comprehensive Economic Region and Middle Reaches of Yellow River Comprehensive Economic Region have central provinces: Sichuan Province and Henan Province. Within these regions, the local governments surrounding the center provinces issue significantly more policies than other regions and provinces, with a remarkable neighborhood effect, stemming from competition with and studying each other. With the Southwest Comprehensive Economic Region as an example, Sichuan Province issued Application Opinion About Support for rural returnees and Farmers in 2015. This was the earliest guide of the Southwest Comprehensive Economic Region to focus on related policies in order to encourage returnees to start a business in their hometown. It was followed by Sichuan Province, Chongqing, Yunnan Province and Guizhou Province, which also issued relevant policies. By 2016, all provinces in Southwest Comprehensive Economic Region had issued relevant policies. In the following years, Sichuan Province consistently took the lead in issuing policies and other provinces in this region followed, continuously releasing relevant policies with their own characteristics to promote returnee entrepreneurship. For example, Chongqing took the initiative in proposing a policy to adjust the interest subsidy for key returnee entrepreneurship enterprises. Thus, it can be seen that the policy attention evolution of Chinese local governments to promote returnees to start a business in their hometown on the space dimension has neighborhood effects, reflected in the interaction and development of policy attention of different local governments in the same region linked to geographical relationship, competition and learning behaviors. Moreover, the competition and learning behaviors among local governments in the same region will also promote the optimization of policies related to entrepreneurship of returnees as well as having a ripple effect for the policy attention distribution pattern from other angles.

5.3. The Policy Attention Evolution on Specific Policy Domains Will Gradually Spread and Diversify with the Change of Local Governance Environment and Stage Goals

Based on the main categories and corresponding categories obtained from the programmatic grounded theory, this paper statistically analyzed the number and proportion of relevant policy texts (

Table 6. The number and proportion of relevant policy release.

Specific Policy Areas		The Number of Policy Releases	The Proportion of Policy Releases (%)
Completing infrastructure	Developing business incubators	59	24.28%
	Enhancing the development of online and offline infrastructure	12	4.96%
	Enhancing the infrastructure construction of grassroots service platform	21	8.68%
Innovating service system	Intensifying the enterprise training	24	9.92%
	Completing public service	16	6.61%
	Improving intermediary services	5	2.07%
	Guiding the connection of “mass innovation”	6	2.48%
Improving policy guarantee system	Lower standard for starting business in hometown	8	3.31%
	Implementing policies to cut taxes and fees	20	8.26%
	More financial support	20	8.26%
	Effective financial service for starting business in hometown	21	8.68%
	Completing support system of business incubators	17	7.02%
Practicing organization and implementation	Improving organization and coordination	5	2.07%
	Strengthening the demonstration effect	3	1.24%
	Making great efforts to publicize and guide	4	2.16%

) in order to discover key domains of policy attention evolution of Chinese local governments to promote the entrepreneurship of returnees in specific policy areas.

Table 6 indicates that policy attention evolution in specific policy areas will gradually spread and diversify with the change of local governance environment and stage goals. At present, Chinese local governments mainly focus on the infrastructure construction of returnee entrepreneurs, especially the development of business incubators. The reason is that in the initial stage of encouraging returnees to start a business, the phased policy goal of Chinese local governments is to build and improve the related material infrastructure, to provide a high-quality platform for returnee entrepreneurship, including completed industrial parks and business incubators. It is also provided favorable development conditions and environment for entrepreneurship, to improve their success rate [51]. In addition, business incubators with an industrial agglomeration effect can prompt new enterprises to improve their own strength, encouraging them to provide better high-quality products and services for the local areas, to improve the production and living conditions of the hometown, and to make contributions to the optimization and sustainable development of the local industrial structure [52]. At the same time, the development of the online and offline infrastructure as well as grassroots service platform is also conducive to providing a more complete infrastructure guarantee and platform for returnees to start a business to encourage their sustainable development.

While the construction of the returnee entrepreneurial infrastructure is gradually improving, the governance environment of Chinese local governments to promote returnee entrepreneurship is getting better. Phased policy goals have shifted to improve the construction of the entrepreneurial infrastructure while optimizing the entrepreneurial environment. As can be seen from Table 6, except for completing infrastructure, the proportion of policy releases related to improving the policy guarantee system is the highest, reaching 35.53%, including measures such as lowering the standard for starting a business in a returnee’s hometown, implementing policies to cut taxes and fees, and providing more financial support. Such policies will have a positive influence on optimizing policy support, financial services, social and economic factors, and science and technology environments for returning entrepreneurs, thus providing them with a good business environment [53]. At this stage, local governments in China are gradually stimulating the willingness of returnees to start a business by building high-quality entrepreneurship platforms, providing a superior entrepreneurship environment, and gradually improving the quality of their entrepreneurship.

Finally, with the gradual improvement of the infrastructure construction and entrepreneurial environment, the phased policy goals of Chinese local governments to encourage returnees to start businesses in their hometown has also begun to shift. In addition, the number of policy releases in innovating the service system and practicing organization and implementation has increased. This also reflects the fact that policy attention is saturated in the policy areas of completing infrastructure and guaranteeing policy, and policy attention is also increasing in other diversified policy areas. To sum up, the policy attention evolution of Chinese local governments to encourage returnees to start a business in their hometown in relation to specific policy areas is gradually spreading and diversifying with the change in the local governance environment and stage goals.

5.4. The Evolution Trend of Collaborative Social Network among Policy Objects Has the Characteristic of “Loose-Central-Loose”

Based on the above social network analysis results of policy text released by Chinese local governments to encourage returnees to start a business in their hometown, this study takes 2015 as the cut-off point. It divides changes in the cooperation relationship among policy subjects in promoting the entrepreneurship of returnees by Chinese local governments into three phases: The period from 2003 to 2015 is the initial exploration phase, the period from 2015 to 2016 is the promote development phase, and from 2016, there is the deepening development phase. The visual network diagrams are drawn according to the analysis result.

In the initial exploration phase (Figure 5), the breadth index is 7, and the network cohesion index is 0.913. This shows that policy subjects’ participation in promoting the entrepreneurship of returnees mainly include people’s governments at all levels, as well as the Ministry of Human Resources and Social Security and administration committees of industrial parks. The policy subjects involved are few and the breadth is low. In terms of cooperative processes, local governments issue leading guide policies, and other policy subjects implement policies according to policy requirements. The whole collaboration social network shows the characteristics of giving priority to local governments, with administrative innovations being the main way of implementing policies. Although there is a high intensity relationship among the policy subjects in this phase, it is still in essence an administrative, task-oriented, and loose cooperation network.

In the promotion development phase (Figure 6), the breadth index is 23, and the network cohesion index is 0.98. A total of 23 policy subjects were involved. It is the stage in which the largest number of policy subjects participate, and the intensity relationship between them is extremely significant. Improvement mainly benefits from two aspects: First, after the initial exploration phase, the Chinese local governments’ policy goal to encourage returnees to start a business in their hometown is clear, and as the improvement of returnees’ entrepreneurial infrastructure and the construction of foundation platform develops, the work can penetrate into broader policy fields. Second, central government has continuously issued relevant policies to promote the entrepreneurship of returnees, encourage a positive response of local governments and integrate the promotion of returnee entrepreneurship with local governance and sustainable development. In this phase, the policy subjects and application fields involved in promoting returnees’ entrepreneurship gradually diversify, showing characteristics of multi-center radiating outward in the cooperation social network. In addition, compared with the initial exploration phase, the participation of financial departments such as the banks, the Ministry of Finance and the Tax Bureau, reflects policy attention to promote the entrepreneurship of returnees, gradually shifting from improving the infrastructure to innovating service system and improving policy guarantees. In brief, the cooperation social network among policy subjects in this phase shows characteristics of high-intensity, high-breadth, and multiple-balanced policy.

In the deepening development phase (Figure 7), the breadth index is 17, and the network cohesion index is 0.426. This shows that the policy objects and practices participating in promoting returnees’ entrepreneurship are in decline. A total of 17 policy subjects are involved, and the relationship among them is significantly lower, suggesting that the policy attention of promoting returnees’ entrepreneurship gradually disperses. However, from the perspective of specific policies, although the lower intensity relationship means the reduction of cooperation among policy subjects, it also means the improvement of the independence of policy subjects. This reduction trend is the result of the policy attention of local governments to promote the entrepreneurship of returnees into a wide range of people’s livelihood applications and government service. Therefore, at this phase, the cooperation social network shows the characteristics of independence and dispersion, an important development reflecting a shifting policy attention toward accurate practice and application. In addition, compared with the above two phases, colleges and vocational schools as the policy subjects are also involved, indicating that the process of returnees’ starting a business in their hometown has grown from traditional returnees, as the main body, to include university graduates and new professional farmers. This development embodies a new direction of evolution of the policy attention of Chinese local governments to encourage returnees to start a business in their hometown.

To sum up, before 2016, the policy attention evolution of collaborative social network among policy subjects had the characteristic of highly fragmented and segmentary relationships, high-strength, low-breadth, simple and loose as the country’s emphasis and policy support on promoting returnees’ entrepreneurial skills. Chinese local governments and policy subjects responded positively, with their collaborative social network showing a diversified and balanced pattern of high intensity and high breadth. After 2016, with the gradual improvement of the relevant infrastructure and the entrepreneurial environment, Chinese local governments dispersed policy attention on promoting returnee entrepreneurship and moved to other policy and social fields; therefore, the collaborative social network consists of policy subjects with an overall trend of “loose-central-loose”.

6. Discussion

6.1. Theoretical Contributions

First, the extensive studies on the impact of policies on entrepreneurial behavior are rich, such as implementing strategic management [54]. It illustrates the beneficial and value-oriented ways to promote the quality of entrepreneurship [55], and thus drive regional economic as well as sustainable development. There are, however, few studies from the perspective of policy attention evolution. This paper argues that, as a scarce resource, the distribution pattern and evolution rules of policy attention will significantly reflect the strength of support and critical focus area of local governments for returnees’ entrepreneurial behavior. The aim of this research, based on this perspective, is effectively to close the research gap in relevant fields.

Second, this paper innovatively combines programmatic grounded theory with social network analysis, and includes both the quantitative and qualitative research perspectives so as to systematically explore the evolution rules of policy attention.

Finally, in recent years, based on the “push–pull” theory, studies of migration movements and their entrepreneurial behavior have grown in number. Most of the existing studies involve independent research on the “push” and “pull” forces of the immigratory and migratory places. Previous studies have shown that the “pull forces” of economic backward areas, such as their hometowns, an improved entrepreneurial environment [56], stronger policy support [57] and lower business startup costs [58], will attract growing numbers of returnees to start businesses. The “push force”, such as worse living conditions and environment, higher human resource costs [59], fiercer competition [60] and lack of identity [61] in some developed areas, reduces entrepreneurial behavior. However, this paper believes that the decision of returnees to start a business in their hometown is not due to the above “pull” or “push” force alone, and consequently, analyzes entrepreneurial behavior from a systematic perspective. This paper constructs a theoretical framework of returnees’ influencing factors to start a business in their hometown from the dimensions of both the “push force” and “pull force” of the immigratory and migratory place in order to explain this behavior further. This approach can also provide a reference point for Chinese local governments to further improve their political practice of returnee entrepreneurship motivation, quality and sustainability.

6.2. Research Limitation

There are still some limitations in this study. First of all, due to the differences in the political system, economic development level and cultural values between China and other countries, especially some developed countries, there are some research paradigm differences on the policy attention evolution of local governments in promoting the entrepreneurship of returnees. Specifically, “returnees” is a special group generated in the context of China’s development process and governance. As a result, many other countries have very little research results, lacking a consensus definition and research paradigm. In addition, China and other countries are different in the definition and understanding of the concept of “local governments” due to differing forms of administrative division [62]. Therefore, it is possible that results reflect the weak generalization and referability of sustainable governance in other countries or regions around the world. Second, the selection of research samples in this research has certain limitations. The selection of research samples in this paper is based on a relatively single cultural background, focusing on policy documents issued by Chinese local governments on promoting the entrepreneurship of returnees. As a result, the nature of the research samples is relatively simple and single. Finally, the sample size and diversity still have much room for improvement.

6.3. Research Prospects

First, as an innovative way to solve social problems, social entrepreneurship is booming worldwide, gradually evolving into innovative thinking that breaks through market and government failures as well as imbalances in social development. Among those who have returned to their hometowns to start businesses, there are some entrepreneurs who carry out entrepreneurial activities for the purpose of realizing social value, devoting themselves to improving the social and economic development as well as living conditions of local residents. For example, “farming cooperatives”, as a unique economic organization with both enterprise and social attributes, are currently considered one of the most important forms of social entrepreneurship in China, helping to enhance the attractiveness of local returnees’ entrepreneurship [63]. However, current academic research on social entrepreneurship is in the initial stage of theoretical construction. Therefore, the policy attention distribution pattern of Chinese local governments to promote the entrepreneurship of returnees in the field of social entrepreneurship needs to be further optimized and improved.

Second, in terms of research methods, existing studies mainly use questionnaires to collect research data and establish influencing factor models to quantitatively describe and analyze the evolution of local governments’ policy attention. Yet their research objects obviously reflect both geographical and group specificity, as well as differences in quantitative methods for the same factors. This leads to results appearing to be inconsistent situation, while fully embodying differences in the policy attention distribution structure of Chinese local governments in promoting returnees’ entrepreneurial skills. This requires that future research choose typical cases with a specific area and group features to study and highlight actual conditions of the relevant environments. This would help to promote improved research value and significance, and also help provide relevant experiences for other countries to help narrow the gap between individuals and regions and enhance sustainable development capacity.

7. Suggestions

Taken together, the conclusions of this study have a number of implications.

1. Giving play to and enhancing the initiative of all local governments in the distribution pattern of policy attention. Local governments at all levels are the administrative subject to promote returnees entrepreneurial as well as the main sectors of promoted policy [64]. Therefore, various functional departments in local governments should study and judge the difficulties and dangers faced by local returnees sufficiently within the scope of their respective responsibilities, actively explore a more effective governance path to promote returnees’ sustainable development capacity and strive to turn the advantages of returnees’ entrepreneurship into governance efficiency that drives local sustainable development. In addition, local governments should give full play to their initiative, improve the policy attention distribution pattern, better collaborate between departments in the spirit of innovation and independence, aim to construct a strategic pattern of local sustainable development and optimize the distribution pattern of policy attention.

2. Promoting the transformation of collaborative governance model among policy subjects in the policy attention distribution pattern. Local governments do not work independently in promoting the entrepreneurship of returnees but need to work toward the full participation of all policy subjects [65] and form a good interactive relationship to promote the transformation of collaborative governance model of various subjects in the policy attention distribution pattern of returnee entrepreneurship policy. This involves moving from the single-center model dominated by the local government to the multi-agent cooperation model with the full participation of market forces. For this purpose, the first task is to define the functions and roles of each policy subject. Local governments should make clear the basis and development direction of returnees’ entrepreneurship in their local area and focus on creating a good policy support and cooperation environment. Banks and other financial institutions should focus on innovating financial service channels, improving service methods, and increasing credit supply to provide continuous and high-quality financial supports for returnees to start their own businesses. Other social organizations should also focus on providing various aspects of support for returnees to start a business, such as collecting relevant entrepreneurial information, organizing entrepreneurial skills training, and researching entrepreneurial culture [66]. Second, all policy subjects should strengthen the overall development, coordinate and promote deep and sustained cooperation, reduce the fragmented situation in promoting entrepreneurship of returnees, and form an extensive linkage cooperation relationship. Finally, under the guidance of national macroscopic policy, policy subjects should draw from their own characteristics and reality to focus on developing a policy implementation model with characteristics of guidance, operability and developmental, instead of mechanistic policy implementation.

3. Optimizing the policy attention distribution pattern according to the local situation and enterprise development stages. After nearly 20 years of exploration and practice, Chinese local governments have gradually improved the construction of entrepreneurial infrastructure, entrepreneurial environment and service system. At present and in the future, the focus of policy attention should shift to the exploration of ways to improve the entrepreneurial quality of returnees, step by step. However, there are still great differences among Chinese local governments in resource endowment, population status, financing environment and the conditions of entrepreneurship. In addition, the problems faced at different development stages are not the same for enterprises started by migrant workers returning to their hometown. There is both generality and individuality in the paths, methods and intensity of policy implementation to improve the quality of entrepreneurship in the hometown, closely related to the local situation and enterprise development stages. For example, newly established enterprises mainly encounter problems of capital shortage and narrow sale channels, while enterprises in the growth stage need to solve the problems of land use, employment, technological innovation and management efficiency. Enterprises started by returnees located in relatively closed areas are mainly faced with the problem of how to open product sales channels, while enterprises located in relatively developed and open areas are mainly concerned with how to achieve brand management. Therefore, specific paths to improve the entrepreneurial quality of returnees require conducting both theoretical and practical exploration, based on local situations and enterprise development stages.