1. Introduction
Rural entrepreneurship has injected new vitality into rural revitalization and, as the main force to comprehensively promote rural revitalization, promoting entrepreneurship of new agricultural management subjects and accelerating the participation of new agricultural management subjects in industrial integration, can effectively promote the modernization of agriculture and rural areas; it is of great significance to the construction of a strong agricultural country. Since the implementation of the strategy of “mass entrepreneurship and innovation” in September 2018, China’s rural entrepreneurial environment has been continuously optimized and the entrepreneurial activity of farmers has been continuously improved. The dual-creation strategy provides an entrepreneurial environment, scientific and technological innovation, market opportunities and policy support for new agricultural management subjects, and effectively promotes the upgrading of the structure of the agricultural industry and the quality of farmers. It effectively promotes the upgrading of the agricultural industry structure and the quality of farmers; provides strong support for the achievement of agricultural modernization, rural revitalization, and urban–rural integration; and promotes the sustainable development of the national economy. New agricultural management subjects (hereinafter referred to as “new agricultural subjects”) are agricultural economic organizations based on the family management system, with a large scale of operation; they are compatible with modern agriculture and the market economy, and their economic main bodies include family farms, agricultural cooperatives, agricultural enterprises, etc. [
1]. New agricultural subjects have played an important role in improving the efficiency of agricultural production, optimizing the structure of the agricultural industry, and improving the quality of the agricultural industry. In the context of the full implementation of the rural revitalization strategy, the new agricultural subjects of the organization and management level and the ability to collaborate in the development of the new agricultural subjects significantly improved and strongly support the development and growth of the agricultural industry chain links; the new agricultural subjects have become a new driving force for agricultural and rural development and provide a necessary way to lead the high-quality development of agriculture and rural areas, to achieve rural revitalization and the construction of modern agriculture [
2]. Since the concept of new agricultural subjects was put forward in 2014, the new agricultural subjects have developed rapidly and their number has grown rapidly; in China’s case, as of 2022, family farms alone have reached 3.9 million
①, farmers’ cooperatives have reached 2.227 million
②, and there are more than 90,000 agricultural industrialized leading enterprises above the county level
③. The areas of their coverage have also been extended to the whole industrial chain.
As an important part of agricultural entrepreneurship, analyzing the influencing factors affecting the level of agricultural entrepreneurship based on previous research can play a good role in improving the level of entrepreneurship of new agricultural subjects. In the process of agricultural development, agricultural production environment [
3], national policy land [
4], financial support [
5], and other factors can play an important role and agricultural entrepreneurship are an important force to promote the development of agriculture. There have been scholars in the field of agricultural entrepreneurship that have carried out a large number of studies. Bielby [
6] found that an enhanced interaction between farmer entrepreneurs and other stakeholders can help to enhance their entrepreneurial ability and management level, while Bouichou [
7] found that agribusiness financing constraints can have an important impact on agricultural entrepreneurship willingness. Meutia [
8] pointed out that different countries, according to their own economic structure and development stage, adopt different financing strategies to support agricultural entrepreneurship. Soleymani [
9], on the other hand, constructed rural entrepreneurship indicators based on the Delphi method. In addition, some scholars found that the new generation of digital technologies [
10], rural network broadband [
10], Internet use [
11], and other emerging technologies can facilitate agricultural entrepreneurship. Coupled with the unique form of development of new agricultural subjects, the entrepreneurship of new agricultural subjects faces problems such as low survival rate, low entrepreneurial willingness, lack of entrepreneurial knowledge, and high entrepreneurial risk.
The pilot policy for innovative cities (hereinafter referred to as the “innovation policy”) is an important decision made by the Chinese government to improve the comprehensive competitiveness of cities and to build an innovative country, with the impetus of institutional innovation; it is also a key strategy to enhance the capability of independent innovation. As an important initiative to support China’s innovative development strategy, the implementation of the innovation policy has undergone a series of pilots and has been continuously expanded. Up to now, innovation pilot cities cover 78 cities (districts) in 31 provinces across the country, which is a centralized embodiment of China’s innovation-driven development strategy. With the gradual implementation of innovation pilot cities, many scholars have evaluated their policy effects, and current studies have mainly explored the innovation effect of innovation policies [
12] and the green development effect [
13,
14]. In addition, existing studies have shown that innovation policies can significantly increase the level of urban entrepreneurial activities and the policy effects are more obvious in cities with higher administrative levels, geographic location advantages, and non-productive service industries [
15]. Unfortunately, however, existing studies have not paid enough attention to whether innovation policies have an impact on agricultural entrepreneurship. If there is an impact, what is its transmission mechanism? Based on this, this paper takes innovation policy as a quasi-natural experiment, based on a multi-period double-difference model, to study the impact of innovation policy on the level of entrepreneurship of new agricultural subjects, as well as a heterogeneity analysis, with a view to providing theoretical references and practical references for the implementation and formulation of innovation policy, as well as the promotion of the entrepreneurship level of new agricultural subjects and, in this way, showing the impact of innovation policy on rural entrepreneurship.
The possible marginal contributions of this paper mainly include the following: (1) Constructing a quasi-natural experiment with innovation policy and evaluating and demonstrating the impact of innovation policy on the entrepreneurial level of new agricultural subjects and the spillover effect, after a series of robustness tests. (2) From the perspectives of city size, science and education level, and different types of new agricultural subjects, we carefully analyze the heterogeneous impact of innovation policy on the entrepreneurial level of new agricultural subjects under different city sizes, different levels of science and technology investment, and different types of new agricultural subjects. (3) Taking the level of scientific and technological input and the level of credit support and the level of scientific and technological progress as the transmission mechanism, we explore in depth the intrinsic mechanism of innovation policy affecting new agricultural subjects, enriching the research literature on the impact of entrepreneurial activities in policy evaluation. It provides strong theoretical support for the implementation of innovation policies and how to improve the level of entrepreneurship of new agricultural subjects and, at the same time, provides inspiration for relevant government departments to formulate effective entrepreneurship policies.
3. Study Design and Data Description
3.1. Research Methods and Analytical Tools
Referring to the approach of the article by Peráček [
24], this paper points out that it is necessary to choose suitable scientific methods for the research and that these will lead us to the expected results of the research, as well as the fact that the choice of methods is determined by the main content of the research. The purpose of the research in this paper is to explore the impact of innovation policy on the level of entrepreneurship of new agricultural subjects and its transmission path, as well as the heterogeneous differences that exist between different city sizes, levels of science and education, and the level of entrepreneurship of new agricultural subjects, in order to reflect the role of innovation policy on rural entrepreneurship, in line with the characteristics of the method of empirical analysis and the method of generalization. Empirical analysis is a research method that recognizes objective phenomena and provides people with real, useful, certain, and precise knowledge, which is used to obtain experience through observation and then summarize the experience into theory, usually adopting the inductive method, focusing on the experience close to the reality, and focusing on the problem of “what” the phenomenon itself is. It attempts to transcend and exclude value judgment, revealing only the intrinsic constituents of the objective phenomena and the universal connection of the factors, summarizing the essence of the phenomena and the operating rule, and its main purpose is to explain the relationship between various independent variables and a dependent variable, which can be used to validate the existing theories or to summarize the new theories from observation. Therefore, this paper adopts empirical analysis and inductive methods to carry out scientific research.
As for the empirical analysis, the double-difference method, as a policy effect assessment method recognized by a wide range of scholars, has been widely used in recent years. The principle of the use of this method is to regard the implementation of a certain policy as a natural experiment and to examine the net effect of the policy implementation on the object of analysis, by adding a control group of those unaffected by the policy into the sample and comparing the analysis with the sample points that were originally affected by the policy to form an experimental group. This analytical method consists of benchmark regression, balanced trend test, and placebo test components, which are in line with the scenario of the impact of innovation policies on the level of entrepreneurship of new agricultural subjects of this research question. Therefore, this study uses the multi-period double-difference method as the specific research method of empirical evidence. The double-difference method needs to satisfy the following three hypothetical premises: (1) parallel trend hypothesis—the trend of the outcome effect of the control group and the experimental group is the same before the policy is implemented; (2) the individual treatment stability hypothesis—the policy intervention affects the experimental group only and does not have an interaction effect on the control group; and (3) the linear conditional hypothesis—the potential outcome variable satisfies a linear relationship with the treatment and time variables, implying that each unit of change in the treatment variable has a fixed effect on the outcome variable. Based on the data composition and empirical methods used in this paper, the commonly used econometric statistical software is selected for calculations in this paper.
3.2. Modeling
This paper analyzes the spillover effects of innovation policies on entrepreneurship of new agricultural subjects through a multi-period DID approach. In 2008, Shenzhen officially became a pilot city for innovation policies, while the second, third, fourth, and fifth batches of pilot cities were approved in 2010, 2011, 2012, 2013, and 2018.This paper constructs a quasi-naturalistic-based experiment. The final experimental group includes 71 innovative pilot cities and the control group includes the other 213 cities.
Innovation policies are gradually promoted in batches, while the traditional DID method is only applicable to assess a single policy point in time. For this reason, this paper draws on the work of Autor [
25] and Yuan [
26], to construct a multi-period DID model, with pilot cities assigned a value of 1 and non-pilot cities assigned a value of 0. The policy implementation time dummy variable (treat_policy) is set, which is 0 before the implementation time of the policy in the pilot cities, and is set to 1 for the year of implementation and subsequent years. The multi-period DID model is constructed as follows:
where
enterpit is the explanatory variable number of new types of subjects,
treat_policyit represents the innovation policy, and its coefficient reflects the policy effect of the innovation policy;
controlit represents the control variables; year
k and
μind represent the time dummy variables and individual city fixed effects; and
ℇit represents the random error term. The model effectively controls the characteristic differences and trends in time change between pilot and non-pilot cities.
3.3. Variable Setting and Data Description
Explained variable. Regional entrepreneurship level (
enterp). The level of regional entrepreneurship is generally examined in terms of the number of self-employed persons and the number of start-ups. The World Bank defines entrepreneurial activity as the behavior of individuals or groups participating in formal economic sector activities in the form of a legal business, so the newly registered limited liability companies are used as a measure of the level of entrepreneurial activity [
27]. In order to better accomplish data collection and comparison, this paper uses the number of new types of new agricultural subjects added as an explanatory variable.
Core explanatory variable. Innovative city pilot policy as a dummy variable, 0 for pilot cities before the policy is implemented, and 1 for the current year and subsequent years.
Control variables. Drawing on the studies of Xu [
28] and Li [
29], this paper controls for the following factors affecting the level of entrepreneurship of the new main body: The level of regional economic development (
pgdp), measured as the real GDP per capita of the prefecture and city; population density (
logurl), measured as the ratio of the total population of the region to the area of the region, with logarithmic treatment; number of people employed in the primary industry (
logalf), measured as the population of the labor force engaged in the production of the primary industry, with logarithmic treatment; human capital (
logahcl), measured as the number of students enrolled in the general institutions of higher education, with logarithmic treatment; the level of financial development (
fin), measured as the prefecture’s and city’s ratio of financial institutions’ loan balance to GDP at the end of the year; digital inclusive finance (
dif), digital Inclusive Finance Index; agricultural mechanization level (
aml), measured as the ratio of total power of agricultural machinery to the area of arable land; and digital rural construction (
drc), measured as the number of mobile telephones owned by the average rural resident per 100 households at the end of the year.
Mediating variables. Science and technology investment (logfse), measured by the number of government investments in science and technology; credit support (logcse), measured by the balance of loans from financial institutions at the end of the year in the prefecture and municipalities; and scientific and technological progress (loganypag), measured by the number of patent applications by new types of subjects.
The data in this paper come from the
China Urban Statistical Yearbook,
China Agricultural Statistical Yearbook, and the ZJU Carter-Enterprise Research China Agricultural Research Database (CCAD). In view of the differences in dimension and order of magnitude of the indicators in the evaluation index system, it is necessary to logarithmically process some of the data in order to eliminate the influence of heteroskedasticity on results. The definitions of the various variables are shown in
Table 1.
3.4. Descriptive Statistics of Variables
The panel data in this paper contain data including 284 prefecture-level cities from 2005 to 2020, of which there are 71 pilot cities and 213 non-pilot cities; the descriptive statistics are shown in
Table 2.
5. Conclusions and Recommendations
As the strategy of “mass innovation and entrepreneurship” continues to deepen, the role of farmers’ entrepreneurship and agricultural entrepreneurship in the rural revitalization and the integrated development of the agricultural industry is increasing and the Chinese government has introduced a series of policies to promote innovation and entrepreneurship in agricultural and rural areas. Based on the panel data of 284 prefecture-level cities in China from 2005 to 2020, this paper constructs a multi-period double-difference model by treating the pilot policy of innovative cities as a “quasi-natural experiment”. Using various statistical methods such as linear regression, parallel trend test, placebo test, and PSM-DID, we empirically investigate the impact and mechanism of the innovative city pilot policy on the entrepreneurial level of new farmers and verify the proposed theoretical hypotheses. It also analyzes the heterogeneous differences generated by the innovation policy on the entrepreneurship level of new agricultural subjects of different city sizes, different levels of science and education, and different types, in order to reflect the impact of the pilot policy of innovative cities on rural entrepreneurship. Specific conclusions and recommendations are as follows:
5.1. Conclusions
The findings of this paper show that innovation policies significantly increase the level of entrepreneurship in new agricultural subjects and this conclusion still holds true after a series of robustness tests using a propensity score matching method, placebo test, and replacement of explanatory variables; Hypothesis 1, proposed above, is verified. While Po-Chi [
38] showed that technological innovation can have a significant impact on agricultural productivity growth, Carolan [
39] found that the use of digital platforms and technologies can help agriculture manage resources more efficiently and improve agricultural productivity; Mann [
40] pointed out the effectiveness of the U.S. SBIR policy in guiding the innovation activities of both rural and urban firms; Raissa [
41] found that perfecting mobile agricultural advisory services and using innovative technologies can provide small farmers with more market information and agricultural expansion suggestions, thus optimizing the efficiency of agricultural production. The conclusions of the above scholars echo the findings and models of this paper. Innovation has the characteristics of knowledge spillover and technology diffusion; active urban innovation activities promote technological upgrading and, thus, this technological progress spreads to the countryside, promoting the development of rural entrepreneurial activities. While innovation centers in developing countries tend to be located in research institutes in cities, relying on the incentives and support of urban innovation policies, innovation agents in developed countries are more widely distributed, but the phenomenon of knowledge diffusion is still widespread, although the direction of knowledge flows may not be as uniform as in developing countries. Therefore, the model proposed in this paper is more applicable when used in developing countries and, if applied to developed countries, the regional distribution of the location of technology centers may have to be considered and it also has some value for policy making and agricultural entrepreneurship practices in developed countries.
At the same time, the research in this paper can also conclude that the mechanism test shows that there are three main paths of the innovation policy on the entrepreneurship of new agricultural subjects. First, through increasing scientific and technological inputs to effectively enhance the efficiency of innovation and the rate of transformation of scientific and technological achievements, to create a more suitable entrepreneurial environment; second, through the enhancement of the effect of credit support, to alleviate the pressure on the entrepreneurial capital of farmers, which, in turn, enhances the entrepreneurial activity; and third, through the promotion of scientific and technological progress to promote the upgrading of agricultural technology, to bring more opportunities for entrepreneurship in agriculture, to attract more entrepreneurial talents to join, and, thus, to improve the entrepreneurship level of the new agricultural subjects in the pilot region. Hypotheses 2–4 proposed above are verified through the analysis. Heterogeneity test found that innovation policy significantly promotes the entrepreneurial level of new agricultural subjects in cities with a high level of science and education, as well as small and medium-sized cities, and does not play a significant role in the entrepreneurial level of new agricultural subjects in cities with a low level of science and education, as well as large cities. In addition, innovation policy has a significant effect on increasing the entrepreneurship level of agricultural cooperatives and agricultural enterprises, and the policy effect is stronger for agricultural enterprises, while there is no significant effect on family farms.
5.2. Recommendations
Based on the findings of this paper, the following policy implications are drawn:
First, increase investment in agricultural science and technology and actively create an agricultural entrepreneurship service platform to promote the entrepreneurship of new agricultural subjects. Innovation policy as a complex systematic project, in order to better play its role in promoting the entrepreneurial level of New agricultural subjects. On the one hand, cities should further increase investment in agricultural science and technology; promote agricultural science and technology research and development, to ensure that agricultural science and technology innovations continue to produce output; and promote the transformation of innovation results on the ground. On the other hand, the relevant departments should actively create agricultural entrepreneurship service platforms to provide farmers with scientific and technological, information, capital, and other support and enhance the willingness of farmers to start their own business. In addition, it is necessary to further improve the infrastructure and related public services, tilt the administrative services towards agricultural entrepreneurship, formulate a targeted and differentiated support system, create a favorable entrepreneurial environment and development space, and promote the creation of new agricultural business entities. At the same time, governments at all levels need to strengthen interaction in policy formulation, implementation, and optimization; summarize experiences; and gradually expand the scope of implementation in the original pilot cities as the center, so as to drive the development of the surrounding areas and maximize the effectiveness of the policy.
Second, the implementation of innovation policies should be tailored to local conditions and scientifically planned, so as to make policy implementation more flexible and inclusive. For cities with a high level of science and technology, as well as large cities, the urban entrepreneurial environment should be further optimized and the policy dividends brought about by innovation policies should be utilized continuously to promote urban agricultural entrepreneurship, while the effects of the policies should be radiated to the neighboring cities, so as to achieve high-quality development. As for small and medium-sized cities and cities with a low level of science and education, they should give full play to their “latecomer’s advantage”, tap entrepreneurial potential through innovative policies, promote the concentration of urban innovation factors, and facilitate the emergence of new opportunities and technologies, thereby increasing the entrepreneurial vitality of new agricultural subjects.
Third, explore the multidimensional path of innovative policies to promote the entrepreneurial level of new agricultural subjects and optimize the effect of pilot policy implementation. First of all, for different types of new agricultural subjects to take different measures to help, such as agricultural cooperatives and agricultural enterprises, can be encouraged to integrate agriculture and emerging technology, vigorously guiding the application of information technology such as big data, 5G technology, Internet of Things, cloud computing, and other information technology, combined with the development of the agricultural industry and agricultural dual-creation, to help family farms to solve the existing lack of scientific and technological support, and to improve the adaptability of science and technology and agricultural production. Secondly, new agricultural subjects in different regions should be adapted to local conditions and suitable development paths should be selected to enhance the effect of policy implementation.
5.3. Research Limitations
There are still shortcomings in the research process of this paper. Firstly, there may be multiple policies affecting the level of entrepreneurship of new agricultural subjects implemented simultaneously in various regions and, although this paper takes into account the impact of policies such as entrepreneurial cities, smart cities, and low-carbon city pilot policies, it still cannot completely exclude the competing explanations of other policies; it is worthwhile to further explore how to more accurately identify the impact of innovative policies on the entrepreneurship of new agricultural subjects. Second, due to the limitation of the completeness of the data on the registration of new agricultural subjects, this study fails to fully assess the individual variability of the impact of innovation policies on new agricultural subjects and, with the increasing richness and improvement of data resources, future studies can explore this issue in greater depth. Third, this study mainly explores the impact of innovation policies on new agricultural subjects’ entrepreneurship in prefecture-level cities. In the future, if more refined data can be obtained, the specific impact of innovation policies on new subjects’ entrepreneurship at the county and township levels can be explored. It better captures the impact of innovative city pilot policies on rural entrepreneurship.