Transfer or Surrender Rural Homestead Land? An Investigation of farmers’ Preferences in China

AbstractWith rapid industrialization and urbanization in China, a large rural population has migrated to cities and left their rural homestead land (RHL) idled. Land use policy on RHL utilization is crucial to the efficient use of land resources. However, due to farmers’ limited knowledge of regulations and lack of participation in rural land governance, their preference is sometimes overlooked in the policy-making process. The purpose of this paper is to examine farmers’ preferences for the two primary RHL utilization programs, i.e., transfer and surrender. We conducted a farmer household survey of 405 households in rural areas in Changchun at Jilin Province, China in 2018 and estimated farmers’ willingness to participate in the mechanisms using a bivariate ordered probit model. Various options under the two mechanisms are also investigated. We find a positive correlation between rural households’ willingness to participate in these two mechanisms. This is the first comprehensive study taking into consideration the correlation and tradeoff in farmers’ choice between the two mechanisms. The results shed light on the heterogeneity of households’ needs and interests in RHL utilization, an important component to be considered in rural development policy-making decisions.Keywords: Rural homestead landfarmers’ preferenceChina Disclosure statementThe authors report there are no competing interests to declare.Notes1 Village collective is a notion traced to the socialist era of agricultural collectivization associated with the Maoist period. Communist rule under Mao redefined the Chinese village as a socialist collective, both an administrative and economic unit. Under collectivization, rural resources were collectively owned, meaning that rural resources were owned not by individual peasants or households, but by “a community body that is seen as the legitimate entity representing the interests of all its constituents” (Kan, Citation2016; Yep, Citation2015).2 The bivariate ordered probit model, even for only two dimensions, does not easily extend to ordered logit model framework due to computation complications (Greene & Hensher, Citation2010). While for multinomial choices, the probit model has limited use due to the need to evaluate multiple integrals of the normal distribution. Instead, the logit model has been used widely in various fields (Greene, Citation2012).3 We also collected 19 observations at a labor market in Changchun and place them in the Other category.4 The survey implementation and data collection were led by the second author whose university does not have an institutional review board for ethics approval of research involving human subjects. The university does not require ethics approval of research. However, to ensure ethical practices, we took certain steps. First, at the beginning of the survey questionnaire, we clearly stated the purpose of the survey and the estimated time needed to complete it. We described that the survey is anonymous, participation is voluntary, and data confidentiality will be protected. Further, all survey enumerators followed a standardized script when approaching farmers and seeking their consent to participate in the survey. Prior to starting the survey, participants were asked for their oral consent.5 The sample size was determined based on statistical modeling power and research budget constraints.6 According to China’s 7th census in 2020, in Changchun, male accounts for 50.05% and female accounts for 49.94%. People of 0-14 years old, 15-59 years old, and 60 years old and above account for 12.14%, 67.01%, and 20.85%. People who are 65 years or older account for 14.15%. 1.01% of the population have no education, 20.42% have elementary school, 35.75% have middle school, 16.05% have high school, and 22.22% have college education. The average years of education is 10.69 years (Changchun News, Citation2021). Since we drew our sample from rural areas of Changchun, the demographics of respondents in our sample show a higher male ratio, an older family member composition, a lower education level, and a larger household size, which is consistent with the rural vs. urban demographics difference.7 We also tested the correlation between farmers’ willingness to transfer RHL and to surrender RHL by estimating a system of two equations. Specifically, we included farmers’ willingness to surrender RHL as an explanatory variable in the willingness to transfer RHL equation, included farmers’ willingness to transfer as an explanatory variable in the willingness to surrender equation, and estimated the simultaneous equation system using a three-stage least square method. Both the coefficient on surrender in the transfer equation and the coefficient on transfer in the surrender equation are positive and significant at the 1% level, indicating a positive correlation between the two willingness to participate variables.8 Since the sample is composed of seven districts/areas and the housing locations may affect farmers’ decisions on their willingness to transfer or surrender RHL, we conducted a robustness check by adding the location dummy variables to the model. We found that the coefficients of the location dummy variables are insignificant and adding the dummy variables doesn’t change the coefficients of variables of interest much.9 Since the marginal effect calculation in bivariate ordered probit model is complex, we calculate the marginal effects based on univariate ordered probit model, aiming to provide an approximate estimate on the magnitude of the impacts.Additional informationFundingThe study was supported by Jilin Province Philosophy and Social Science Planning Fund Office, China (Program Number 2018BS20).