**2. Theoretical Analysis and Research Hypotheses**

Theoretically, the income level of rural households and the level of social security entitlement are the two fundamental factors that influence rural households' consumption [1]. Farmland, as the largest livelihood, influences the general social structural characteristics of most rural households in China [21]. Since the reform and opening up, with the rapid progress of urbanization, industrialization and agricultural modernization, farmers have broken free from the shackles of farmland to gradually enter the cities and towns for nonagricultural employment. The small-scale, loose and fragmented farmland management model is no longer able to meet the needs of rural economic development [34]. For this reason, the Chinese government has been making efforts in top-level design, formulating and issuing a series of relevant policy documents to strengthen rural households' residual claims to farmland, relax the management rights of farmland and guarantee the realization of farmland transfer benefits for rural households. Undoubtedly, the transfer of farmland can promote the optimal allocation of land resources and moderate scale operation of agriculture, allow the rational and adequate allocation of rural factors of production such as land, labor, technology and capital, effectively promote the development of the rural economy and the improvement of the income level of rural households and completely activate the productive and property functions of farmland [9–12]. However, it must be acknowledged that the social security system in China's rural areas is not yet sound, and the transfer of farmland can indeed change the fate of the rural households concerned to a certain extent. In addition, farmland is increasingly becoming a basic survival guarantee for vulnerable groups of rural households who lack the ability to move to urban areas [35–37]. Of course, the security function of farmland is the most important and basic function of farmland for rural households. Whether it is the "further" property function of farmland or the "step back" productive function of farmland, when rural households realize the functional differentiation of farmland, not only does the security function of farmland not become lost [1] but also it increasingly strengthens the property and productive functions of farmland [38,39].

At present, the transfer of farmland is gradually becoming a new way for rural households to accumulate original capital. Based on the perspective of rural household livelihood of farmland transfer-out, the income effect of farmland property function can be divided into direct effect and indirect effect [40]. Among them, the direct effect is the rental income brought by the lease or transfer of farmland management rights to rural households of farmland transfer-out, and the income from the direct effect is agricultural income. However, at this time, the dependence of rural households on the agricultural income brought by farmland transfer-out is weak. While the indirect effect refers to the wage income obtained by farmland transfer-out to promote the transfer of surplus rural labor to the non-agricultural sector for employment. The rental income of farmland is an important part of transfer income, which has the characteristics of temporary income and rural households will be more casual in spending [3]. Wage income has a permanent character, and rural households prefer to use this income as a recurrent consumption expenditure [1]. It is obvious that the income structure of rural households is enriched and diversified by the property function of farmland. In addition, the theory of "psychological accounts" suggests that rural households can allocate different incomes to different accounts, which cannot be filled by each other, and that rural households have different consumption tendencies for different sources of income [41]. The enrichment of rural households' income structure is essentially the division of their holistic income into numerous units, which will greatly strengthen the perception of subjective wealth increase [1]. Therefore, the change in income structure brought about by the transfer-out of farmland can stimulate the consumption of rural households with both rental income and wage income [42]. At the same time, as a component reflecting the hierarchy of rural households' needs and the order of their satisfaction, food consumption is a demand dominated by rural households' physiological requirements, while non-food consumption is a pursuit of rural households' convenience and performance needs and personal enjoyment and development needs [43]. As the transfer-out of farmland gives rural households a richer income structure and brings them a higher level of subjective income, they will gradually reduce food consumption to satisfy their physiological needs and increase non-food consumption of goods and services for convenience and performance needs as well as personal enjoyment and development needs [44]. Thus, the transfer-out of farmland can lead to an increase in the non-food consumption capacity of rural households, which in turn helps to optimize the consumption structure of rural households.

The perspective of the rural household livelihood strategy is based on farmland transfer-in, rural households mainly focus on agricultural production, and farmland has become one of the most important means of production for them. At this time, the dependence of rural households on agricultural income is very strong, and they will do everything possible to expand the scale of farmland to increase agricultural productive income. Therefore, farmland transfer promotes the rational and optimal allocation of land resources, allowing ordinary rural households to acquire relatively concentrated farmland, which is beneficial to a certain extent to the development of the agricultural production of rural households of farmland transfer-in in the direction of moderate scale, intensification, specialization or marketization and rural households realize income growth in the realization of productive functions of farmland and continuously improve their consumption capacity [40]. However, along with the basic completion of China's farmland titling and certification work, on the one hand, farmland titling has properly solved the problems of inaccurate area of contracted land of rural households parcels and unclear four directions, and the "public domain" of farmland property rights has been priced into the market, and rural households of farmland transfer-in have lost the organizational space to earn the "public domain" of property rights [9–12]. On the other hand, the stable property rights of farmland encourage the impersonalization and high rent of farmland transfer among acquaintances [45], which enhances the bargaining position and bargaining power of farmland transfer-out transactions of rural households. As a result, farmland transfer-in of rural households based on the productive use of farmland is faced with the dilemma of increasing production expenditure due to the expansion of production scale, while the income from single-structure agricultural production increases. The theory of "loss aversion" suggests that rural households feel more strongly when weighing losses than gains [1]. In addition, farming is a tough occupation, and rural households value hard-earned income [3]. In view of this, rural households are reluctant to increase their non-food consumption of goods and services for convenience and performance needs as well as personal enjoyment and development needs [44], resulting in a slowdown or even a decrease in the growth rate of non-food consumption expenditure [46]. Instead, rural households tend to increase their spending on food consumption to satisfy physiological needs [44], which is not conducive to optimizing the consumption structure of rural households.

Based on the above theoretical analysis, this paper constructs an analytical framework for the impact of farmland transfer on rural households' consumption (see Figure 2). Meanwhile, the following research hypothesis 1, hypothesis 2 and hypothesis 3 are proposed.

**Figure 2.** Analytical framework of the impact of farmland transfer on rural household consumption.

**H1:** *The transfer-out of farmland from rural households stimulates their consumption and makes them more willing to increase their non-food consumption expenditures, and then contributes to the optimization of their consumption structure.*

**H2:** *The transfer-in of farmland from rural households stimulates their consumption and makes them more willing to increase their food consumption expenditure, and then does not contribute to the optimization of their consumption structure.*

**H3:** *The transfer of farmland (farmland transfer-out or farmland transfer-in) from rural households affects their income under different livelihoods, and then affects their consumption and consumption structure.*

#### **3. Research Design**


Yunnan Province, located in southwest China, is an important part of the Yunnan-Kweichow Plateau and a relatively underdeveloped area in China. Compared with Guizhou Province, where 62% of the land area is karst landform, Yunnan Province has its comparative advantage in landform, which makes the farmland in Yunnan Province more valuable for circulation. In addition, although Yunnan Province is not the main grain-producing area in China, it is of great practical significance to study the impact of farmland transfer behavior on rural household consumption in plateau mountainous and underdeveloped areas. Based on the above explanations, we selected Yunnan Province as the final research point of this paper.

#### 3.1.2. Data Sources

To explore the impact of farmland transfer on rural households' consumption, in November 2021, teachers, doctoral students and master's students of related majors from the School of Economics of Guizhou University and the School of Economics of Yunnan University, formed a relevant subject group to conduct rural household surveys in 16 prefecture-level cities or autonomous prefectures in Yunnan Province. In order to reduce sampling bias, the research team used a stratified random sampling method, stratified according to the administrative vertical relationship of the city (state)—county (city, district)—township (town)—village in turn. One county (city or district) was randomly selected in each prefecture-level city or autonomous prefecture, two townships or towns were randomly selected in each county (city or district), one to two villages were randomly selected in each township or town, and 10 to 15 questionnaires were randomly distributed to rural households in each village under investigation. In addition, this paper takes 2020 as a unit time cycle and a key time node of this survey, so as to facilitate the interview of relevant issues and data collection and collation by the members of the research group. Finally, in this survey, 650 questionnaires were distributed in 50 villages, and 600 questionnaires were recovered with a recovery rate of 92.31%. In addition, out of the 600 questionnaires collected, the questionnaires with obvious errors, repeated relevant content and inconsistent with the research theme of this paper were discarded. Finally, 537 valid questionnaires were obtained, involving 50 villages, with an effective rate of 82.62%.

#### *3.2. Variable Settings*

#### 3.2.1. Explained Variables

The China National Bureau of Statistics categorizes the consumption of rural residents in China into eight major types, including food, clothing, housing, household equipment and supplies, transportation and communication, education and entertainment, health care and other consumption. Besides this, rural household and family are both organizational concepts. Unless otherwise specified, the number of rural households and family members in this paper is consistent. We use the 2020 per capita household consumption expenditure (logarithmicized) to represent the total consumption expenditure of rural households according to China's national statistical caliber and drawing on the research practices of Geng et al. [1], Chen et al. [19] and Yang et al. [21]. However, unlike Cai et al. [47] who covered per capita household consumption expenditure divided into three types of expenditure: subsistence consumption, developmental consumption and productive consumption to measure the consumption expenditure and consumption structure of rural households, this paper uses this feature of the Engel coefficient to reflect changes in consumption structure to dichotomize per capita household consumption expenditure in 2020 into two types of expenditure: per capita rural household food consumption and per capita rural household non-food consumption. As we all know, the Engel coefficient refers to the proportion of food consumption expenditure in the total consumption expenditure of the family. It was put forward by Engel, a German statistician in the 19th century, on the change of consumption structure based on empirical statistical data. Generally speaking, the smaller the Engel coefficient, the better the consumption structure of the family. In other words, the more the family spends on non-food consumption, the better the consumption structure of the family will be.
