Tourism Development and Rural Land Transfer-Out: Evidence from China Family Panel Studies

Abstract

For a long time, the decline in agricultural comparative returns and the urban–rural development gap in China have prompted the outflow of rural labor. Land transfer policies, which allow farmers to retain their land contracting rights while transferring their management rights, were instituted to mitigate the impact of labor outflow on land use and agricultural production. In recent years, tourism has contributed to the diversification of the rural economy and has had an essential impact on the urban–rural allocation of elements such as labor. In this paper, we adopt a probit model to investigate the impact of tourism development on rural land transfer-out by using data from the China Family Panel Studies. The results show that the marginal effect of tourism development is significantly negative, indicating that the probability of rural land transfer-out was significantly reduced with tourism development. The results are still valid after a series of robustness tests. A mechanism analysis indicates that tourism development inhibits land transfer by enhancing local vitality, such as increasing the local employment of rural labor and promoting participation in agricultural production. Moreover, from the perspective of rural welfare and asset prices, further research finds that tourism development contributes to poverty alleviation and increases land value. These results suggest that tourism development inhibits land transfer while promoting rural sustainable development, helping to understand the impact of tourism on rural land use and household asset allocation from a more comprehensive perspective.

1. Introduction

Since China’s reform and opening up, its rural economy has remained predominantly agricultural. However, the decline in comparative agricultural returns and the widening urban–rural development gap has driven significant rural-to-urban labor migration [1,2], thus leading to land use issues. In China, rural land belongs to village collectives, from which farmers contract farmland for cultivation. Land transfer policies, which allow farmers to retain their land contracting rights while transferring their management rights, were instituted to mitigate the impact of labor outflow on land use and agricultural production [3,4]. Research shows that the off-farm employment of migrant farmers and low returns from agriculture are essential factors that lead to land transfer [1,5], and that household characteristics and institutions also play an essential role [5,6,7]. Industry convergence has become a practical approach to promoting rural revitalization in recent years. Changes in rural industrial structure significantly impact the allocation of factors such as labor and capital [8]. However, the existing research does not include rural industrial factors in the discussion of revitalization, and it especially neglects the rapidly growing tourism.

Tourism has become a key activity in promoting rural economic diversification [9,10]. The development of tourism can promote employment, increase income, and narrow the gap of wealth, and as an “inclusive business”, it has an anti-poverty effect [11,12,13]. Studies have found that tourism, while promoting large-scale land management and guiding the flow of capital to rural areas, also leads to problems such as land use shifts, increased pressure on cultivated land, and ecological impacts, which have a profound and complex influence on rural land use and agricultural production [14,15,16]. However, research directly examining the impact of tourism development on rural land transfer-out is still scant. Although some valuable qualitative discussions exist, quantitative studies remain scarce, and their conclusions are often limited to specific regions. The relationship between tourism and agricultural land transfer is still complex, as the existing research details. On the one hand, local agricultural products and specialty foods are vital to attracting tourists, thus leading to a positive link between tourism and agricultural production [17,18]. In addition, as a labor-intensive industry, tourism can create more local jobs [19], lead to the return of rural labor [20], and promote the local employment of rural labor [11,20]. Given that the land transfer decision is primarily made by migrant workers who can no longer consider agricultural production as a source of income, tourism may inhibit rural land transfer-out by increasing the enthusiasm for agricultural production and promoting local employment. On the other hand, tourism can also destroy agricultural production by competing for land and rural labor [9,21,22], thus promoting rural land transfer-out. Therefore, the impact of tourism development on rural land transfer needs to be further clarified.

To fill this gap, this paper utilizes nationwide survey data from across 25 provinces (cities and autonomous regions) in China to examine the impact of tourism development on rural land transfer-out. This study has made several minor contributions to this area of research: First, the rural industrial system is an important factor driving agricultural modernization and is the key to solving rural problems. However, the existing land transfer research has mainly focused on micro-characteristics and institutions, neglecting the impact of industrial change on land use. This study examines the impact of tourism on land transfer from an industrial perspective, thus extending the existing research on the driving factors of land transfer. Second, this study analyzes the impact mechanism of tourism development on land transfer from the perspectives of local employment and agricultural production, which contribute to a more comprehensive understanding of the complex relationship between tourism and rural land transfer-out. Lastly, this study further examines the welfare effect of tourism from the perspectives of poverty alleviation and rural land value. This not only helps us to explore the role of tourism in improving rural welfare, but also provides useful insights into how to promote sustainable rural development.

2. Research Background and Hypothesis

2.1. Research Background

Rural land in China belongs to village collectives. Villagers contract farmland from the village collective for agricultural production. In 2014, the Chinese government introduced a policy about transferring agricultural land, proposing a separation of the ownership, contracting rights, and management rights of agricultural land. Land contracting rights now belong to the members of the village collectives, while management rights belong to the land users. Rural land transfer-out refers to farmers retaining their land contract rights and transferring their management rights to economic organizations or other farmers. Since China’s reform and opening up, the continuous widening of the gap between urban and rural incomes, consumption, and infrastructure, as well as the decline in agricultural comparative returns, have led to a large number of rural laborers migrating to cities [2,8,23]. The rural population and employment rate in rural areas in China have continuously declined from 2011 to 2020. Coupled with the problem of an aging population, the labor shortage has led to severe land utilization problems in rural China [24]. In 2017, the percentage of fallow cultivated land in China’s main grain-producing areas was 5.85%, with the most severely fallow areas being northeastern Heilongjiang and northwestern Jilin [25]. The government is trying to solve the land utilization and agricultural production problems through land transfer.

Scholars have extensively discussed the drivers of land transfer, with some of them placing off-farm employment at the core of their research. A questionnaire survey conducted on 1032 farmer households in 10 provinces in China showed that labor migration and low agricultural income are the top two factors leading to land transfer [5]. Improving farmers’ off-farm income will reduce their dependence on their land and encourage them to transfer it [26,27]. However, the impact of non-farm employment on land transfer cannot be generalized. For example, the land transfer effect of non-farm employment differs according to the gender of the employee [28], while the distance from employment also plays a crucial role in land transfer decisions [29]. Another strand of research focuses on micro-characteristics such as households, household heads, and communities [7,30,31]. Moreover, institutional arrangements such as land ownership reform and public pension schemes significantly impact land transfer [6,32]. At the same time, tourism’s industrial linkages and relationship with the rural economy have become a hot topic of research. By the end of 2018, China had created 388 national demonstration counties for leisure agriculture and rural tourism. Tourism has profoundly changed the rural industrial system, promoted the diversification of the rural economy, and significantly impacted the allocation of elements such as labor in rural and urban areas. However, research directly discussing the relationship between tourism and rural land transfer-out is still lacking. This paper aims to fill this gap by discussing the impact of tourism development on land transfer.

2.2. Research Hypothesis

Labor outflow is a central reason why farmers transfer land [24]. When middle-aged and young laborers work outside their hometowns for a long time, their families may choose to transfer their farmland due to constraints on their labor force. Contrarily, local labor employment can take care of the area’s agricultural production. Based on a survey of 243 villages in Southwest China, it was found that farmers’ non-farm employment in the local area reduced the scale of land transfer [26]. Compared with those outside the province, farmers who work in their village, outside their village in the local town, and outside their town but in the local county are less willing to transfer land [29,33]. In China, a common social phenomenon is that migrant workers tend to choose to work short distances non-farm jobs while also taking care of agricultural production [23]. According to economic geography, tourism is more attractive to geographically closer residents. Therefore, an increase in non-farm employment in the local area can negatively affect land transfer [29].

As a labor-intensive industry, tourism can create many employment opportunities [19]. According to statistics, the proportion of farmers who are in the leisure agriculture and rural tourism workforce has reached more than 60%. Tourism not only provides opportunities for laborers to work near where they live [11], but also significantly increases the willingness of laborers to return to the area [20]. The job creation effect of tourism is both indirect and direct. In order to accommodate tourism, local areas may need to invest in infrastructure construction, such as roads, transportation, and hotels [34]. The construction and maintenance of infrastructure usually require a large amount of labor, which will indirectly encourage farmers to participate in tourism. In addition, farmers can directly participate in tourism by engaging in hospitality and catering. Research also shows that rural tourism projects can enhance residents’ pride in their rural culture, thereby strengthening their emotional connection to local areas [35]. Therefore, tourism development can help create employment opportunities, enhance the sense of community among farmers, and promote local employment for rural laborers, thus inhibiting land transfer.

Agricultural production is also affected by tourism development. First, in addition to tourism development, a mutually complementary relationship has been formed between agricultural production and non-agricultural employment in rural areas. Due to the localization of the employment of rural labor and the time difference between the tourism and farming seasons, tourism allows farmers to balance non-agricultural activities with agricultural activities, thus flexibly combining farming, working, and tourism employment [36,37]. Second, as tourism expands, the demand for local agricultural products will increase, which helps stimulate residents’ enthusiasm for participating in agricultural production [38]. Some studies have shown that tourists’ interest in specific foods is crucial to their choice of destination [17,39], thus increasing the demand for local food [17,40,41]. By providing catering services or cooperating with tourism enterprises, agricultural products can be directly or indirectly sold to tourists, bringing added value to agricultural and sideline products and thus promoting agricultural production [42,43]. Therefore, tourism development can also promote farmers’ participation in agricultural production, thereby inhibiting land transfer.

In general, tourism has injected new vitality into the rural economy. The promotion of local employment and the enhancement of farmers’ enthusiasm for agricultural production have made farmers more dependent on their land to create value. Despite this, existing discussions have not focused on the impact of tourism on rural land transfer-out, with a particular lack of empirical studies based on national survey data. Based on our previous analysis, the following research hypotheses are proposed:

Hypothesis 1.

Tourism development significantly reduces the probability of rural land transfer-out.

Hypothesis 2.

Tourism development inhibits rural land transfer-out through two mechanisms—the local employment of rural labor and participation in agricultural production.

3. Methodology

3.1. Data Sources

The data used in this paper come from the China Family Panel Studies (CFPS). The CFPS is a national, large-scale, and multidisciplinary social tracking survey project conducted by the Institute of Social Science Survey (ISSS) of Peking University. It adopts a three-stage, unequal, probability clustered sampling design. The survey covers 25 provinces, municipalities, and autonomous regions, including Beijing, Tianjin, Hebei, Shanxi, Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangxi Zhuang Autonomous Region, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, and Gansu, which account for about 95% of China’s total population. Therefore, the CFPS can be considered a well-representative sample of the national population. We use ArcGIS 10.8 to map the sample provinces, as shown in Figure 1. Among its studies, a household questionnaire inquired, in detail, into the households’ land transfer situation, which provides supporting data for this paper to examine the impact of tourism development on land transfer. Since land transfer is usually decided at the household level, the sample in this paper is limited to rural households who have contracted rural land from village collectives. All household and individual data used in this paper are from CFPS 2018 and CFPS 2020, and macroeconomic data are from the EPS database. Since the CFPS questionnaire after 2010 no longer investigates the tourism status of communities, we refer to the existing literature based on the CFPS and obtain community-level data from CFPS 2010 [11], including tourism development and community location. Typically, community locations do not change over the 2010–2020 period. The influence of a long merging span of tourism development will be explained in depth and addressed in the Section 3.2.2 and Section 4.2. Except for the community-level variables, which come from CFPS 2010, the other variables include two years of data from 2018–2020. We merge and match community, household, individual, and macroeconomic data, thus constituting unbalanced panel data of 10,659 observations.

3.2. Definitions of the Variables

3.2.1. Explained Variable

Rural land transfer-out (RLTO) is defined based on the question “In the past 12 months, did your family rent out any of the collectively distributed land?” in the questionnaire. If the answer is yes, then the value of RLTO is 1; otherwise, it is 0.

3.2.2. Explanatory Variable

Drawing on previous research [11], the value of tourism development (TD) is assigned based on the question “Is your village/residential community a tourist destination (including the reception of tourists)” in the questionnaire. If the answer is “yes”, then the value of TD is 1; otherwise, it is 0. The tourism development variable in this paper is derived from CFPS 2010, but the long merging span may lead to estimation bias. This effect is reflected in the exit of communities within scenic spots from tourism and the entry of communities outside of scenic spots into tourism after 2010. Since villages within scenic spots in the sample are mostly located in national 4A or 5A scenic spots, they usually stay in tourism. In addition, there usually may not be a significant increase in the number of communities with a newly established or opened scenic spot in the sample. To mitigate estimation bias as much as possible, samples from provinces with a high proportion of newly established or opened 4A or 5A scenic spots from 2010 to 2020 are excluded in the Section 4.2.

3.2.3. Control Variables

Based on existing research [11,24,33,44], this paper introduces a series of control variables, including household head characteristics, household economic characteristics, and village location characteristics, as shown in

Table 1. Definitions of the Variables.

Variables	Definition	Number	Mean	Standard Deviation
Rural land transfer-out (RLTO)	Whether the household rent out their land (Yes = 1; No = 0)	10,659	0.1982	0.3987
Rural land transfer-out with a fee (RLTF)	Whether the household rent out their land for a fee (Yes = 1; No = 0)	10,659	0.1625	0.3689
Rural land transfer-out income (RLTI)	Income from the rural land transfer-out	10,659	368.2536	1224.8480
Tourism development (TD)	Whether the community is a tourist destination (Yes = 1; No = 0)	10,659	0.0460	0.2094
Household migrant income	The logarithm of the total income from helping other farmers with farm work and working outside	10,659	5.9078	5.0621
Household net assets	Household net assets (ten thousand yuan)	10,659	171.9425	337.5504
Scale of agricultural machinery	The logarithm of the total value of the household’s agricultural machinery	10,659	3.0506	4.0077
Household size	The total number of family members	10,659	3.9508	1.9248
NPEH	The number of people eating at home	10,659	3.4072	1.6941
Land expropriated	Whether the household experienced land expropriation (Yes = 1; No = 0)	10,659	0.0721	0.2586
Household head’s education	Household head’s educational level (illiterate or semi-literate = 1; primary school = 2; junior high school = 3; senior high school = 4; above college = 5)	10,659	2.4057	1.0997
Household head’s marriage status	Household head’s marital status (Married = 1, Unmarried = 0)	10,659	0.8695	0.3369
Household head’s age	The age of household head	10,659	51.4674	13.4070
Household head’s health	Household head’s health status (Very healthy = 5; Healthy = 4; Relatively healthy = 3; Average = 2; Unhealthy =1)	10,659	2.8697	1.2600
Household head’s gender	Household head’s gender (Male = 1; Female = 0)	10,659	0.5878	0.4923
Community location	The travel time from the community to the county seat (h)	10,659	1.1402	1.6091
Per capita GDP	Per capita GDP of each province in that year (ten thousand yuan)	10,659	5.6498	2.1512
Population	Population of each province in that year (millions)	10,659	61.5692	31.3125

. Household head characteristics include age, gender, educational attainment, marital status, and health status. Household economic characteristics include migrant income, net assets, the value of agricultural machinery, household size, and number of people eating at home. Land expropriation is a common phenomenon in China due to tourism development, and one we control for. Moreover, the community location characteristics are also controlled for. Finally, we control for the per capita GDP and population of the provinces.

3.3. Model Setting

3.3.1. Baseline Model

Since the explained variable is a binary dummy variable, this paper adopts a probit model to examine the impact of tourism development on rural land transfer-out

$$\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{b}\left({\mathrm{R}\mathrm{L}\mathrm{T}\mathrm{O}}_{\mathrm{i},\mathrm{t}}\right)=(\mathsf{\alpha }+\mathsf{\beta }{\mathrm{T}\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathsf{\gamma }{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}}_{\mathrm{p}}+{\mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}})$$

(1)

where ${\mathrm{R}\mathrm{L}\mathrm{T}\mathrm{O}}_{\mathrm{i},\mathrm{t}}$ indicates whether the farmer has transferred land out. We also use ${\mathrm{R}\mathrm{L}\mathrm{T}\mathrm{I}}_{\mathrm{i},\mathrm{t}}$ and ${\mathrm{R}\mathrm{L}\mathrm{T}\mathrm{F}}_{\mathrm{i},\mathrm{t}}$ as explained variables in our robustness test. ${\mathrm{T}\mathrm{D}}_{\mathrm{i},\mathrm{t}}$ represents local tourism development. ${\mathrm{X}}_{\mathrm{i},\mathrm{t}}$ is a set of control variables, including household head characteristics, household economic characteristics, community location, and provincial variables. ${\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}}_{\mathrm{p}}$ and ${\mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}}_{\mathrm{t}}$ are provincial fixed effects and annual fixed effects respectively. ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ is a random disturbance term.

3.3.2. Mechanism Test

We use a form of mediation effects for mechanism analysis. First, we examine the impact of tourism development on the mediating variable (MV) using model (2). Then, we examine the impact of the mediating variable on rural land transfer-out using model (3).

$${\mathrm{M}\mathrm{V}}_{\mathrm{i},\mathrm{t}}=\mathsf{\alpha }+\mathsf{\beta }{\mathrm{T}\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathsf{\gamma }{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}}_{\mathrm{p}}+{\mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(2)

$$\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{b}\left({\mathrm{R}\mathrm{L}\mathrm{T}\mathrm{O}}_{\mathrm{i},\mathrm{t}}\right)=(\mathsf{\alpha }+{{\mathsf{\beta }}_1\mathrm{M}\mathrm{V}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{T}\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\mathsf{\gamma }{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{v}\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{e}}_{\mathrm{p}}+{\mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}})$$

(3)

where ${\mathrm{M}\mathrm{V}}_{\mathrm{i},\mathrm{t}}$ represents mediating variables, including the local employment of rural labor and participation in agricultural production. Other variables are as set in model (1).

4. Results

4.1. Baseline Results

We first group the samples according to the tourism development (TD) variables, and then examine the mean difference in values between these two groups. The results are shown in

Table 2. Mean comparison test.

Variables	TD = 0	TD = 1	Mean Difference	T Value
RLTO	0.2013	0.1347	0.0666	3.61 ***
RLTF	0.1652	0.1061	0.0591	3.46 ***
RLTI	374.8722	230.898	143.9742	2.54 ***

Note: Table 2 reports the results of the t-test for the mean differences between groups. *** represents the significance level of 1%. As shown in Table 2, the proportion of households involved in land transfer in scenic areas is relatively lower.

. The results show that the proportion of households involved in land transfer in scenic areas is 13.47%, while in non-scenic areas it is 20.13%. The mean difference between the two groups is 6.66%, which is statistically significant at the 1% level. Moreover, the proportion of households in scenic areas that transferred land for a fee and the land transfer income of farmers in scenic areas are both significantly lower than those in non-scenic areas. These findings suggest that tourism development significantly inhibits land transfer, thus supporting Hypothesis 1.

Table 3. Tourism development and rural land transfer-out.

Variables	(1)	(2)	(3)
Tourism development	−0.0664 ***	−0.0635 ***	−0.0757 ***
	(0.0231)	(0.0228)	(0.0227)
Household migrant income	0.0022 ***	0.0034 ***	0.0032 ***
	(0.0008)	(0.0008)	(0.0008)
Household net assets	0.0001 ***	0.0001 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)
Scale of agricultural machinery	−0.0152 ***	−0.0142 ***	−0.0140 ***
	(0.0011)	(0.0011)	(0.0011)
Household size	−0.0115 ***	−0.0102 ***	−0.0093 ***
	(0.0034)	(0.0034)	(0.0033)
NPEH	−0.0041	0.0015	0.0008
	(0.0036)	(0.0036)	(0.0036)
Land expropriated	−0.0286 *	−0.0299 *	−0.0331 **
	(0.0156)	(0.0155)	(0.0156)
Household head’s education		0.0196 ***	0.0176 ***
		(0.0042)	(0.0042)
Household head’s marriage status		−0.0597 ***	−0.0608 ***
		(0.0117)	(0.0116)
Household head’s age		0.0020 ***	0.0018 ***
		(0.0004)	(0.0004)
Household head’s health		−0.0148 ***	−0.0149 ***
		(0.0032)	(0.0032)
Household head’s gender		−0.0171 **	−0.0134
		(0.0083)	(0.0083)
Community location			−0.0346 ***
			(0.0056)
Per capita GDP			−0.0123
			(0.0123)
Population			−0.0052
			(0.0077)
Province FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
Observations	10,659	10,659	10,659
Pseudo R2	0.0706	0.0805	0.0858

Note: Table 3 presents the regression results for model (1). In column (1), the model includes only household economic characteristics. Column (2) expands upon this baseline by incorporating household head characteristics, building upon the model in column (1). Finally, column (3) presents the full model, with all control variables included. As is evident in Table 3, the results demonstrate an inhibitory effect of tourism development on rural land transfer. Robust standard errors, clustered at the family level, are shown in parentheses. Significance levels are denoted by *, **, and *** for 10%, 5%, and 1%, respectively.

reports the regression results of model (1). Column (1) only includes household economic characteristics, while column (2) further introduces household head characteristics and is based on column (1). Column (3) includes all control variables. As shown in column (3), the average marginal effect of TD is −7.57%, which is statistically significant at the 1% level. The results indicate that tourism development significantly reduces the probability of land transfer.

The results for the control variables are as expected. The coefficient of household net assets is significantly positive, indicating that households with better economic conditions are more likely to transfer land. Increasing migrant income reduces the households’ dependence on rural land and promotes RLTO. The coefficient of the household head’s education level is significantly positive, indicating that households with higher education levels are more likely to transfer land. The household head’s health status coefficient is negative, indicating that a better health status inhibits land transfer. In other words, poor health de-incentivizes agricultural production and tends to lead to the transfer of rural land. The coefficient of the scale of agricultural machinery is significantly negative, indicating that households with a larger agricultural machinery input are not likely to rent out their rural land.

4.2. Robustness Tests

We then conduct a series of robustness tests: Firstly, the results show that TD inhibits RLTO significantly. However, it may lead to estimation bias due to potential changes in the TD variable. Since communities in scenic spots are primarily located in national 4A and 5A scenic spots [11], these samples will generally not exit tourism between 2010 and 2020. However, communities outside of scenic spots may enter tourism between 2010 and 2020, thus causing estimation bias. Therefore, we need to eliminate the sample outside of scenic spots likely to enter tourism. In order to identify the net effect of TD as much as possible, we choose to exclude the samples located in the provinces with a high proportion of newly established or opened 4A or 5A scenic spots from 2010 to 2020. For this reason, we manually collect the years when 4A or 5A scenic spots were established or opened. Statistics show that about 70% of the scenic spots were established or opened before 2010. However, 30% of scenic spots were established or opened between 2010 and 2020, and nearly 90% of them are located in 15 provinces, such as Shandong. Then, we exclude the samples from these 15 provinces and end up with a sample of 2804. Among these samples, the TD variable is less likely to change. The regression results are shown in column (1) of

Table 4. Robustness tests.

Variables	(1)	(2)	(3)	(4)	(5)
	Sample Exclusion	Logit Model	RLTF	RLTI	Cultivated Land
Tourism development	−0.0859 **	−0.0795 ***	−0.0798 ***	−175.1049 ***	−0.0763 ***
	(0.0434)	(0.0238)	(0.0208)	(48.7993)	(0.0232)
Household migrant income	0.0025	0.0031 ***	0.0035 ***	10.9605 ***	0.0033 ***
	(0.0017)	(0.0008)	(0.0008)	(2.5950)	(0.0008)
Household net assets	0.0001 ***	0.0001 ***	0.0001 ***	0.3858 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0657)	(0.0000)
Scale of agricultural machinery	−0.0153 ***	−0.0143 ***	−0.0109 ***	−23.5043 ***	−0.0142 ***
	(0.0023)	(0.0011)	(0.0010)	(3.2254)	(0.0011)
Household size	−0.0044	−0.0091 ***	−0.0053 *	−16.8976	−0.0095 ***
	(0.0079)	(0.0034)	(0.0031)	(10.4615)	(0.0034)
NPEH	−0.0111	0.0005	0.0013	15.1410	0.0006
	(0.0081)	(0.0036)	(0.0033)	(10.7578)	(0.0037)
Land expropriated	−0.1014 ***	−0.0367 **	−0.0285 *	−80.4468 *	−0.0307 *
	(0.0358)	(0.0162)	(0.0146)	(42.0009)	(0.0158)
Household head’s education	0.0192 **	0.0179 ***	0.0158 ***	42.0514 ***	0.0186 ***
	(0.0090)	(0.0042)	(0.0039)	(14.1637)	(0.0043)
Household head’s marriage status	−0.0840 ***	−0.0585 ***	−0.0333 ***	−78.0374 *	−0.0634 ***
	(0.0238)	(0.0114)	(0.0110)	(42.9698)	(0.0118)
Household head’s age	0.0024 ***	0.0018 ***	0.0016 ***	6.4409 ***	0.0018 ***
	(0.0008)	(0.0004)	(0.0003)	(1.1394)	(0.0004)
Household head’s health	−0.0107 *	−0.0149 ***	−0.0125 ***	−14.1615	−0.0150 ***
	(0.0063)	(0.0032)	(0.0029)	(10.3095)	(0.0032)
Household head’s gender	−0.0353 **	−0.0141 *	−0.0037	−11.3157	−0.0134
	(0.0160)	(0.0083)	(0.0076)	(26.7620)	(0.0084)
Community location	−0.0545 ***	−0.0349 ***	−0.0369 ***	−24.6219 ***	−0.0341 ***
	(0.0118)	(0.0058)	(0.0055)	(4.8003)	(0.0057)
Per capita GDP	−0.0798 ***	−0.0123	−0.0107	−71.6741	−0.0124
	(0.0233)	(0.0122)	(0.0110)	(45.2470)	(0.0125)
Population	0.0942 ***	−0.0044	−0.0032	−52.8301 **	−0.0063
	(0.0332)	(0.0076)	(0.0069)	(26.5297)	(0.0079)
Province FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
Observations	2804	10,659	10,659	10,659	10,388
Pseudo R2/Adj. R2	0.1269	0.0859	0.0879	0.0701	0.0862

Note: Table 4 reports the estimation results of the robustness test. In the first robustness test, we exclude the samples from Hebei, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Guangdong, Guangxi, Sichuan, Guizhou, Shaanxi, Gansu, and Hunan. The robust standard errors, clustered at the household level, are in parentheses. *, **, and *** represent significance levels of 10%, 5%, and 1%, respectively. The probit model reports the average marginal effects of all the explanatory variables.

, and the coefficient of TD is still significantly negative at the 5% level. Secondly, we use the logit model to create an estimation. Model selection can have a substantial impact on the estimated results. By using the logit model for estimation, this paper aims to ensure that the results are not overly sensitive to the choice of the regression model. The logit model can also be used for regression analysis with a binary dependent variable. However, it differs from the probit model in that the logit model assumes that the residuals follow a logistic distribution function. In contrast, the probit model assumes that they follow a normal distribution function. As shown in column (2) of Table 4, the results of this paper still hold. Thirdly, we use land transfer for a fee as the dependent variable to test the impact of TD on rural land transfer-out. A 3.57% presence of samples transferring agricultural land free of charge may affect this causal identification. To address this, RLTF (rural land transfer-out for a fee) is used as the dependent variable. The results are shown in column (3) of Table 4, and the coefficient of TD is also significantly negative at the 1% level. Fourthly, in order to avoid limiting this study due to the dependent variable being a dummy variable, we choose rural land transfer income as the dependent variable for regression. As shown in column (4) of Table 4, the coefficient of TD is still negative. Finally, to further ensure the robustness of our results, the sample is limited to households that contract cultivated land. As shown in column (5) of Table 4, the coefficient of TD is still negative at the 1% level. After the robustness tests, the conclusions remain robust.

4.3. Mechanism Analysis

The results above show that tourism development is essential to inhibiting rural land transfer-out. This paper will next discuss how this works.

4.3.1. Local Employment of Rural Labor

An outflow of rural labor is essential for land transfer [5,33]. The existing research suggests that the non-agricultural employment of farmers in the local area negatively affects the scale of land transfer [29]. Tourism can create local employment opportunities and entice farmers to return [20,45]. Tourism can also strengthen residents’ emotional connection with the countryside and reduce farmers’ willingness to move to cities [35,46]. Research shows that tourism significantly promotes the local employment of farmers [11]. Therefore, the local employment of rural labor is a potential mechanism by which tourism development inhibits land transfer. We test this mechanism using a mediation effect model. The results showed that tourism development significantly promoted the local employment of rural labor, which in turn negatively affected RLTO, indicating that it is an essential mechanism by which tourism development can inhibit land transfer.

4.3.2. Agricultural Production Participation

The departure of farmers from agricultural production is a direct cause of land transfer. Tourism is closely related to the local demand for agricultural and sideline products [17,41], thus increasing farmers’ land-based agricultural income and making them more dependent on their land. By creating food demand and promoting local employment [11,41], tourism can promote agricultural production, thus inhibiting land transfer. We test the mechanism of agricultural production using a mediation effect. The results are shown in

Table 5. Mechanism Analysis.

Variables	Local Employment of Rural Labor		Participation in Agricultural Production
	(1)	(2)	(3)	(4)
Tourism development	0.1328 ***	−0.0699 ***	0.0449 **	−0.0598 ***
	(0.0495)	(0.0227)	(0.0222)	(0.0220)
Local employment of rural labor		−0.0377 ***
		(0.0044)
Participation in agricultural production				−0.1958 ***
				(0.0079)
Household migrant income	0.0055 ***	0.0035 ***	−0.0012	0.0031 ***
	(0.0019)	(0.0008)	(0.0009)	(0.0008)
Household net assets	0.0001 **	0.0001 ***	−0.0000 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Scale of agricultural machinery	0.0314 ***	−0.0128 ***	0.0311 ***	−0.0067 ***
	(0.0025)	(0.0011)	(0.0011)	(0.0011)
Household size	0.1717 ***	−0.0031	0.0423 ***	−0.0000
	(0.0092)	(0.0033)	(0.0039)	(0.0031)
NPEH	0.0782 ***	0.0036	−0.0049	−0.0002
	(0.0095)	(0.0036)	(0.0041)	(0.0034)
Land expropriated	0.0582 *	−0.0312 **	−0.0299 *	−0.0406 ***
	(0.0350)	(0.0155)	(0.0153)	(0.0154)
Household head’s education	0.0047	0.0178 ***	−0.0373 ***	0.0087 **
	(0.0100)	(0.0042)	(0.0044)	(0.0041)
Household head’s marriage status	0.3892 ***	−0.0458 ***	0.1037 ***	−0.0334 ***
	(0.0258)	(0.0117)	(0.0125)	(0.0113)
Household head’s age	−0.0012	0.0017 ***	0.0026 ***	0.0023 ***
	(0.0009)	(0.0004)	(0.0004)	(0.0003)
Household head’s health	0.0309 ***	−0.0137 ***	0.0057 *	−0.0132 ***
	(0.0074)	(0.0031)	(0.0034)	(0.0030)
Household head’s gender	0.0964 ***	−0.0086	0.0265 ***	−0.0064
	(0.0195)	(0.0082)	(0.0089)	(0.0079)
Community location	−0.0065	−0.0356 ***	0.0066 *	−0.0298 ***
	(0.0080)	(0.0056)	(0.0034)	(0.0054)
Per capita GDP	−0.1031 ***	−0.0153	−0.0043	−0.0133
	(0.0296)	(0.0122)	(0.0139)	(0.0121)
Population	−0.0922 ***	−0.0085	0.0001	−0.0051
	(0.0200)	(0.0077)	(0.0084)	(0.0076)
Province FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Observations	10,659	10,659	10,659	10,659
Pseudo R2/Adj. R2	0.2865	0.0938	0.1703	0.1425

Note: Table 5 reports the results of our mechanism analysis. Columns (1) and (2) test the mechanism of the local employment of rural labor (measured by the number of employed laborers living at home). Column (1) uses OLS regression to find that the effect of TD on the local employment of rural labor is positive, and column (2) shows that the local employment of rural labor is negatively correlated with land transfer. Columns (3) and (4) test the mechanism of participation in agricultural production, which is measured by whether the household is involved in agriculture (yes = 1, no = 0). In column (3), the probit model demonstrates that the effect of TD on farmers’ participation in agricultural production is positive, and column (4) shows that the effect of farmers’ participation in agricultural production on RLTO is negative. These regression results show that tourism development inhibits RLTO by promoting the local employment of rural labor and participation in agricultural production. The robust standard errors, clustered at the household level, are in parentheses. *, **, and *** represent significance levels of 10%, 5%, and 1%, respectively. The probit model reports the average marginal effects of the explanatory variables. The OLS model reports Adj. R², and the probit model reports Pseudo R².

. The results show that the estimated effect of tourism development on agricultural production is positive, indicating that it has increased the probability of farmers participating in agricultural production. The estimated effect of agricultural production on land transfer is negative, indicating that tourism development has inhibited land transfer by promoting agricultural production. These results indicate that tourism development inhibits land transfer by promoting farmers’ participation in agricultural production.

4.4. Tourism Development and Rural Welfare

The results above show that, by promoting local employment and agricultural production, tourism inhibits land transfer aimed at improving rural welfare. However, it is not clear how tourism affects rural welfare. Next, we will directly examine the welfare impact of tourism from the perspectives of poverty alleviation and rural land values.

4.4.1. Poverty Alleviation

Tourism is a labor-intensive industry with low employment thresholds and significant inclusiveness. Even people with weak employment competitiveness can undertake tourism service work after simple training. In 1999, the UK Department for International Development (DFID) introduced the term “pro-poor tourism” [47,48]. Based on rural areas’ unique resources, and supported by tourism products with market appeal, pro-poor tourism aims to benefit low-income residents. By integrating local communities and small farmers into the supply chain, tourism can help people escape poverty [47,49,50]. We identify households from the data with a per capita income below the poverty line as poor and assign them a value of 1; otherwise, we assign them a value of 0. As shown in

Table 6. Further research: poverty alleviation and rural land value.

Variables	(1)	(2)
	Poverty Alleviation	Rural Land Value
Tourism development	−0.0367 **	0.7275 *
	(0.0170)	(0.3723)
Household migrant income	−0.0226 ***	−0.0921 ***
	(0.0006)	(0.0131)
Household net assets	−0.0003 ***	0.0035 ***
	(0.0001)	(0.0004)
Scale of agricultural machinery	−0.0011	0.3558 ***
	(0.0008)	(0.0180)
Household size	0.0024	0.2507 ***
	(0.0027)	(0.0450)
NPEH	0.0171 ***	−0.0150
	(0.0030)	(0.0490)
Land expropriated	−0.0005	−0.4444 **
	(0.0126)	(0.1882)
Household head’s education	−0.0248 ***	−0.0562
	(0.0034)	(0.0579)
Household head’s marriage status	−0.0295 ***	0.6740 ***
	(0.0091)	(0.1411)
Household head’s age	0.0014 ***	−0.0076
	(0.0003)	(0.0047)
Household head’s health	−0.0040	0.1407 ***
	(0.0025)	(0.0457)
Household head’s gender	0.0000	0.5103 ***
	(0.0066)	(0.1127)
Community location	0.0040 ***	−0.0099
	(0.0015)	(0.0324)
Per capita GDP	−0.0024	−0.4058 **
	(0.0148)	(0.1993)
Population	0.0169 *	−0.4438 ***
	(0.0091)	(0.1355)
Province FE	Yes	Yes
Year FE	Yes	Yes
Observations	10482	10,659
Pseudo R2/Adj. R2	0.2290	0.1700

Note: Table 6 reports the impact of tourism development on household poverty and rural land value. Since the dependent variable, household poverty, is a dummy variable, this paper adopts the probit model to examine it. As shown in column (1) of Table 6, the effect of TD on household poverty is significantly negative, indicating that tourism can significantly alleviate household poverty. We adopt an OLS model to examine the relationship between tourism development and rural land value. As shown in column (2) of Table 6, the effect of TD on land value is significantly positive, indicating that tourism can improve land value. The robust standard errors, clustered at the household level, are in parentheses. *, **, and *** represent significance levels of 10%, 5%, and 1%, respectively. The probit model reports the average marginal effects of the explanatory variables.

, the coefficient of TD is significantly negative, indicating that tourism can help rural households escape from poverty.

4.4.2. Rural Land Value

The findings above indicate that the tourism industry has significantly impacted rural economies. By promoting the involvement of rural households in agricultural production and creating more local employment opportunities, the tourism industry has reduced the transfer of rural land and rural poverty levels. Through influencing the supply and demand of land and the local demand for agricultural products, the impact of the tourism industry on land value may be pronounced. However, there is still limited discussion in the existing literature on how the tourism industry specifically influences rural land value. In this context, this study further examines the impact of the tourism industry on rural land value (ten thousand yuan). Using land value as the dependent variable, the regression results are as shown in column (2) of Table 6. The results show that the coefficient of TD is significantly positive at the 10% level, suggesting that the tourism industry significantly promotes an increase in land value.

5. Discussion

5.1. Theoretical Contributions

First, previous studies have mainly focused on institutional factors and farmer characteristics while paying less attention to industrial development. We find that tourism development has a negative impact on rural land transfer-out, which is significant at the 1% level. This finding provides a new perspective for land transfer research and could expand the research on the economic impact of tourism development.

Second, this study analyzes how tourism affects land transfer from the perspectives of labor migration and agricultural production. Previous studies have argued that tourism could lead to land transfer due to tourism competing with agriculture for land and labor [22,29,41,45], or that it could help inhibit land transfer by creating local employment and stimulating agricultural production [29,41,45]. This study finds that tourism development promotes the local employment of rural labor, encouraging workers to stay and participate in agriculture, thus inhibiting land transfer. This finding helps clarify the complex relationship between tourism and land use.

Third, this study examines the impact of tourism development on welfare from the perspective of poverty alleviation and land value. Our results show that tourism can help lift farmers out of poverty and enhance their land value; it has potential to improve their welfare. Therefore, although tourism does not promote RLTO, it can help to improve the overall interests of farmers and rural communities with respect to local employment, agricultural production, and poverty alleviation.

Overall, the findings of this study enrich our understanding of the impact of tourism on land use, rural livelihoods, and regional development. They provide valuable insights for future research and policy making.

5.2. Practical Contributions

Firstly, tourism can promote local employment and poverty alleviation. As a labor-intensive industry, tourism can provide more employment opportunities locally and encourage more residents to remain in the countryside, endorsing the sustainable development of rural communities. Moreover, it can also help rural households out of poverty, making it an “inclusive” industry that promotes social equity. Therefore, it is recommended that local governments should develop tourism projects featuring local characteristics to attract tourists, taking into account local conditions and relying on local resources such as nature, culture, and history. At the same time, it is recommended that governments should strengthen their tourism infrastructure to provide a favorable environment for the development of tourism.

Secondly, tourism can facilitate agricultural production. Tourism provides more opportunities for rural labor. Not only can tourism directly create jobs, but it can also help farmers improve their income through increased demand and value-added agricultural products. Farmers can meet the needs of tourists by innovating and producing traditional agricultural products, thereby boosting their value. Governments and businesses can work together to provide training and support for farmers, assisting in enhancing their agricultural production techniques and the production of traditional agricultural items that meet tourist needs. In addition, it is suggested that the government should encourage farmers to develop leisure agriculture, rural tourism, and other industries by introducing policies to broaden their sources of income.

Lastly, studies have shown that tourism can promote sustainable rural development. Tourism development can further improve economic development in rural areas, alleviate the development gap between urban and rural areas, and enhance the sustainability of the rural economy. Therefore, it is suggested that the government should actively promote the upgrading of regional industries based on local natural and cultural resources while also promoting the integrated development of primary, secondary, and tertiary industries. Furthermore, the government should strengthen the training of human resources in rural areas to provide strong support for tourism development.

6. Conclusions

Using nationwide survey data from China, we examined the impact of tourism development on rural land transfer-out and found a significant inhibition, with a marginal effect of −7.57%, indicating that tourism is an essential factor affecting rural land transfer-out. This conclusion remained valid after adopting a series of robust tests. This study further discussed the mechanisms by which tourism development affects rural land transfer-out. This revealed that tourism development inhibits land transfer by promoting the local employment of rural labor. In addition, we found that tourism development can increase the probability of rural households participating in agricultural production, which in turn inhibits land transfer. Moreover, tourism also contributes to poverty alleviation and increases land value.