Aging population, farm succession, and farmland usage: Evidence from rural China☆

Contents lists available atScienceDirect

Land Use Policy

journal homepage:www.elsevier.com/locate/landusepol

Aging population, farm succession, and farmland usage: Evidence from rural China

^☆

Baoling Zou

^a

, Ashok K. Mishra

^b,⁎

, Biliang Luo

^a

aNational School of Agricultural Institution and Development, College of Economics and Management, South China Agricultural University, Guangzhou, 510642, China

bMorrison School of Agribusiness, W. P. Carey School of Business, Arizona State University, 7271 E Sonoran Arroyo Mall, Mesa, AZ 85212 USA

A R T I C L E I N F O

Keywords:

Aging farming population Farm succession Farmland usage Multinomial logit China Leasing out Farmland

A B S T R A C T

The aging farming population has a significant influence on production agriculture, succession planning, suc- cessors, and farmland usage. Given recent trends in urban migration and increased opportunities for off-farm work, aging farmers increasingly face problems with farmland succession and usage in China. This study in- vestigates the usage of farmland in the absence of a farm successor. Using multinomial logit regression model and data from rural households, wefind that aging farmers without successors tend to have options for farmland usage. Specifically, the presence of a grain subsidy increases the likelihood of keeping the farmland in agri- culture—albeit by hiring labor or leasing out farmland and decreases the likelihood of pooling farmland into farming cooperatives as shareholders. Off-farm work decreases the likelihood of using hired labor and leasing out farmland. Rich rural households are less likely to keep farmland idle. Large farm operators are more likely to lease out farmland. Finally, rural people with pension plans are more likely to pool their land in land co- operatives—a less-risky option.

1. Introduction

The issue of an aging society has become a highly prevalent social problem not only in developed countries (Fichtner, 2018) but in de- veloping countries (Masters, 2013). The most significant impact of an aging population, however, occurs in production agriculture. Produc- tion agriculture is back-breaking work that relies on labor, and farm operators tend to be older than average worker in the overall work- force. Older workers, on average, tend to be less productive than younger workers. Moreover, due to small land holdings, large family sizes, and higher off-farm wages, young and middle-aged individuals prefer to participate in off-farm labor activities (Rizwan et al., 2017).

Older farm operators want to retire from farming, and retirement and succession decisions in family farms are interrelated (Kimhi and Lopez, 1999). As a result, older farm operators face a farm succession problem.

In many cases, older farm operators end up with no successors and then have to decide on the future of their farmland—whether to rent it out, hire labor, or participate in a profit-sharing venture by pooling their

farmland with others.¹

The prosperity and sustainability of agriculture depends on an ef- fective succession of farmland, according to several studies in the lit- erature (Bertoni and Cavicchioli, 2016;Joosse and Grubbström, 2017;

Mishra et al., 2010). However, the children of older farmersethe new generation of potential farmerseare more likely to participate in the off-farm labor market because off-farm income is higher and stable and offers a strategy for exiting agriculture (Corsi and Salvioni, 2017). In addition, the lack of a successor may lead to declines in productivity, financial performance, and innovation in farming (Harris et al., 2012;

Sottomayor et al., 2011). Therefore, current old-age operators without successors face the dilemma of farmland usage.

China, a developing country with the world’s largest population, has about 35% of its labor force in agriculture, compared to 2.5% in the United States. According to the third National Agricultural Census, China had about 314 million agricultural operators in 2016. The age distribution of the agricultural operators is interesting. For example, the population under age 35 was about 60 million and accounted for just

https://doi.org/10.1016/j.landusepol.2018.06.001

Received 17 April 2018; Received in revised form 26 May 2018; Accepted 1 June 2018

☆The authors would like to thank the anonymous reviewers for their helpful and constructive comments that greatly contributed to improving thefinal version of the paper. They would also like to thank the Editor for support during the review process.

⁎Corresponding author.

E-mail addresses:bzou2@asu.edu,Zoubl90@163.com(B. Zou),Ashok.K.Mishra@Asu.edu(A.K. Mishra),luobl@scau.edu.cn(B. Luo).

1In China, land shareholding cooperatives system tend to pool farmland (on share basis) in villages and cultivate the land. Profits are then shared with shareholders.Chen (2016)notes that“the purpose of Land Shareholding Cooperative System is to achieve economies of scale for agricultural production through reconcentration of village land.”Under the LSCS farmers give up operational land use rights and is entrusted to the cooperative. The cooperative can entrust the land use right to the broker (operator) through the bidding. The broker will perform the production and management of the cooperative's land and ensure that the farmers can share the satisfactory dividends (cash income).

Available online 15 June 2018

0264-8377/ © 2018 Elsevier Ltd. All rights reserved.

T

19% of the total agricultural operators; the population ages 36–54 was about 149 million and accounted for about 47% of total agricultural operators; and the population age 55 and above was about 106 million and accounted for 34% of total agricultural operators (NBS, 2016). Like the U.S., China suffers from an aging farming population. Since 93% of China’s total agricultural operators are involved in crop production, farmland succession is a serious social problem. Note that in China the state technically owns and controls farmland, but farm families are allocated land that they operate and pass on to the next generation.² However, farm families can lease their farmland to others but cannot sell it. Land reforms under the Deng Xiaoping era have allowed in- dividuals to lease land from villages.

Readers also should note two salient features of the Chinese farm- land structure and ownership. First, under the China’s Household Contract Responsibility System and households have only three rights:

to transfer the rights of operation of the farmland, only rural house- holds, as members of the collective village, have the original con- tractual and operational rights (Wang and Zhang, 2017). Although farmland leasing is emerging in China, farming households can only transfer the rights of operation of farmland, return the contractual rights of the farmland to the collective village, or pass the farmland to their family members. Second, China’s dual-track structure of socio- economic development requires households in mainland China to reg- ister as either a Rural Hukou or an Urban Hukou. This law also restricts household migration (and by extension, labor migration) between Hu- kous. Migration from rural areas to urban areas is possible. Because of Hukou law requirement, however, the opposite is not realistic because only members of a Rural Hukou can receive rights to operate farmland in a village.³ Considering these two features and the rising share of aging farmers, farm operators are concerned about their farmland usage decisions. The issue is paramount in cases where farm operators lack successors.

Therefore, this study’s objective is to investigate the factors af- fecting farmland usage decisions in the absence of successors in China.

We use farm-level data from nine Chinese provinces (see Fig. 1) and multinomial logit modeling technique to study this objective. The survey was conducted in 2015. This study’s contributions are unique in several ways. First, the study addresses a timely issue faced by a country with the largest number of farmers, with smallholdings, and with a large share of farm operators ready to retire. Second, the issue of farmland usage in the absence of a successor is interesting in a country where a farmland market⁴does not exit and where economic and po- litical systems (defined Hukous per household) discourage the move- ment of urban people to rural areas. Finally, this is thefirst study, to best of our knowledge, to investigate farmland usage by farm operators in the absence of successors.

The rest of this paper is organized as follows. The next section discusses the background of farmland, property rights of farmland, and farmland usage in China. Section3presents the survey data and de- scriptive statistics. The empirical framework and estimation procedure are developed in Section4. Section5analyzes and discusses the em- pirical findings. The last section concludes the study and provides policy implications.

2. Background

Land is a key factor of production and the most important asset of livelihood. This is particularly true in China, where resources are dis- tributed unequally between urban and rural areas. Farmland in China takes the role of social security, providing livelihood security, old-age security, and employment security (Yao, 2009). According to China’s Contracting of Rural Land Law (Wang et al., 2015)⁵, rural households are designated as contractors and have the right to the use the land for growing crops, to sell those crops and to make a profit. However, as contractors they are obligated to maintain the farmland and keep it for production agriculture. To our knowledge, farmland has granted live- lihood opportunities to the villagers, but such opportunities are influ- enced by the availability of non-farm employment opportunities (Shirai et al., 2017). Furthermore, based on China’s Measures for the Admin- istration of Circulation of Rural Land,⁶the contractor (the land’s op- erator) also can subcontract the operational rights of the farmland.

Changes in operational rights become important for old-age rural farming households. In this case, old-age operators havefive options for farmland usage. These include designating a successor, leaving farm- land idle⁷, hiring labor to cultivate the farmland, leasing out farmland for rental income, and joining a land shareholding cooperative. In the last option, rural farm families can pool farmland into a cooperative, essentially becoming shareholders in the cooperative, which in return shares the profits with the shareholders.

In China, farmland is an important symbol of membership in a village. Rural households are able to trace their families’history for several generations (Lobley and Baker, 2012). Rural households also dominate production agriculture, and traditionally, farm succession within the family is preferred (Baker et al., 2016). Recent studies reveal that farms with designated successors may have higher farm invest- ments (Calus et al., 2008) and increased borrowing capacity to invest in farming (Stiglbauer and Weiss, 2000). Conversely, farms without suc- cessors tend to take up other activities and diversification schemes (Potter and Lobley, 1992), disinvest or enter a static management mode (Inwood and Sharp, 2012).

In China, with its booming non-farm economy, migration to urban areas, and limited family members, aging farm operators face a decision on farmland usage. This problem is severe when aging farmers have no successors to take over the farming business. In the absence of succes- sors, many rural Chinese households leave their farmland idle (Lin and Wang, 2014). In a study, (Sottomayor et al., 2011) noted that the choice to leave farmland idle is directly related to the absence of a successor.

This is consistent with the fact that aging farmers have deeper feelings than their children have for their farmland (Zhong and Luo, 2013).

Other factors that could encourage aging Chinese farmers without successors to leave farmland idle include the high transaction cost of leasing farmland and the absence of a mature and perfect farmland market, which leaves farmland price undefined (Wu, 2017).

Technological innovation and off-farm work opportunities have increased labor migration from rural to urban areas. The emergence of off‐farm employment has a significant and positive impact on stimu- lating rural households to lease out their farmland (Che, 2016). For example,Kung (2002), using data from 824 rural households concluded that active participation in off-farm labor markets (measured by number of days worked) resulted in reduced amount of land rented.

2China has about 120 million hectares of farmland—making the size of the average plot per rural family less than 0.5 hectares.

3Rural Hukou grants rural families the rights to operate the allocated, by village col- lective action, piece of farmland. However, the opposite is not true because the total land in the village has already been divided among the families living in the village. New residents will not be able to get to work the land for a living because they belong to urban Hukou.

4Recall that unlike in many market economies, farmland in China cannot be bought or sold in the market (Ma, et al., 2015).

5Law on the Contracting of Rural Land of the People’s Republic of China (promulgated by the Standing Committee of the National People’s Congress, August 29,2002, effective March 1,2003), material source available at: http://www.lawinfochina.com/Display.

aspx?lib=law&Cgid=41762.

6Measures for the Administration of Circulation of Rural Land (promulgated by the Ministry of Agriculture, January 19, 2005, effective March 1, 2005), http://www.la- winfochina.com/display.aspx?lib=law&id=3933&CGid=.

7Though it is illegal to leaving the land idle based on the Law on the Contracting of Rural Land, much of the land still become vacant because the lack of effective supervision in China.

Research Center,Che (2016)found that more members participating in either migration or local off-farm work are more/less likely to lease out/in land. In addition, literature shows that other factors may impact aging farmers’decision to hiring farm labor, including seasonal demand (Benjamin, 1992), farm attributes, off-farm income, family composition (Bedemo et al., 2013), and size of farmland (Kimhi, 2009). Land leasing, or leasing out, is an effective solution for addressing the shortage of on-farm labor and for increasing old-age security for rural Chinese households. For example, the survey data collected from el- derly farmers in China’s Hubei Province showed that about 47% were willing to lease out their farmland to make profits (Nie, 2018). How- ever, others in the same survey did not believe in leasing out their farmland because of the loss of autonomy in production and of a rural way of life. This was especially true for old-age operators for whom the farmland was their only old-agefinancial security. External factors such as land lease policies (Li et al., 2017), land tenure security, and ethni- city (Min et al., 2017) also affect rural households’decisions on the usage of their farmland.

The Chinese government’s program of direct subsidy for grain pro- duction, which is related to historical grain production or contracted land areas in order to improve grain producers’income (Yi et al., 2015), also affects rural households’decisions on their farmland usage.⁸Since farmland is a valuable asset for old-age security, our survey also queried rural households about whether they have pension plans. The notion here is that if the rural households have pension payments, they are more likely to lease out their farmland than rural households without such plans (Nie, 2018). To control for the heterogeneity in economics among households, as the rural households’wealth usually is used in research (Krueger et al., 2016).⁹

Rural households may consider available options for farmland usage and relevant information (Groeneveld et al., 2017). Hence, attributes of their farmland, such as the land area, soil fertility, and irrigation con- ditions, are important variables in rural households’choice of farmland usage (Felson, 2017). Village development and village traffic conditions also may impact farmland values and rental rates. In addition, rural households’experience with land leasing and the Land Shareholding Cooperative System (LSCS) may affect their decisions on farmland usage. We argue that experience with these two options may affect farm household’s decision on farmland usage in old age. The LSCS has the advantages of large-scale modern farming, sustainable farmland usage, and increased income to farmers (Ren et al., 2017).¹⁰The communal institutions and rules governing the access to farmland are a way of sharing risk and allowing a minimum level of subsistence (Promsopha, 2018). Chinese farmers, and aging farmers in particular, may choose to join the LSCS (Lin and Gu, 2017) and receive profits as shareholders in the LSCS. Joining the LSCS is also a strategy for managing risk in the presence of incomplete and asymmetric information in the farmland leasing market. However, the above literature fails to shed light on factors that affect the decisions of aging Chinese farm households without successors on farmland usage. The literature also falls short in recognizing that the real estate market in China is unique in that farmland is not readily bought and sold. If the aging farmer dies without a successor and without having decided what to do with the land the land will revert back to the village and would be reallocated among existing village farm families.

3. Data and descriptive statistics

We use the survey data from rural Chinese households on rural land Fig. 1.Provinces of China and surveyed provinces.

8Because there are differences in the way the grain subsidy program is structured, we queried operators“to whom has the government offered the direct subsidy for grain production.”

9However, due to lack of data in this study we include four dummy variables that capture different income classes based on family income in 2014—a year before the

(footnote continued)

survey data was collected (2015).

10Note that under LSCS there is no change in the collective ownership of the farmland.

and related factors of the markets. The survey was conducted at the beginning of 2015. The selection of research sites for this survey was based on clustering analysis of six indicators (total population, per ca- pita GDP, total area of farmland, share of farmland, agricultural po- pulation, and share of agricultural output). Finally 54 counties within nine provinces and regions (Fig. 1) were selected. The survey collected information from 2704 rural households (Table 1). However, since the study’s objective is to assess the factors affecting farmland usage by aging farm operators without successors, we include only farm opera- tors 51 or older. As a result, we have 878 observations for the present study. Using a structured questionnaire, the survey collected informa- tion on farm and farmer characteristics, cropping patterns, economics of cultivation, participation in a pension system, grain subsidy, soil fertility, and the irrigation condition of farmland. The survey queried farm operators on their choice of farmland usage. Specifically, farmers were asked,“If you were too old to farm and there is not a successor to inherit your farmland, what would you prefer to farm?”The choices were:

(1) leave farmland idle; (2) hire farm labor; (3) lease out farmland; (4) pool the land in the Land Shareholding Cooperative System (LSCS) for investment; and (5) Unknown (meaning they had not yet decided what to do with their farmland).Table 2describes the distribution of the ages of the heads of household and what their farmland usage decision might be in the absence. The table also shows that about one-third of the rural Chinese households are over age 50. Among the rest, households are most willing to lease out their farmland. Pooling of farmland as an investment becomes the second choice, followed by leaving farmland idle and hiring farm labor.

Table 3reports the definition and summary statistics of the variable used on our analysis. Note that about 20% of the sample, farm operators without successors do not know what they will do with their farmland in the future. Our variable of interest, the choice of farmland usage in the absence of a successor, is quite interesting.Table 3reveals that 44%

of respondents want to lease out the farmland, followed by 22% who do not know what to do with the farmland. Additionally, 14% of re- spondents want to leave the farmland idle, and 12% want to pool the farmland into the LSCS as shareholders and receive profits for their

share of farmland. Only 7% want to hire labor to engage in production agriculture. About 62% of the family workers have either part-time off- farm jobs (24%) or full-time off-farm jobs (37%). As a result, off-farm income as a share of total household income is about 64%, suggesting a strong reliance on off-farm income. However, this share is significantly lower than the share of off-farm income in the United States, where Mishra and Sandretto (2002)found that off-farm income is about 90%

of total household income for small farms. On the issue of grain sub- sidies,Table 3shows that 45% of landowners reported receiving a grain subsidy directly from the government, and 40% of the operators (those leasing land from owners) reported receiving a grain subsidy from the government.

The average age of farmers in our sample is about 60, 68% percent are male operators, and the majority (53%) have a primary school education. The average farm operated is about 1.67mu(muis unit of land size in China, 1mu= 0.165 acres). About 46% of the operators had off-farm work experience, and 50% of the farm families reported having access to pensions. Regarding family income,Table 3reports that 36% of the farm families reported income of 10,000–29,000 Yuan, followed by 24% with income of 30,000–50,000 Yuan, and 23% with income of more than 50,000 Yuan. Land market participation ethe leasing in or leasing out of farmlandeare two ways in which retiring or old farm operators can increase their farms’scale of operation and se- cure additional income in old age.Table 3shows that 22% of the rural farm families reported leasing out farmland, perhaps for rental income, and 11% reported experience in leasing in farmland.

Finally, the bottom half ofTable 3shows some land and village attributes we considered in this study. For example, 53% of the farm families reported operating farmland with mediocre soil fertility, and only 1% reported operating farmland with good or excellent soil ferti- lity. The majority of the rural farm families (74%) reported farmland with mediocre to poor irrigation facility, but 26% reported farmland with excellent or high irrigation facility. Lastly, about 60% of the rural farm families in our survey reported their village’s economic develop- ment as medium, and 87% reported their village’s traffic condition as mediocre or better. Economic development and traffic conditions pro- vide information on accessibility to transportation, shops, healthcare facilities and opportunities for recreational activities.

4. Empirical framework and estimation procedure

Let us assume that the rural Chinese household maximizes utility and one-period optimization and under the assumption of risk-neu- trality is obtained by solving the following optimization problem:

= I C L H τ

Maximize U U( , ; , , ) (1)

subject to the constraints:

= + + − +

P C_c OFI R_L (P Q_q W X_x ) V (2)

=

Q Q F X H ϕ( , ; , ) (3)

Table 1

Sample distribution.

Province City Country Village Household

observation percent observation percent observation percent observation percent

Guangdong 15 21.74 15 14.29 112 18.82 547 20.23

Guizhou 11 15.94 27 25.71 100 16.81 239 8.84

Henan 5 7.24 7 6.67 24 4.03 230 8.51

Jiangsu 6 8.70 6 5.71 34 5.71 239 8.84

Jiangxi 10 14.49 15 14.29 132 22.18 587 21.71

Liaoning 6 8.70 10 9.52 53 8.91 221 8.17

Ningxia 4 5.80 7 6.67 41 6.89 226 8.36

Shanxi 5 7.25 6 5.71 28 4.72 201 7.43

Sichuan 7 10.14 12 11.43 71 11.93 214 7.91

Total 69 100 105 100 595 100 2704 100

Table 2

Age distribution of operators and farmland usage after retirement.

Operator age distribution Farmland usage decision Age group Number

of sample

Percent decision Number of

sample

Percent

18-30 691 25.55 Leave farmland idle 332 12.28

31-40 382 14.13 Hire farm labor 202 7.47

41-50 753 27.85 lease out farmland 1229 45.45

51-60 498 18.42 Pool farmland in land cooperatives

404 14.94

above 60 380 14.05 Unknown 537 19.86

Total 2704 100.00 Total 2704 100.00

whereUin Eq. (1) is the farm household’s utility (or welfare) function;

Cdenotes consumption goods purchased in the market;LandHdenote hours of leisure and stock of human capital by the operator, respec- tively; andτrepresents other factors such as age and number of children.

In Eq. (2),Pcis the price of consumption goodC,OFIis off-farm income earned by the household,R_Ldenotes rental income from land,Pqis the price of farm outputQ,Wxis the price of purchased inputsXincluding hired farm labor, andVdenotes other non-labor income of the house- hold, including pension income and profit shares from pooling land in the LSCS. Eq. (3) describes the technology of farm production, with ϕdepicting location-specific characteristics such as local climate, eco- nomic conditions, traffic, irrigation facilities in the village, and soil quality. The utility functionUin Eq. (1) and the non-stochastic pro- duction functionQin Eq. (3) are assumed to be concave, continuous, and twice differentiable. The maximizing of Eq. (1) is subject to the restrictions as defined by Eqs. (2) and (3). Equivalently, the optimality conditions can be obtained by first forming the following Lagrange function( )l:

= = + + + − +

−

l I C L H τ λ OFI R P Q F X H ϕ W X V P C

U U( , ; , , ) [ ( , ; , )

]

L q x

c (4)

Solving Eq. (4) for first-order conditions once can derive an

estimable form of farmland usage models in old age for rural Chinese households. The discernment of how rural Chinese households without successors would optimally choose farmland usage in retirement will depend on the rental rate of land, operator and household character- istics, land size, land and local economic conditions, off-farm income, pension income and farmers’experience with leasing market and par- ticipation in the LSCS. Since farmland in China may not be bought and sold in the market, old-age farm operators without successors essen- tially havefive choices of farmland usage: (I1) unknown; (I2) leaving farmland idle; (I3) leasing out farmland; (I4) hiring labor to farm the land; (I5) pooling the farmland in the LSCS.

In this study, a multinomial logit (MNL) model is used to examine the determinants of farmland usage of retiring farm households who are classified amongfive distinct strategies (M) based on their choice of farmland usage. In thefirst strategy, I1, farmland is left idle; in the second,I2, retiring or old-age operators hire farm labor to farm the land;

in the third,I3, farmland is leased out; in the fourth, I4, farmland is pooled in the LSCS; and in thefifth,I5, farmland usage in unknown (the base group in the present analysis). LetYj take the value 1 if thejth household chooses theqth choice of farmland usage); 0 otherwise. The relative odds (P) of farmland usage choices are expressed using the followingMNLmodel:

Table 3

Definition and summary statistics of variables used in the choice of farmland usage model, China (farmers > 50 years).

Variable Definition Mean standard deviation

Farmland usage Farmland usage decision by the current operator with no successor:

Pool farmland in Land Shareholding Cooperative System 0.120 0.325

Lease-out farmland 0.444 0.497

Hire labor to do farming 0.074 0.262

Leave farmland idle 0.144 0.351

Unknown (base group) 0.219 0.414

Dependence on farmland

Part-time worker Share of part-time workers to total family labor, with off-farm job 0.243 0.325

Full-time worker Share of full-time workers to total family workers with off-farm job 0.371 0.328

Off-farm income Share of off-farm income to total family income 0.643 0.338

Grain subsidy_01 =1 if the grain production subsidy goes to the contracted farmer (farmer who has the original contractual right or right of disposal); 0 otherwise

0.449 0.498

Grain subsidy_02 =1 if grain production subsidy goes to the operator; 0 otherwise 0.401 0.490

Grain subsidy_03 =1 if the household doesn’t know the distribution of grain production subsidy; 0 otherwise (base group) 0.150 0.358

Insurance =1 if the contracted owner of the farmland has pension; 0 otherwise 0.499 0.500

Household characteristics

Age Age of head of household (years) 60.407 7.498

Sex =1 if head household (HH) is male; 0 otherwise 0.677 0.468

Education_01 =1 if education of HH is primary, 0 else (base group) 0.526 0.500

Education_02 =1 if education of HH is junior high school; 0 otherwise 0.349 0.477

Education_03 =1 if education of HH is high school or above; 0 otherwise 0.125 0.331

Off-farm work =1 if HH has the experience working off-farm; 0 otherwise 0.457 0.498

Family income_01 =1 if household income over 50 thousand yuan; 0 otherwise 0.227 0.419

Family income_02 =1 if household income between 30-50 thousand yuan; 0 otherwise 0.235 0.424

Family income_03 =1 if household income between 10-29 thousand yuan; 0 otherwise 0.359 0.480

Family income_04 =1 if household income less than 10 thousand yuan; 0 otherwise (base group) 0.180 0.384

Lease-out =1 if the household had ever leased out farmland; 0 otherwise 0.216 0.412

Lease-in =1 if the household had ever leased-in farmland; 0 otherwise 0.107 0.309

Participation =1 if the household is the member of the Land (shareholding) Cooperative; 0 otherwise 0.055 0.227 Land characteristics

Farmland area Area of contracted farmland (mu) 1.660 0.690

Fertility_01 =1 if land fertility of contracted farmland is excellent or good; 0 otherwise 0.123 0.329

Fertility_02 =1 if land fertility of contracted farmland is mediocre; 0 otherwise (base group) 0.530 0.499

Fertility_03 =1 if land fertility of contracted farmland is poor; 0 otherwise 0.347 0.476

Irrigation_01 =1 if irrigation condition excellent or high; 0 otherwise 0.255 0.436

Irrigation_02 =1 if irrigation condition mediocre; 0 otherwise (base group) 0.371 0.483

Irrigation_03 =1 if irrigation condition is poor; 0 otherwise 0.374 0.484

Village characteristics

V_Development_01 =1 if village economic development within the town is high; 0 otherwise 0.202 0.401

V_Development_02 =1 if village economic development within the town is medium; 0 otherwise (base group) 0.599 0.490

V_Development_03 =1 if village economic development within the town is low; 0 otherwise 0.199 0.400

V_ traffic_01 =1 if traffic condition of the village is excellent or good; 0 otherwise 0.438 0.496

V_ traffic_02 =1 if traffic condition of the village is mediocre; 0 otherwise (base group) 0.433 0.496

V_ traffic_03 =1 if traffic condition of the village is poor; 0 otherwise 0.129 0.335

Province Dummy variable for nine provinces

⎜ ⎟

⎛

⎝

⎞

⎠

=Z′ +Θ +ε j= n q= − log P

P^jq (1, ...., ), (1, ..., M 1)

jM

j q j

(5) where log is the natural logarithm, Z is a vector of exogenous ex- planatory,Θis a vector of parameters to be estimated, andεis a random disturbance term. The means of explanatory variables as defined by vectorZand based on the distinctMstrategies of choice of farmland usage are presented in Table 3. The conditional probability for the choice q is derived as in the following [for more detail, seeGreene (2002), p. 721]:

= = = ′

∑ ′ =

=

Z Θ Z Θ

P Prob(Y 1) exp( )

exp(^{j q} ), q (1, ...,M-1)

k M

j q

jq jq

1 (6)

which, alternatively, can be written as:

= =

=

′

′

′ + ∑

+ ∑

=

−

=−

P , q (1, ...,M-1)

P .

Z Θ Z Θ

Z Θ jq

exp( )

1 exp( )

jM 1

1 exp( )

j q k

M j q

kM j q 1

1

11

(7) Since we cannot interpret the parameters of multinomial models directly, we can use marginal effect of parameters for estimation (Williams, 2012). For the rural householdi, the marginal effect of a change in thekth independent variable on the probability that decision of farmland operationjis the outcome as follow.

=∂ =

∂ = ∂

ME Y j ∂

X

F X Z X

Pr( ) ( , )

ijk i

ik

j i

ik (8)

Therefore, we presented the MEs (marginal effects) corresponding to the probabilities of the four choices of farmland usage by rural Chinese households in Table 4, except for the base group (unknown usage).

5. Results and discussion

Table 4 reports estimation results for theMNL model which also were reached based on maximum likelihood and robust variance esti- mation methods. The reference choice of farmland usage category for the MNL model isI5, reflecting the decision by the farm operator not knowing what to do with the farmland.¹¹ Accordingly, and as men- tioned earlier, the estimated coefficientsΘ_q(q = 1,…, M-1) measure the marginal effect of the regressors in vectorZj(see Eq. (5)) on the like- lihood of a farm operator’s choice of farmland usage categoryI1,I2,I3, orI4relative toI5. A positive regression coefficient means that an in- crease in the explanatory variable is associated with increased prob- ability of a categoryq(i.e.,I1,I2,I3orI4) relative to categoryM(i.e.,I5).

Before expounding on the results, it is worth mentioning that the esti- mated model demonstrated a good predictive capability as indicated by a McFadden pseudo-R²value of 0.11.¹²

Turning our attention to the variables inTable 4, the coefficient on the share of full-time off-farm workers in the household has a positive and statistically significant effect on the choice of leasing out farmland.

The marginal effect of 0.083 suggests that with a higher share of full- time off-farm workers in the rural Chinese household, the retiring op- erator without a successor is 8.3% more likely to lease out farmland for future usage than households without full-time off-farm workers.

Direct subsidy for grain production also can help improve rural households’income. The coefficient on Grain subsidy_02 is positive, negative and statistically significant for the choice of leasing out farmland and pooling farmland in the LSCS, respectively. Findings

indicate that in cases where the renter of the land receives the grain subsidy, the contracted operator (the retiring farmer), is about 11%

more likely to lease out the farmland and about 7% less likely to pool the land for investment in the LSCS than operators who don’t know who receives grain subsidies. A plausible explanation for this may be that rents received from leasing out farmland may be higher than income received from investing the farmland in the LSCS. Perhaps the grain production subsidy has been capitalized into the farmland’s rental value, making it more attractive. Similarfindings on the capitalization of government payments into farmland rental value have been reported in the United States (Goodwin et al., 2011;Lence and Mishra, 2003).

Against a background of an aging population, economic crisis, and the process of urban-rural integration, the Chinese government in 2009 created the New Rural Social Pension (NRSP) program, in which gov- ernment subsidies were the main source of funds (Li et al., 2018). Ad- ditionally, the pension insurance induces an income effect for rural elders, reducing their dependence on farmland (Liu et al., 2015). Re- sults in Table 4reveal that old-age operators without successors but with access to pension plans are about 6% less likely to choose leasing out their farmland than operators without access to pension plans. Old- age operators without successors but with access to pensions also are about 5% more likely to pool their farmland in the LSCS and receive the profits as income than operators without access to pensions. Findings here may reveal that old-age operators are risk-averse, choosing a less- risky option for income generation, and may not like to negotiate leasing rates with potential renters.

Results inTable 4also show that male old operators without suc- cessors are about 6.2% less likely to leave farmland idle than are old- age operators who do not know what to do with the farmland. This finding is not surprising because the majority of the farm operators are male and those significant farming experiences may want to leave the farmland idle(Lin and Wang, 2014). Results inTable 4show that the coefficient on off-farm work experience (off-farm work) is negative and significant at the 5% level of significance. Results suggest that farm operators (or the heads of household, in our case) with off-farm work experience are 3.8% less likely than those without such experience to hire labor to cultivate the farm (Column 5,Table 4) or to continue farming. Ourfinding is consistent withWu et al. (2017), who argue that off-farm work experience improves households’ cognitive, decision- making ability, and the choice of livelihood-working on the farm or both on and offfarm jobs.

Results inTable 4 reveal that the coefficients on family income categories for incomes greater than 50,000 Yuan and incomes of 10,000–29,000 Yuan are negative and significant at the 5% and 10%

level of significance, respectively. Result suggests that compared to rural Chinese households with less 10,000 Yuan in income, rural Chi- nese households with incomes of 50,000 or more and of 10,000–29,000 Yuan are 7.6% and 4.6% less likely to leave the farmland idle, re- spectively.

Table 4reveals that rural Chinese households that have experience with leasing out farmland are about 13% more likely to lease it out (Column 6,Table 4). Further, the coefficient on leasing in is negative and significant at the 10% level of significance. Result suggests that retiring rural Chinese households without successors are 8.4% less likely to leave farmland idle than households without leasing-in ex- perience. The coefficient on participation in cooperatives is positive and significant at the 5% level of significance. Finding suggests that old-age rural Chinese households with a membership in land cooperatives and without successors are 6.5% more likely to pool their land in the LSCS for investment and receive the share of the profits than households that do not belong to cooperatives.

The coefficient on farmland size is negative and statistically sig- nificant at the 10% level of significance for the choice of leaving farmland idle, but it is positive and statistically significant at the 10%

level of significance for the choice of leasing out farmland. These findings suggest that old-age rural Chinese households with large

11In estimating the MNL model as described in Eq. (5), the equation for the odds ofq versusMis:

12McFadden pseudo-R², which is suggested by McFadden (1973, p. 122), can be ap- plied to any model estimated by maximum likelihood methods as in the cases of theMNL and Tobit regression models discussed above. As defined, it is a scalar measure which varies between 0 and 1 and is computed as follows:

farmland but without successors are 4% less likely to leave the farmland idle, compared to rural Chinese households without successors and with unknown usage of farmland. On the other hand, old-age rural Chinese households with large farmland but without successors are about 5%

more likely to lease out farmland compared to rural Chinese households without successors and with unknown usage of farmland. Findings here suggest that large farms are more likely to use the farmland in pro- duction agriculture and that old-age farmers are rent seekers.

Turning our attention to the land and village characteristics, the results show that, compared to mediocre-quality land, poor-quality farmland reduces the likelihood of pooling of farmland as investment in

the LSCS by about 5%. Village economic development conditions also have a significant effect on farmland usage of old-age rural Chinese households. For example, the coefficient on village development_01 is negative and significant at the 10% level of significance in the choice of leasing out farmland usage. Thisfinding suggests that old-age rural Chinese households without successors and located in villages with high economic development conditions are about 7% less likely to lease out farmland than households in villages with low economic develop- ment.¹³A plausible explanation could be that farmers in these villages Table 4

Multinomial logit estimates of factors affecting farmland usage, farm operators without successors, China.

Variable Leave farmland

idle

Hire farm labor

Lease-out farmland

Pool farmland in land cooperative

coefficient MEs coefficient MEs coefficient MEs coefficient MEs

Part-time worker −0.086 (0.417)

−0.013 0.180 (0.485)

0.012 0.133

(0.306)

0.049 −0.382

(0.440)

−0.045 Full-time worker 0.595

(0.430)

0.025 −0.215

(0.561)

−0.043 0.622^* (0.331)

0.083^* 0.592

(0.452)

0.020

Off-income 0.524

(0.449)

0.062 −0.226

(0.529)

−0.018 −0.068 (0.323)

−0.040 0.064 (0.442)

0.005 Grain subsidy_01 0.123

(0.360)

−0.011 0.459 (0.497)

0.017 0.350

(0.303)

0.054 0.084

(0.399)

−0.016 Grain subsidy_02 −0.252

(0.359)

−0.036 0.024 (0.515)

0.000 0.295

(0.302)

0.108^** −0.550

(0.421)

−0.066^*

Insurance −0.383

(0.253)

−0.036 0.321 (0.314)

0.029 −0.238

(0.191)

−0.060^* 0.329 (0.266)

0.047^**

Age 0.367

(1.118)

0.036 −0.747

(1.438)

−0.060 −0.086 (0.871)

−0.083 1.261 (1.220)

0.130 Sex −0.509^*

(0.274)

−0.062^** 0.259 (0.363)

0.019 0.048

(0.212)

0.022 0.163

(0.300)

0.018

Education_02 0.275

(0.296)

0.009 0.460

(0.360)

0.018 0.275

(0.227)

0.029 0.087

(0.315)

−0.013

Education_03 −0.583

(0.428)

−0.032 −0.132 (0.506)

0.014 −0.526^*

(0.303)

−0.076 −0.132 (0.402)

0.023

Off-farm work 0.402

(0.270)

0.022 −0.303

(0.334)

−0.038^** 0.354^* (0.204)

0.043^* 0.459

(0.279)

0.024 Family income 01 −0.808^*

(0.427)

−0.076^** 0.233 (0.565)

0.030 −0.276

(0.325)

−0.030 0.087 (0.459)

0.031 Family income_02 −0.471

(0.396)

−0.032 0.235 (0.542)

0.032 −0.342

(0.316)

−0.047 −0.227 (0.445)

0.000 Family income_03 −0.612^*

(0.359)

−0.046^* 0.103 (0.493)

0.025 −0.389

(0.282)

−0.054 −0.061 (0.395)

0.021

Lease-out −0.269

(0.335)

−0.040 −0.458 (0.433)

−0.035 0.368 (0.236)

0.129^*** −0.343

(0.351)

−0.045

Lease-in −0.496

(0.487)

−0.084^* 0.420 (0.475)

0.017 0.367

(0.314)

0.070 0.515

(0.405)

0.034

Participation −0.525

(0.621)

−0.077 −0.033 (0.728)

−0.009 0.205 (0.432)

0.039 0.699

(0.549)

0.065^**

Farmland area −0.222

(0.221)

−0.037^* 0.269 (0.267)

0.014 0.193

(0.162)

0.047^* 0.010

(0.222)

−0.009

Fertility_01 −0.145

(0.408)

−0.029 0.059 (0.530)

−0.003 0.131 (0.310)

0.012 0.446

(0.418)

0.038

Fertility_03 0.296

(0.313)

0.029 0.037

(0.375)

−0.001 0.139 (0.236)

0.036 −0.408

(0.336)

−0.051^*

Irrigation_01 −0.628^*

(0.340)

−0.022 −0.943^**

(0.458)

−0.034 −0.531^**

(0.255)

−0.026 −0.610^* (0.363)

−0.016

Irrigation_03 0.007

(0.322)

0.002 0.152

(0.384)

0.011 −0.117

(0.251)

−0.047 0.293 (0.346)

0.033 V_Development_01 0.296

(0.332)

0.044 0.162

(0.419)

0.015 −0.246

(0.256)

−0.074^* −0.039 (0.345)

0.004 V_Development_03 0.452

(0.331)

0.030 1.085^***

(0.382)

0.060^** 0.056

(0.268)

−0.067 0.442 (0.358)

0.025

V_ traffic_01 0.338

(0.275)

0.004 0.366

(0.339)

0.004 0.438^**

(0.209)

0.049 0.389

(0.289)

0.008

V_ traffic_03 0.322

(0.382)

0.040 −0.183

(0.495)

−0.013 −0.087 (0.301)

−0.037 0.114 (0.404)

0.013

Province YES YES YES YES YES YES YES YES

Number of observations = 878, LR chi²= 262.76, Pseudo R²= 0.106, Log likelihood=−1113.785.

Notes:Statistically significant at: *10, * *5 and * * *1 percent levels; standard errors are report in the parentheses.

ME(dy/dx) for factor levels is the discrete change from the base level.

13Compared to rural Chinese households located in a mediocre economic

may be opting for alternative uses of farmland. Finally, results in Table 4show that old-age rural Chinese households without successors and located in villages with low economic development conditions are about 6% more likely to hire farm labor to do farming than households in villages with high economic development. A plausible explanation could be that these villages have surplus labor willing to work at lower wages.

6. Conclusions and policy implications

China has a peculiar farmland market, and the populations living in urban and rural areas are strictly defined by residence or Hukou. The government allocates operating rights to farmland to rural Chinese households. The farmland is not a marketable asset, though households can pass it on to the next generation. Farmland is considered a retire- ment security for many rural Chinese households. However, with the share of household income from farming decreasing and with techno- logical development and labor migration to urban areas increasing, rural China faces several problems that need attention from researchers, residents, academics, and policymakers. One such issue is old-age rural Chinese households that face an unknown future of farmland usage in the absence of successors. In China in particular, rural farmland mar- kets face institutional constraints that may threaten the sustainable development of agriculture.

Aging farmers’choices in succession planning will affect the amount of farmland in production, the prosperity and sustainability of pro- duction agriculture, and ultimately a nation’s food security. This study examined the factors affecting farmland usage by old-age farmers in the absence of successors. We used multinomial regression and farm-level data from aging rural Chinese households to investigate the objective of this study. Results from this study reveal that increased off-farm work by family members increases the likelihood of farmland being leased out by 8.3%. In cases where the grain subsidy goes to operator of the land (not to the contracted family), the likelihood of leasing out farmland increases, but the likelihood of pooling farmland into co- operatives as a shareholder decreases. Findings here may reveal capi- talization of grain subsidies in the rental value of the farmland. Off-farm work experience by the old-age rural Chinese household decreases the likelihood of using hired labor but increases the likelihood of leasing out farmland. Thisfinding may be signaling a moral hazard problem and the monitoring costs of hiring labor. This study also reveals that old-age rural Chinese households’experience in the farmland leasing market and the LSCS are important factors in determining future farmland usage.

Findings from this study reveals that old-age farmers without suc- cessors are less likely to leave farmland idle and more likely to lease out farmland. Experience with the LSCS is likely to increase farmland usage in farming as households pool their farmland as investments in LSC and receive profits as shareholders. Findings here provide evidence to pol- icymakers that farmland rental markets, albeit informal, are providing much-needed help to rural Chinese households. If policymakers’goal is to keep the land in agriculture, designing incentives that promote ef- ficiency and productivity in the land market may help the future of agriculture and of old-age farmers. Relatively richer old-age rural Chinese households without successors are less likely to leave farmland idle. Rural Chinese households with pension plans are more likely to choose a less-risky option when deciding on the farmland usage. These households, in the absence of successors, are more likely to pool their farmland in the LSCS.

Finally, this study found that male old-age operators with large farms are less likely to leave farmland idle, suggesting that old-age farmers without successors are rent seekers.

To encourage sustainable production agriculture and increase the nation’s food security, policymakers could use these results to design policies aimed at aging farmers who lack successors. Policies based on ourfindings could give these farmers the incentives to make decisions that will keep their farmland in production and prevent farmland from becoming idle or being converted to other uses. Pension plans for rural Chinese household could encourage retiring farmers without successors to pool their farmland in the LSCS, thereby ensuring increased use of land in production agriculture. Lastly, if the policymakers’goal is to increase village economic development conditions, such goals may lead to alternative uses of farmland and take it out of production agriculture.

Conversely, a lack of incentives for village economic development could encourage old-age rural Chinese households to hire labor to farm the land and thereby keep the land in agriculture and increase food se- curity.

Acknowledgments

This paper was sponsored by the key project of National Natural Science Foundation of China: “Research on Rural Land and Relevant Element Market Cultivation and Reform” (No. 71333004); “Farmland Titling: Realistic Background, Policy Objective and Effect Evaluation”(No.

71742003); foreign joint training program for doctoral students of South China Agricultural University (No. 2017LHPY006).

References

Baker, J.R., Lobley, M., Whitehead, I., 2016. Keeping It in the Family: International Perspectives on Succession and Retirement on Family Farms. Routledge.

Bedemo, A., Getnet, K., Kassa, B., 2013. Determinants of household demand for and supply of farm labour in rural Ethiopia. Aust. J. Labour Econ. 16, 351.

Benjamin, D., 1992. Household composition, labor markets, and labor demand: testing for separation in agricultural household models. Economet.: J. Economet. Soc. 287–322.

Bertoni, D., Cavicchioli, D., 2016. Farm succession, occupational choice and farm adap- tation at the rural-urban interface: the case of Italian horticultural farms. Land. Use Policy 57, 739–748.

Calus, M., Van Huylenbroeck, G., Van Lierde, D., 2008. The relationship between farm succession and farm assets on Belgian farms. Sociologia Ruralis 48, 38–56.

Che, Y., 2016. Off-farm employments and land rental behavior: evidence from rural China. China Agric. Econ. Rev. 8, 37–54.

Chen, A., 2016. The politics of the shareholding collective economy in China’s rural villages. J. Peasant Stud. 43, 828–849.

Corsi, A., Salvioni, C., 2017. Once part-timer always part-timer? Causes for persistence in offfarm work state of farmers. Bio-Based Appl. Econ. 6, 159–182.

Felson, M., 2017. Rural Settlement and Land Use. Routledge.

Fichtner, J.J., 2018. Global aging and publicfinance. Bus. Econ. 1–7.

Goodwin, B.K., Mishra, A.K., Ortalo-Magné, F., 2011. The Buck Stops Where? The Distribution of Agricultural Subsidies. National Bureau of Economic Research.

Greene, W.H., 2002. LIMDEP: Version 8.0: Econometric Modeling Guide. Econometric Software..

Groeneveld, J., Müller, B., Buchmann, C.M., Dressler, G., Guo, C., Hase, N., Hoffmann, F., John, F., Klassert, C., Lauf, T., 2017. Theoretical foundations of human decision- making in agent-based land use models–A review. Environ. Modell. Softw. 87, 39–48.

Harris, J.M., Mishra, A.K., Williams, R.P., 2012. The impact of farm succession decisions on thefinancial performance of the farm. 2012 Annual Meeting, August. pp. 12–14.

Inwood, S.M., Sharp, J.S., 2012. Farm persistence and adaptation at the rural–urban in- terface: succession and farm adjustment. J. Rural Stud. 28, 107–117.

Joosse, S., Grubbström, A., 2017. Continuity in farming-not just family business. J. Rural Stud. 50, 198–208.

Kimhi, A., 2009. Demand for on-farm permanent hired labour on family holdings: a comment. Eur. Rev. Agric. Econ. 36, 447–452.

Kimhi, A., Lopez, R., 1999. A note on farmers’retirement and succession considerations:

evidence from a household survey. J. Agric. Econ. 50, 154–162.

Krueger, D., Mitman, K., Perri, F., 2016. Macroeconomics and Household Heterogeneity, Handbook of Macroeconomics. Elsevier, pp. 843–921.

Kung, J.K., 2002. Off-farm labor markets and the emergence of land rental markets in rural China. J. Comp. Econ. 30 (2), 395–414.

Lence, S.H., Mishra, A.K., 2003. The impacts of different farm programs on cash rents.

Am. J. Agric. Econ. 85, 753–761.

Li, J., Rodriguez, D., Tang, X., 2017. Effects of land lease policy on changes in land use, mechanization and agricultural pollution. Land. Use Policy 64, 405–413.

Li, Q., Wang, Y., Zhao, Y., 2018. The impact of China’s new rural pension program on elderly labor, grandchild care, and old-age support. Fem. Econ. 1–23.

Lin, L., Gu, Q., 2017. Analysis on decision-making and performance evaluation of farmers’

participation in the rural land shareholding cooperative: based on survey of 1831 rural households in Jiangsu Province. J. Agrotech. Econ. 49–60 (in Chinese).

Lin, H., Wang, H., 2014. Problems of family farms in China and countermeasures. Asian (footnote continued)

development,

Agricultural Res. 6, 12.

Liu, H., Han, X., Xiao, Q., Li, S., Feldman, M.W., 2015. Family structure and quality of life of elders in rural China: the role of the new rural social pension. J. Aging & Soc.

Policy 27, 123–138.

Lobley, M., Baker, J.R., 2012. Succession and Retirement in Family Farm Businesses.

Keeping It in the Family: International Perspectives on Succession and Retirement on Family Farms. pp. 1–20.

Masters, W., 2013. Urbanization and Farm Size in Developing Countries: Implications for Agricultural Research. Synthesis of a Foresight Study of the Independent Science and Partnership Council. CGIAR Independent Science and Partnership Council (ISPC), Rome Italy.

Min, S., Waibel, H., Huang, J., 2017. Smallholder participation in the land rental market in a mountainous region of Southern China: impact of population aging, land tenure security and ethnicity. Land Use Policy 68, 625–637.

Mishra, A., El-Osta, H., Shaik, S., 2010. Succession Decisions in U.S. Family Farm Businesses. J. Agric. Resour. Econ. 35, 133–152.

Mishra, A.K., Sandretto, C.L., 2002. Stability of farm income and the role of nonfarm income in U.S. agriculture. Appl. Econ. Perspect. Policy 24 (1), 208–221.

National Bureau of Statistics, 2016. (NBS) of the People’S Republic of China. Accessed on 24 May, 2018. http://www.stats.gov.cn/tjsj/tjgb/nypcgb/qgnypcgb/201712/

t20171215_1563599.html.

Nie, J., 2018. Security dependence, old-age stage and the willingness of rural elderly to transfer farmland management rights: based on the questionnaire survey of rural elderly in Hubei Province. J. Nanjing Agric. Univ.(Soc. Sci. Ed.) 41–53 in Chinese.

Potter, C., Lobley, M., 1992. Ageing and succession on family farms: the impact on de- cision‐making and land use. Sociol. Ruralis 32, 317–334.

Promsopha, G., 2018. Risk‐Coping, Land tenure and Land markets: an overview of the literature. J. Econ. Surv. 32, 176–193.

Ren, Y., Bian, Y., He, T., 2017. Characterizing the land shareholding cooperative: a case study of Shanglin Village in Jiangsu, China. Sustainability 9, 1175.

Rizwan, M., Deyi, Z., Nazir, A., Osati, R., Traore, L., Sargani, G., 2017. Determinants and

choices of off-farm work among rice farmers in a developing country. JAPS: J. Anim.

Plant. Sci. 27, 1993–2002.

Shirai, Y., Fox, J., Leisz, S.J., Fukui, H., Rambo, A.T., 2017. The influence of local non- farm employment on rural household structure in Northeast Thailand. J. Rural Stud.

54, 52–59.

Sottomayor, M., Tranter, R., Costa, L., 2011. Likelihood of succession and farmers’atti- tudes towards their future behaviour: evidence from a survey in Germany, the United Kingdom and Portugal. Int. J. Sociol. Agric. Food 18, 121–133.

Stiglbauer, A.M., Weiss, C.R., 2000. Family and non-family succession in the Upper- Austrian farm sector. Cahiers d’Economie et de Sociologie Rurales 54, 5–26.

Wang, H., Riedinger, J., Jin, S., 2015. Land documents, tenure security and land rental development: panel evidence from China. China Econ. Rev. 36, 220–235.

Wang, Q., Zhang, X., 2017. Three rights separation: China’s proposed rural land rights reform and four types of local trials. Land. Use Policy 63, 111–121.

Williams, R., 2012. Using the margins command to estimate and interpret adjusted pre- dictions and marginal effects. Stata J. 12, 308.

Wu, Y., 2017. The pricing mechanism innovation of Land remediation in hollow villa- ge—investigation and thinking on Houzhai River Basin in Guizhou Province. Open. J.

Soc. Sci. 5, 196.

Wu, Z., Li, B., Hou, Y., 2017. Adaptive choice of livelihood patterns in rural households in a farm-pastoral zone: a case study in Jungar, Inner Mongolia. Land Use Policy 62, 361–375.

Yao, J., 2009. Social benefit evaluation on regional land consolidation based on social security function of land-a case of Nanjing City. Asian Agric. Res. 1, 37–41.

Yi, F., Sun, D., Zhou, Y., 2015. Grain subsidy, liquidity constraints and food secur- ity—impact of the grain subsidy program on the grain-sown areas in China. Food Policy 50, 114–124.

Zhong, W., Luo, B., 2013. Endowment effect, property strength and restraint of agri- cultural land transfer: farmer household’s survey data from Guangdong Province.

Issues Agric. Econ. 3, 6–15 in Chinese.