Is Land-based Resettlement Still Appropriate for Rural People in China? A Longitudinal Study of Displacement at the Three Gorges Dam

Transcription

1 Is Land-based Resettlement Still Appropriate for Rural People in China? A Longitudinal Study of Displacement at the Three Gorges Dam Brooke Wilmsen ABSTRACT Now complete, the effects of the Three Gorges Dam on rural livelihoods can be observed. This article presents the first longitudinal analysis of the livelihoods of two groups of rural households that were first surveyed in 2003, just after inundation, and again in It finds remarkable gains in income and social well-being along with moderate improvements in food security and income equality. This study explores the livelihood strategies associated with these outcomes, namely land consolidation, agricultural specialization, livelihood diversification and migration. These strategies are discussed in the context of the dramatic changes that have taken place in the Three Gorges Resettlement Area over the last decade and across China more broadly. The article concludes that while non-farm jobs are increasingly important, land remains an essential resource in the reconstruction of rural livelihoods, allowing rural households to respond to their drastically altered environment more effectively and at a pace of their own choosing. INTRODUCTION Conservative estimates suggest the number of people displaced by large dams worldwide is between 40 and 80 million people (Scudder, 2005; World Commission on Dams, 2000), the overwhelming majority of whom are poor, uneducated and often powerless rural people (Scudder, 2005). The rural nature of large dam displacements is unsurprising. Dams are normally located on large rivers running through steep gorges or valleys places often not conducive to urban living. The author would like to thank Dr Andrew van Hulten for statistical analysis and comments on earlier drafts and the three anonymous reviewers for constructive feedback. The author would also like to thank Professor Duan Yuefang of China Three Gorges University (CTGU) and Johanna Higgs of La Trobe University for assistance with fieldwork. The author gratefully acknowledges the receipt of Australian Research Council grant DE that fully supported the research reported in this article. Development and Change 49(1): DOI: /dech C 2018 International Institute of Social Studies.

2 Land-based Resettlement in Rural China 171 The impacts of land acquisition, displacement and resettlement on livelihoods are typically adverse. The literature documents the negative effects of dams on rural people, which includes fewer options for economic activities (Kura et al., 2014); land scarcity, lower land quality and changes to control over resources (Dao, 2016); high rates of unemployment and enduring poverty (Aiken and Leigh, 2015); and forced share-cropping due to land becoming unaffordable (Fujikura and Nakayama, 2013). These are just a small selection of the more recent publications from a catalogue that began in 1956 when Colson (1960) and Scudder (1962) produced the earliest studies of the Gwembe Tonga. This literature illustrates the enduring cost of large dams for rural people. Resettling rural people poses a particular set of challenges for social development specialists. Whilst urbanites risk losing a house, a business or a job and social networks, rural dwellers also risk losing farmland, grazing lands and access to natural resources such as forest products, wildlife and fisheries the foundations of their agricultural livelihoods (IFC, 2012). Indeed, land loss is the principal form of decapitalization and pauperization of displaced people as it takes away natural and human-made capital. This is a central observation of the Impoverishment Risks and Reconstruction (IRR) model (Cernea, 1997), which heavily informs the praxis of large development banks and scholarly research. Rural households are susceptible to all eight risks identified by this model: landlessness, joblessness, homelessness, marginalization, food insecurity, loss of access to common property resources, increased morbidity and community disarticulation (see ibid.: 1569 for discussion). International policies for land acquisition recommend land-based compensation as a starting point. That is, where livelihoods are land based or where land is collectively owned, replacement land is offered as recompense; see, for example, the IFC s Performance Standard 5C (IFC, 2012); the World Bank s Operational Policy 4.12 (World Bank, 2001) and the Asian Development Bank Safeguard Policy Statement 2009 (ADB, 2009). Most recently, Article 14 of the World Bank s Environmental and Social Framework states: Where livelihoods of displaced persons are landbased or where land is collectively owned, the Borrower will offer the displaced persons an option for replacement land... unless it can be demonstrated to the Bank s satisfaction that equivalent replacement land is unavailable (World Bank, 2016: 56). The problems with cash compensation have been extensively documented (see Jaywardene, 1995; Mathur, 1999; Parasuraman, 1999; Partridge, 1989; Tan and Wang, 2003). Moreover, the World Bank s Environmental and Social Framework (Article 35c) states cash assistance alone... frequently fails to provide affected persons with the productive means or skills to restore livelihoods (World Bank, 2016: 60). On the other hand, land offers a tangible resource base for rebuilding livelihoods. If it is of adequate size and quality, securely tenured and located with easy access to markets, land acts as a

3 172 Brooke Wilmsen safety net for rural people attempting to diversify their livelihood strategies in response to new jobs off the farm. However, the biggest barrier to landfor-land compensation is the scarcity of unoccupied fertile land. Therefore, even though land-based compensation is preferable, cash compensation is commonly practised. The resettlement of rural people for the construction of large dams in China legislates a land-based strategy. Article XIII of the Regulations for Resettlement for Large and Medium Hydropower and Water Conservation Projects requires the provision of replacement agricultural land for displaced rural households (State Council, 2006). In principle, the Chinese government takes complete responsibility for resettling displaced persons (Shi et al., 2012: 224), but in practice the resources devoted to resettlement and livelihood reconstruction vary between projects and locations (Wilmsen, 2016a). In some instances, the Chinese government has offered rural people employment in government enterprises, support to improve low-yield land, living allowances, and a range of other post-resettlement income supports (Shi et al., 2012). Although the Regulations for Resettlement for Large and Medium Hydropower and Water Conservation Projects (2006) continue to govern the resettlement of rural people for reservoirs in China, change is ahead. Rather than land-based resettlement for rural people, future dam resettlements will require rural-to-urban movements (Chen, 2016). Termed New-Type Urbanization Approach to Reservoir Resettlement, such movements would be predominantly from rural areas to nearby urban areas, rather than to the east coast provinces of China (Chen, 2016; Wilmsen, 2017). This shift dovetails with the Chinese government s broader push to boost urbanization, the services sector and urban consumption, thereby reducing the economy s reliance on investment- and export-led growth. Although there is currently no further information available regarding the administrative, regulatory and legal changes, such a shift in the practice of displacement and resettlement is sure to have implications for the lives of rural people forcibly displaced by dam construction. Given the proposed changes, this article asks: Is land-based resettlement still appropriate for rural people in China?. It begins by presenting the legal and planning principles guiding rural resettlement at the Three Gorges Dam and reviewing the existing literature assessing the outcomes of resettlement for displaced rural households. This is followed by a description of research methods and an overview of the national, regional and local-level changes that have impacted the study population since the initial survey in The longitudinal results are then presented to demonstrate how livelihood outcomes and household strategies have changed between 2003 and The final section discusses these household strategies in more detail, linking them to broader processes of change, and reiterates the importance of land for rural households in light of recently proposed changes to resettlement policy in China.

4 Land-based Resettlement in Rural China 173 PLANNING THE DISPLACEMENT AND RESETTLEMENT OF RURAL HOUSEHOLDS AT THE THREE GORGES PROJECT The Three Gorges Project (TGP) on the Yangtze River is the world s largest dam. As such, its social, cultural, environmental and economic effects have been extensive (Wilmsen, 2016b). Some 1.4 million people were displaced by the dam s construction, of which around 87 per cent were rural people (Heggelund, 2004). For farmers, the greatest challenge to their livelihoods was the radical reduction of the natural resource base in the reservoir area. The dam inundated some 238 km 2 of farmland and 50 km 2 of orange groves (Chau, 1995). Over the past two decades, those affected by the TGP have been busily responding to the extensively altered environment. From the outset, rural resettlement at the TGP was land based (Shi et al., 2012). Under the first iteration of the Regulations on Resettlement for the Construction of the TGP on the Yangtze River (henceforth the 1993 Regulations), rural resettlement relied on clearing reclaimable land (land not in productive use) to replace inundated farmland, and the intensification of agricultural practices. 1 In the mid-1990s, resettled rural households began to convert steep forested land into farmland. However, when floods in 1998 revealed that deforestation had increased the risk of flooding and erosion, the Chinese government enacted the National Forest Conservation Programme (NFCP). This restricted deforestation in mountainous areas and required the reforestation of the steep land. 2 In 2001, a new iteration of the 1993 Regulations (Decree No. 299) was promulgated that included plans to relieve the pressure on the rural areas of the Three Gorges Resettlement Area (TGRA) (see State Council, 2001). The 2001 Regulations encouraged rural people to move off the land. About 190,000 rural residents (15 per cent of the total) were resettled out of the TGRA to 11 provinces through a government-organized programme (Xu et al., 2013: 116). In addition, up to 200,000 rural residents were moved into nearby towns (Wu and Liao, 1999). However, it is the rural people who remained in the village and who lost their land, house, or both that are the focus of this study. Their displacement 3 was either physical, economic or both. 1. Decree 126, 19 August 1993 Regulations, Clause 9 of 43: Rural resettled people should devote major efforts to develop agriculture, develop reclaimable land and reform low-yield land. They should cultivate high-yield land, economic gardens, forestry, animal husbandry and fisheries, etc.. 2. Interview, Professor Duan Yuefang, China Three Gorges University, Yichang City, 16 August The term displacement is used in this article to refer to physical and/or economic displacement.

5 174 Brooke Wilmsen RURAL LIVELIHOODS IN 2003 Previous research reveals multiple losses associated with dam constructions around the world (for a description of these losses see Cernea, 1997; Chakrabarti and Dhar, 2010; Gupte and Mehta, 2007; Kothari, 1996). The first survey and interviews I conducted in early 2004 in the TGRA generated similar findings. The lives of the rural households displaced by the TGP were incredibly difficult. Losses typically included land, housing, productive resources, savings and social networks. The dam submerged the most fertile land in 2003 and the rural households interviewed unanimously mentioned the decreased fertility and increased incline of the remaining farmland. Rural people, lacking education and formal opportunities to learn new skills, responded to the displacement as best they could. They took temporary jobs linked directly to the dam s construction as well as the dismantling, relocation or replacement of submerged infrastructure. To bridge the gap between the cost of relocation and the compensation provided by the state, 4 households took out loans to finance new houses and businesses and to cover basic needs. These responses did not stave off impoverishment. Incomes declined and poverty increased (both relative and absolute) following displacement (Wilmsen et al., 2011). With depleted savings and increased debt, rural households were vulnerable to further shocks and stresses. At that time, investments in the region by the government and the subsidiaries of east coast firms did little to offset these effects. Those opportunities for employment that did exist were often not open to rural people. Put simply, rural development was largely ignored by government policy and farmers were flailing. Considering the timing of my first survey in early 2004, the difficulties facing rural households were unsurprising. Although more than a decade after the earliest relocations, many of the rural households in this sample had only recently lost land 5 or resettled up hill (in the late 1990s and early 2000s). The most fertile farmland had been submerged six months before the survey. The TGP itself was still under construction and the region was in a state of destruction and reconstruction. The hills of Badong and Zigui echoed with the sound of pick axes and the traditional work songs that accompany heavy labour. Major infrastructure was yet to be built for example, the Yichang-Wanzhou (Yiwan) railway to Enshi (which opened in 2010), the Yichang-Badong (Yiba) expressway 6 and the three river bridges of Badong county, which now connect the north and south banks of the Yangtze. Rural households were rebuilding and there was still the mat- 4. Compensation was based on quantifying material losses rather than the cost of replacing those losses. 5. First inundation occurred in June Between 1 and 15 June water levels rose from 95 m to 135 m above sea level. 6. It opened in 2014, although the section from Yichang to Ye San Guan was open by 2012.

6 Land-based Resettlement in Rural China 175 ter of future inundation: to 156 m in 2007, 170 m in 2008 and 175 m in Other studies of rural displacement in the TGRA present an equally negative picture, reporting a range of effects. These include high levels of dissatisfaction (Jun, 2000); reduced environmental carrying capacity of the remaining land (Duan and Steil, 2003); the feminization of agriculture (Tan et al., 2005); the difficulties of transferring rural migrants to secondary or tertiary industries (Tan, 2008); and high levels of outmigration (Hwang et al., 2007). A report commissioned by the Chinese government found compensation levels were too low, employment was a problem, training programmes were not always effective, and land to support the displaced farmers was lacking (Duan and Wilmsen, 2012; Tilt, 2015). However, as indicated by my own research, these studies were conducted whilst the TGP was still under construction. To date no independent research has been conducted about the longer-term effects of the TGP displacement for rural households, particularly since the completion of the construction phase. This dearth of longitudinal research on the TGP is characteristic of the broader literature in the field of Development Forced Displacement and Resettlement (DFDR) (Cernea, 1999; Downing, 2002; Scudder, 1997). However, longitudinal research is essential to understanding the cumulative impacts of displacement and resettlement (Scudder, 1997). Indeed, displacement and resettlement scholars suspect that the longer-term social impacts are underestimated (Bartolome and Barabas, 1990 cited in Scudder, 1997; Colson, 1971; Scudder, 1993), although there are few studies to support these suspicions. Moreover, longitudinal research can inform essential theory and improve praxis (Cernea, 1999; Downing, 2002; Scudder, 1997; Wilmsen and van Hulten, 2017). To this end, this research explores the reconstruction of livelihoods beyond the early years of displacement, to provide a better understanding of the responses of displaced rural people through time. METHODS This study applies a mixed-methods approach to data collection. A survey was conducted to capture livelihoods at three points in time (before displacement, 2003 and 2011) in two villages in Hubei province; Leijiaping village in Badong county and Changjiang village in Zigui county (the counties are shown in Figure 1). The surveys were completed in 2004 and These villages were located in counties that had similar levels of development before displacement, measured by GDP per capita, and different levels of development intervention during construction, according to investment in RMB per capita see McDonald et al. (2008) for further details. The physical and economic conditions of both villages were similar prior to displacement, with both cohorts highly dependent on farming for income, in particular citrus and pig production (see Table 1). There was one key

7 176 Brooke Wilmsen Figure 1. Location of Sample Sites Notes: Leijiaping village is 5 km east (downstream) of Badong county town. Changjiang village is 15 km west (upstream) of Zigui county town. Both are on the north bank of the river. Source: McDonald et al. (2008: 89). Table 1. Key Variables by Village (pre- to post-displacement) Changjiang Village (Zigui County) Leijiaping Village (Badong County) Key Variable Before After (2003) Before After (2003) Average land holding (mu/pp) Quality of land (% of land 52% 4% 59% 4% holders reporting very good quality land) Landlessness (% household 13% 10% 0% 11% reporting no land holdings) Primary crop Oranges Oranges Oranges Oranges Primary livestock Pigs Pigs Pigs Pigs Average compensation (yuan/pp) House material Mud and wood Brick and concrete Mud and wood Brick and concrete Floor space of house (m 2 ) Source: Author s own survey.

8 Land-based Resettlement in Rural China 177 difference: whilst Leijiaping village was primarily agricultural, Changjiang village had two village enterprises, a brick-making factory and a coal mine. In this sense, the farmers at Changjiang village could potentially supplement their incomes once their land was submerged. This provided an important point of comparison for the 2003 study (Wilmsen et al., 2011). Land losses at both sites were significant. However, households in Leijiaping village lost more land pre- to post-displacement (1.3 to 0.6 mu/pp) than those from Changjiang village (1.1 to 0.9 mu/pp) as shown in Table 1. In both villages, about 30 per cent of the households lost their houses and land, whilst 70 per cent lost only land. Land was not formally redistributed in either village; however, as will be described later, there was evidence of informal land transfers. In both villages, households reported a significant decline in land quality. For example, before displacement, 52 per cent of respondents reported their land quality as being very good, compared to 4 per cent after displacement (see Table 1). These declines in land quality reflect the fact that farmers were provided with new farmland further uphill following inundation. However, all households were relocated within the same county, meaning relocation distances were kept to a minimum. In Leijiaping village, some farming households were not provided with new farmland, resulting in a small increase in reported landlessness pre- to postdisplacement. In both villages, oranges remained the most important crop grown and pigs the most important type of livestock. Displacement and resettlement proceeded according to the Regulations. This prevented farmers from reclaiming wasteland following resettlement, which many may have attempted to do, to maintain their plot sizes. The Regulations also influenced the amount of compensation received by displaced households. The national standard for calculating compensation for land loss used the average production output value for the previous three years and multiplied it by a factor of 3 to 12. County governments stipulated the exact multiplication factor depending on local conditions. For Changjiang village the multiplication factor was 7.29 and for Leijiaping it was The tight correspondence between production declines and land loss explains why average compensation was higher in Leijiaping village (5757 yuan/pp) than in Changjiang village (3550 yuan/pp), as shown in Table 1. In addition to changes to their farmland, resettled households experienced significant changes to their housing stock. Prior to resettlement, most of the farmers lived in traditional, mud-brick houses with courtyards. At the resettlement sites, they built two-storey concrete structures that were densely packed together. Unlike their previous homes, these new houses were significantly larger and had running water and flushing toilets. As such, many respondents reported an improvement in their housing conditions as a result of resettlement. However, these improvements came at a cost. Government compensation was not enough to finance the construction of their new homes and most resettled households borrowed money from relatives and friends

9 178 Brooke Wilmsen to cover the shortfall. As a result, indebtedness increased dramatically following displacement. Sample Selection In the villages of Leijiaping and Changjiang, 170 rural households lost land to the TGP and were selected to take part in the 2003 survey 113 households in Leijiaping and 57 in Changjiang. In revisiting the villages in 2012 (the questionnaire asked about their 2011 livelihoods), 118 households agreed to complete the questionnaire: 31 in Changjiang village and 87 in Leijiaping village. Taken together, this represents 69 per cent of the original sample. 7 The 2003 questionnaire investigated household livelihoods. It asked respondents to reflect on their household s current situation as well as their recollections of their circumstances before displacement (around the late 1990s). The 2011 questionnaire repeated the same questions, with a few minor additions. In addition to the surveys, semi-structured interviews were conducted with village leaders and government officials at the county, provincial and national levels in 2004, with follow-up interviews conducted in 2012 (county and provincial interviews) and 2014 (national level interviews). As enterprises in the region are important to the Chinese government s development plan, interviews were also conducted with the managers of eight enterprises. These included repeat interviews with six enterprise managers who had been interviewed in 2004, and two interviews with the managers of companies that had been established after As in 2004, in-depth follow-up interviews were also completed in 2012 and 2013 with 15 households reporting the largest increases and decreases in income following displacement, and the lowest and highest overall incomes. CHANGING LIVELIHOOD CONTEXTS Before presenting the longitudinal results, it is important to recognize the important national, regional and local changes that have reshaped the study area since the initial survey was conducted in At the national level, three reforms are of particular relevance: i) the ongoing transition from a 7. Of the non-respondents to the 2011 questionnaire, 27 households could not be contacted because they had not provided any contact details in the 2003 survey in order to remain anonymous. In addition, nine households had moved away permanently and only returned at Spring Festival; 12 households were not at home but still resided in the same house as in the 2003 survey; three householders had passed away and any family members had moved on; and two households refused to participate due to illness.

10 Land-based Resettlement in Rural China 179 socialist to a market economy, ii) the relaxation of the hukou 8 household registration system, and iii) reduced restrictions on rural land transfers. At the regional level, targeted plans were formulated by the State Council to stimulate industrial growth and agricultural production in the TGRA. At the village level, new infrastructure was built that improved access to the county town for residents of Leijiaping, and at Changjiang village, one village enterprise was privatized and the other had closed. Much has been written about China s gradual transition to a market economy over the last three decades. While these economic reforms produced their most visible impacts in urban places (for example, Special Economic Zones [SEZs], rapid urbanization and economic growth) rural places were not immune; cities encroached on farmland, income inequality between rural and urban areas was exacerbated, and rural-to-urban migration increased (Long et al., 2010). The outmigration of working-age people hollowed out many villages, leaving the elderly and women as the primary agricultural workforce (ibid.). Reforms to the hukou system further disinhibited rural to urban migration. For example, in 2001, the State Council extended rights to access health care, education and housing to migrants living in cities with rural hukou (provided they could afford it). 9 Since 2014, rural people can officially settle in small and medium-sized cities (Zhang, 2014). As a result, urban places and urban jobs, both near and far, now play an increasingly important role in the rural lives of the study population. Meanwhile, the ongoing relaxation of the prohibition on land transfers produced important changes at the local level. At the time of the first survey, land use rights could not be transferred easily between farmers to help mitigate the land losses caused by the TGP. Since then, the Chinese government has relaxed restrictions on the transfer of land use rights in an attempt to encourage land consolidation and the industrialization of rural farming practices (see Wilmsen, 2016a for detailed discussion). In doing so, the Chinese government has addressed some of the limitations of the Household Responsibility System (HRS). 10,11 Within the study area, land transfers are now being realized through formal agreements, for example 8. Introduced in 1958, the hukou is a system of control that divides China s populace according to an urban or rural designation (Wilmsen, 2017). 9. It should be noted that the granting of these rights has varied across China s cities (Wing Chan and Buckingham, 2008). 10. Under the HRS, collectively owned land is contracted to individual households to manage. While this model is widely regarded to have fuelled agricultural growth and decreased poverty throughout China (Jia et al., 2012; Lin et al., 1996), it also divided the land into household-sized parcels, limiting the formation of agricultural economies of scale. 11. The most explicit articulation of this occurred in Document No. 1 in 2013, although informal land transfers were already widespread. Document No. 1 is the first major policy document released by the Central Committee of the Communist Party each year and details the priorities of the government for that year.

11 180 Brooke Wilmsen with corporations and village cooperatives, as well as informally between small landholders. Plans were also developed to stimulate the regional economy. Under the 1993 and the 2001 Regulations, displacement and resettlement was planned as a development opportunity. To this end, the State Council designed two main approaches. The first was the Partnership Support Programme (dui kou zhi yuan), the PSP, which encouraged eastern provinces to provide support to specific counties in the Three Gorges (Steil and Duan, 2002). The State Council called for the 20 wealthiest provinces and the municipalities of Beijing, Shanghai and Tianjin to provide financial support to the Three Gorges region. Following this initial investment, private companies were encouraged to invest in the region. County governments designed preferential policies to attract businesses and to develop particular sectors including the citrus industry and tourism. The second scheme was the Development Assistance Fund (DAF) under which the government deposits RMB per kilowatt produced by the Three Gorges Dam into a development fund. This amounts to 42,350, RMB per year for 10 years. Regional economic growth has been rapid since 2003, with the majority of dam construction activities completed by According to county economic data, the gross domestic product (GDP) of Badong and Zigui the two counties encompassing my study area doubled between 2004 and 2009 (the closest census to my two surveys) (Wilmsen, 2016b). Investment in the region also increased with many private enterprises setting up or expanding. 13 In 2012, Zigui was on its way to becoming an important site for logistics and investment. SEZs grew in size and number in both Badong 14 and Zigui 15 counties. In 2009, investment per capita in Badong surpassed Zigui for the first time since 1994 (458.9 RMB per capita in Badong compared to RMB per capita in Zigui) (Three Gorges Project Construction Committee, 2011). For the rural households, rapid regional economic growth created demand for farm produce and new non-farm job opportunities. When compared to 2003, village life also changed in both study sites. Firstly, more land was submerged. 16 In 2003, Leijiaping was more isolated than Changjiang village. However, by 2012 this situation was reversed. The 12. Approximately US$ 6,382,914 per year. 13. The government spent some 8 billion RMB on the construction of a 40,000 mu logistics hub at the southern side of Yangtze River (for more detailed information see Wilmsen, 2016b). 14. The economic development of the new county town continues to be hampered by unstable geology, isolation and complicated construction conditions. Recognizing these limitations, Badong county government encourages investment in several newly established SEZs closer to Yichang city. 15. Zigui s key SEZ grew from 17 large enterprises in 2003 to 40 large enterprises in 2012, including three high-tech enterprises and 12 foreign-backed enterprises (Yichang Government 2012), with a commensurate increase in employment. 16. Of the 1,720 mu of arable land in Changjiang village (Zigui county) in 2003, 22 per cent was submerged by In Leijiaping (Badong county), of the 800 mu of arable land available in 2003, 57 per cent had been submerged by 2011.

12 Land-based Resettlement in Rural China 181 addition of Badong Bridge meant that Leijiaping was directly connected to Badong county town. In 2003, there were two main non-farm employment options in Changjiang village: the brick-making factory and the coal mine. By the time of the second survey, in 2012, the mine was closed and the brickmaking factory privatized. 17 In 2003, Leijiaping had no village enterprise, and this was still the case in However, the new bridge enabled residents to set up small farm restaurants in the village, which were frequented by urban residents from the county town. To attract local tourism the county government provided subsidies to villagers to decorate their houses in the tradition of the Tujia ethnic minority (see Supporting Information Photo S1 in the online publication). They could also travel to the county town for business. 18 In summary, since the survey in 2003, national reforms, regional economic growth and the expansion of local transport connections have reconfigured the opportunities and challenges for livelihood reconstruction for the ruralto-rural households in this sample. When interpreting livelihood outcomes and strategies between 2003 and 2011 in the following sections, it is important to bear these contextual factors in mind. RURAL LIVELIHOOD OUTCOMES IN 2011 In 2011, the State Council publicly acknowledged problems threatening the viability of the displaced communities and released a post-construction working plan (CWRC, 2011). Issues identified included reduced arable farmland, environmental degradation, unemployment, low levels of education, the failure of small businesses and inadequate social security (Duan and Wilmsen, 2012). As the release of the working plan coincided with planning for the second survey, I anticipated that the impoverishment observed in 2003 would have persisted for the rural households. However, I also hoped that the unique commitment of the Chinese government to this resettlement, above all others, in conjunction with a growing regional economy and the broader national changes described above, might positively influence livelihoods. The following section describes the findings capturing four livelihood outcomes: i) increased income; ii) declining inequality; iii) improved food security; and iv) rising social well-being. 17. The mine closed in 2006 due to the tightening of environmental and safety regulations; about 80 workers were laid off, of which 12 to 15 were from resettled households. Among the miners, 20 were re-employed in the brick-making factory and the others left the village to find work. The closure of the mine meant the village could not afford to upgrade the brick factory. Adding to its decline was the slowing demand for bricks locally as the reconstruction of the region ended. In 2009, the brick factory was sold to private buyers without layoffs. 18. Supported by a government tax break, around 400 small businesses opened in Badong county mostly restaurants, hotels, services and transport businesses (Interview, Badong county government, Badong county town, 11 October 2012).

13 182 Brooke Wilmsen County Table 2. Average Household Income per Person (2003 prices) Average Household Income Per Person (RMB) Before Displacement Percentage increase (2003 and 2011) N Changjiang village Leijiaping village Per cent of Average Urban Household Income in Yichang city Changjiang village Leijiaping village Notes: N = total number of households. Sample includes only those households that responded to both 2003 and 2011 surveys, and those that filled out all three income entries. Source: Author s own survey. Increased Income Since 2003, incomes not only recovered but increased dramatically in both rural sites more than two-fold in Leijiaping and four-fold in Changjiang village (see Table 2). Of course, income growth has been high throughout China and Hubei province during this period. To account for this, Table 2 also reports income as a percentage of the average income of a Yichang urban household (Yichang is the closest and second largest city in Hubei province). It shows that income growth in Leijiaping and Changjiang villages between 2003 and 2011 is impressive even by Chinese standards. For example, in 2003, the average income in Changjiang village was only 25.8 per cent of the Yichang urban average, but by 2011, this figure had increased to 78.7 per cent. To understand the determinants of income in the three study periods, a regression analysis was carried out to explore the relationship between household income per person and various explanatory variables. Some of the explanatory variables capture the effect of particular livelihood strategies observed during fieldwork such as the importance of remittances sent home by household members, pensions and other subsidies provided by the government, crop diversity and the reliance on non-farm work. Other variables control for the broader determinants of income such as the village location (locational advantages and disadvantages), level of education (human capital) and age (life-cycle effects). As many of these variables are only relevant to farming households, the landless households are not included in the model. The results are presented in Table 3. Table 3 confirms that land size and quality were a crucial determinant of household income both prior to displacement and in The analysis, for example, predicts that income per person increases by 1,057 RMB for every extra mu of land per person possessed by the household in Put simply, land remains a key determinant of rural incomes. Interestingly, the land size and land quality variables in addition to many other explanatory

14 Land-based Resettlement in Rural China 183 Table 3. Summary of Regression Analysis (pre-displacement, 2003 and 2011) Pre-displacement Explanatory Variable Coefficient p Coefficient p Coefficient p Zigui dummy, 1 = yes ** Land holdings, *** *** mu/person Poor quality land ** (1 = yes) Highest educational attainment in households (omitted dummy: illiterate or primary level) Junior High Senior High Tertiary Person > 60 years in * household (1 = yes) Person < 15 years in household (1 = yes) Retired person in NA a NA a household (1 = yes) Person receiving state NA a NA a subsidies (1 = yes) Person receiving NA a NA a ** remittances in household (1 = yes) Orange main crop ** (1 = yes) 2 or more crops * (omitted: 0 or 1 crop) Non-farming jobs in ** *** household (1 = yes) Constant ** Number of observations R-squared Mean household income pp (RMB) Notes: *, ** and *** denotes the estimated co-efficient is significantly different from zero at the α = 0.1, 0.05 and 0.01 levels, respectively. Non-respondents in 2011 removed from earlier samples and households without land excluded. a Insufficient observations. Source: Author s own survey. variables are not significant immediately after displacement (as shown in Column 2 of Table 3). In 2003, households were in a state of chaos as occupants settled into new homes, established new income streams, drew on loans to cover daily expenses and rebuilding costs, and waited for young trees to begin producing fruit. As such, it is not unexpected that the usual relationship between household income and the determinants of rural income were temporarily disrupted.

15 184 Brooke Wilmsen Figure 2. Mean Quartile Income as a Percentage of Top Quartile (pre-displacement, 2003 and 2011) Source: Author s own survey. Table 3 also depicts other important relationships. First, before and just after displacement, very few households relied on state subsidies or family remittances as a source of income. As a result, these variables were omitted from the regression analysis for the pre-resettlement and 2003 periods. By 2011, however, many more households were relying on income from state subsidies and family remittances. Receiving remittances, in particular, had a significant and positive effect on predicted household income; a household receiving remittances is predicted to have an income 3,439 RMB higher than a similar household that is not receiving remittances. Second, in 2011, producing two or more crops had a statistically significant and negative impact on predicted income; the income per person of a household producing two or more crops is predicted to be 2,770 RMB lower than a household producing zero or one crop. The relationship between crop diversity and income is statistically insignificant in earlier periods. Third, the regression analysis reveals that non-farm work is associated with higher predicted household incomes both prior to displacement and in 2011 (but not in 2003). These statistical relationships are indicative of the kinds of livelihood strategies employed by the rural households, which are explored in more detail in the next section. Declining Income Inequality Income growth is remarkable in the sample sites, but often (and particularly in China) income growth can coincide with rising income inequality. Figure 2 explores whether the reported income gains were captured by a fortunate few or shared more evenly amongst the surveyed villagers. It ranks household incomes into quartiles before displacement, in 2003 and in As such, a household that was in the bottom quartile prior to displacement may appear in another quartile in 2003 or The results demonstrate that overall income inequality measured by comparing the incomes of the bottom

16 Land-based Resettlement in Rural China 185 Figure 3. Percentage of Households Reporting Insufficient Food Sometimes (pre-displacement, 2003 and 2011) Source: Author s own survey. quartiles to the top quartile decreases moderately by This trend is stronger in the village of Leijiaping (Badong county) than in Changjiang (Zigui county). Food Security Rural people rely on land to produce food both for sale and consumption. As such, changes in land size and quality and associated farming techniques can affect the capacity of households to feed themselves. The surveys conducted in 2003 and 2011 explored food insecurity by asking householders if they had enough to eat considering their production and income. These results are summarized in Figure 3 and suggest that the proportion of households reporting that food is sometimes insufficient increased dramatically after displacement. By 2011, these figures had declined significantly, but they still remained higher than pre-displacement levels. Social Well-being Income growth is an important, but imperfect, proxy for improvements in social well-being. An assessment of overall well-being is therefore crucial to understanding the changes in livelihoods. Noting the limitations of subjective measures, this research uses a series of social feelings statements that were noted from the respondents in the 2003 and 2011 surveys and then aggregated into a social well-being index. The respondents were asked to agree, disagree or remain neutral to a range of statements spanning so-

17 186 Brooke Wilmsen Figure 4. Social Well-being Index, Source: Author s own survey. cial, political, economic and environmental factors. Positive responses were coded +1, negative responses 1 and neutral responses 0. Figure 4 displays the changes in the resulting social well-being index across the two villages between 2003 and It shows that subjective well-being has increased in both samples. 19 Leijiaping residents on the whole have higher subjective well-being than those in Changjiang village. Interestingly, Leijiaping village and Changjiang village experienced similar rates of increase in the social well-being index between 2003 and The findings reported in this section are surprising. The 2003 study found widespread impoverishment in the surveyed villages, and suggested that the rural households were not benefiting from broader economic developments. However, eight years on, the gains are great. Incomes have grown dramatically and the households of Changjiang and Leijiaping villages are gaining ground on their more prosperous urban counterparts in nearby Yichang city. People are not only richer, but report improved social well-being. There are also moderate improvements in inequality and food insecurity since Overall, these indicators suggest livelihoods have improved. 19. The first survey, conducted in 2003 in the villages of Leijiaping and Changjiang, did not ask the households to reflect on their social well-being before displacement. There were concerns that post-displacement social well-being would influence accurate memory recall. Therefore, although social well-being improved between 2003 and 2011, no inference can be made relative to pre-displacement social well-being.

18 Land-based Resettlement in Rural China 187 Table 4. Average Land Holdings (pre-displacement, 2003 and 2011) Pre-displacement Average land holding (mu per household) Changjiang a Leijiaping Average land holding (mu per person) Changjiang Leijiaping Note: a there is one large family farm of 50 mu pushing up aggregate land holdings. If removed, average land holdings between 2003 and 2011 are similar at 3.1 mu per household and 1.1 per capita in Source: Author s own survey. LIVELIHOOD STRATEGIES The livelihood effects described in the previous section are a result of the strategies applied by the rural households in response to the broad forces permanently reshaping the TGRA. By analysing these responses, four broad livelihood strategies were identified, each related to the land: land consolidation, agricultural specialization, livelihood diversification and migration. These are discussed in the section that follows. Land Consolidation Despite further inundation since 2003, the majority of households in Leijiaping village (Badong county) and Changjiang village (Zigui county) still hold farmland. Of the 118 households captured in both the 2003 and 2011 surveys, 12 and 18 households are landless, respectively. As such, access to land remains a crucial determinant of livelihood outcomes for the sample population. Average land holdings pre-displacement, after displacement and in 2011 are presented in Table 4. Contrary to expectation, Table 4 shows no aggregate decline in the average farm size between 2003 and 2011, rather a small increase (noting the effect of one large family farm that is discussed later). Moreover, 31 per cent and 39 per cent of households in Changjiang and Leijiaping villages, respectively, increased their land holdings by 10 per cent or more between 2003 and This is surprising as the village committee said there was no formal reallocation of farmland. 20 However, interviews with households in both villages revealed that informal redistribution was widespread after Often informal rental arrangements were negotiated between those who migrated out of the village for work and those who remained on the land. Households also reclaimed uncultivated land or 20. Interview, village secretary, Changjiang village, 23 September 2012.

19 188 Brooke Wilmsen Figure 5. Breakdown of Landholdings by Farm Size (mu) (pre-displacement, 2003 and 2011) Source: Author s own survey. wasteland at higher elevations (which is illegal under the NFCP that was introduced in 1998). To drive rural economic growth, the Chinese government encourages land consolidation throughout rural China. The trend towards larger, specialized farms is observed at Changjiang village, with the emergence of two large farms amongst the sample group, but not at Leijiaping village. In stark contrast to 2003, larger farms in Changjiang village make up nearly half of the total land held by the sample group (see Figure 5). This is not the case in Leijiaping, however, where the composition of land holdings in 2011 is similar to 2003, with farms of 2.5 to 5 mu continuing to make up around half of total farmland. Furthermore, the analysis in Table 3 shows that in 2011, household income grew for every additional mu of land per person. These data suggest that land consolidation is a livelihood strategy employed by a few select households in Changjiang to increase incomes, but not in Leijiaping. Agricultural Specialization The TGP submerged the most fertile land, making the efficient use of the remaining arable land even more vital to rebuilding and improving livelihoods. Oranges are the key agricultural product in the TGRA. Citrus production has a long history in the area. However, the increasing specialization in citrus for commercial purposes is a more recent livelihood strategy, and one explicitly

20 Land-based Resettlement in Rural China 189 Figure 6. Crop Types Source: Author s own survey. Table 5. Number of Crops per Household (percentage of households engaged in farming) No of Crops Village Pre-displacement Changjiang Leijiaping Changjiang Leijiaping or more Changjiang Leijiaping Source: Author s own survey. supported by the third stage of the Chinese government s development plan for the TGRA. Under this plan, farmers are encouraged to plant new orange varieties that are suited to the region s climate and aligned with market demand. After 2003, following this advice, farmers gradually replaced navel oranges with the Valencia, peach and summer varieties. Oranges remain the dominant crop for households in Changjiang and Leijiaping villages, and the proportion of land dedicated to oranges has increased at both sites. This can be seen in Figure 6, which depicts the proportion of land dedicated to different crop types. Corresponding with intensified specialization in oranges is a decline in crop diversity as shown in Table 5. This combination suggests that agricultural specialization is an important livelihood strategy for farmers who are seeking to maximize the production value of their land.

21 190 Brooke Wilmsen Table 6. Farm versus Non-farm Sources of Income (percentage of total reported income sources) Pre-displacement Work Type N = 61 N = 58 N = 93 Changjiang village Farm Non-farm Pensions, remittances and other state subsidies Unemployed Pre-displacement Work Type N = 132 N = 135 N = 206 Leijiaping village Farm Non-farm Pensions, remittances and other state subsidies Unemployed Note: Number of responses (N) increased because respondents cited more than one source of income. Source: Author s own survey. Agricultural specialization is clearly important to income, as shown in Table 3. In 2011, households that produced two or more crops earned significantly less than those that produced zero or one crop (after controlling for other determinants of income). Increased crop specialization, in contrast, does not significantly increase predicted household income in the predisplacement and 2003 samples, suggesting that the rural economy in the area is transforming in such a way as to reward farmers who specialize in particular crops, namely oranges. Livelihood Diversification The growth of the regional economy increased the supply of non-farm jobs, while infrastructure improvements, particularly in Badong, improved access to non-farm work in nearby urban areas. Table 6 shows that income from farming declined considerably between 2003 and 2011 as a proportion of total income sources. This shift coincides with increased reliance on nonfarm work and transfer payments such as pensions, remittances and other state subsidies in Leijiaping village, and an increased reliance on transfer payments in Changjiang village between 2003 and The type of non-farm work that rural households are engaged in is also changing. As anticipated, Figure 7 shows that the completion of infrastructure construction associated with the TGP saw a decline in manual labour, particularly in Leijiaping village. However, the opening of the new bridge increased opportunities for trade and tourism. This is reflected in an increased proportion of non-farm income derived from small business activities (4 per cent in 2003 compared to 21 per cent in 2011). Billions of RMB have been invested in Zigui county town to build its logistics hub and SEZs, but this

22 Land-based Resettlement in Rural China 191 Figure 7. Non-farm Income Sources (pre-displacement, 2003 and 2011) Source: Author s own survey. is not the case in Changjiang village. We do not see a similar increase in small business, which is largely due to the village s ongoing geographic isolation. In fact, after 2003, non-farm work at Changjiang village has declined as a proportion of total income sources in favour of remittances, welfare payments and other state subsidies (see Table 6). Migration The considerable increase in remittances by 2011 suggests that migration of household members is an important livelihood strategy. The Badong and Zigui Resettlement Bureaus 21 said that the most important reason that rural people migrate out of the area is land shortages. The Zigui Resettlement Bureau estimates that 30 per cent of affected Zigui residents moved out of the region. An interview 22 with members of the Changjiang village committee revealed that by 2012, 350 people had moved out of the village permanently, mostly young couples who were directly affected by the TGP. Migrants retain rights to land but often allow friends or family members to farm it. There are a few rare cases in Leijiaping where land is permanently relinquished through an informal purchase arrangement. These findings are also supported by the in-depth interviews in both villages in Of the 14 households interviewed in Changjiang and 21. Interviews conducted with Badong Resettlement Bureau, Badong county town, 10 October 2012; Zigui Resettlement Bureau, Zigui county town, 12 September Conducted in Changjiang village, 23 September 2012.