Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interpreting gentrification in China : the rising consumer society and inequality in the state-facilitated… Yang, Qinran 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_may_yang_qinran.pdf [ 49.15MB ]
Metadata
JSON: 24-1.0343650.json
JSON-LD: 24-1.0343650-ld.json
RDF/XML (Pretty): 24-1.0343650-rdf.xml
RDF/JSON: 24-1.0343650-rdf.json
Turtle: 24-1.0343650-turtle.txt
N-Triples: 24-1.0343650-rdf-ntriples.txt
Original Record: 24-1.0343650-source.json
Full Text
24-1.0343650-fulltext.txt
Citation
24-1.0343650.ris

Full Text

INTERPRETING GENTRIFICATION IN CHINA: THE RISING CONSUMER SOCIETY AND INEQUALITY IN THE STATE-FACILITATED REDEVELOPMENT OF THE CENTRAL CITY OF CHENGDU  by Qinran Yang B.Eng., The University of Chongqing, China, 2009 M.Eng., The University of Chongqing, China, 2012   A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Geography)     THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2017     © Qinran Yang, 2017 ! ii!Abstract  Currently, scholars are debating the epistemological limits of the concept of gentrification as a representation of global urban experiences. The thesis addresses this global debate from the perspective of Chinese urbanisation. In China, socio-spatial upgrading and displacement, which normally define a gentrification process, are most likely prompted by state-facilitated urban redevelopment. The dissertation questions why and how state action attracts middle-class newcomers to the inner city and penalises or reconciles working-class residents. With the research focused on urban China, the thesis also contributes to conceptual and methodological issues on gentrification research on a global scale. A meso-level approach is taken to trace the gentrification process through both structural tendencies and grounded empirical processes in the inner city of Chengdu from 2000 to 2010. Analysis at the two scales in conjunction accomplishes an examination of the existence of gentrification and the explanation of regularities within it. In Chengdu, the state’s mobilisation in developmental strategies concentrating on the built environment allows the case study to make new claims for gentrification knowledge for non-Western cities. Mixed methods, including statistical and spatial analysis, institutional analysis, and extensive ethnographic study, are used to investigate gentrification from a structural perspective, a historical perspective, and as a grounded process within the neighbourhood.  The research reveals that state actions in urban redevelopment direct the cultural and behavioural changes of the middle-class newcomers, so that they are compatible with state strategies in modernisation and real estate boosterism. Working-class groups face varying outcomes in residential relocation. Overall, the process reflects state hegemony over societies that comprise subaltern cultures in the city, incorporating legal and propertied citizens into the frame of consumerism while disenfranchising rural-urban migrants. The study unravels how state domination in urban redevelopment drives social change towards a consumer society, which sharpens social inequities but also, ironically, rebuilds the collectivist ideology from one centring on production to one pressing the ideology and practices of consumption. So an overall gentrification process is formulated in Chengdu, which retains complexities and contingencies in localities. On a broader scale, the thesis urges a meso-level approach to gentrification research in other cities.  ! iii!Preface  This dissertation is an original intellectual product of the author, Qinran Yang. The fieldwork reported in this dissertation was covered by UBC Ethics Certificate number H14-02952.  Parts of the findings from Chapter 5 have been published in China Urban Studies: Vol. 9 (Y. Ning (Ed.), Beijing: China Science Publishing, ISBN: 9787030503848) as the sole-authored article entitled “State-led gentrification?: Inner-city redevelopment and social change in the post-reform cities of China (pp.197-220).”                      !! iv!Table of Contents Abstract ............................................................................................................................... ii Preface ................................................................................................................................ iii Table of Contents ............................................................................................................... iv List of Tables ..................................................................................................................... vi List of Figures ................................................................................................................... vii List of Abbreviations ......................................................................................................... ix List of Characters ................................................................................................................ x Acknowledgements ............................................................................................................ xi Chapter 1 Introduction ........................................................................................................ 1 1.1 The globalisation of gentrification? ....................................................................................... 1 1.2 The scene in China ................................................................................................................. 1 1.3 Research questions ................................................................................................................. 3 1.4 A meso-level approach to gentrification ................................................................................ 6 1.5 The state, institutional change and social change in gentrification in China ......................... 9 1.6 The case study city ............................................................................................................... 11 1.7 Methodology ........................................................................................................................ 14 1.7.1 Fieldwork in Chengdu ................................................................................................... 16 1.8 Thesis overview .................................................................................................................... 25 Chapter 2 Gentrification Studies on a Global Scale ......................................................... 28 2.1 Ontological basis of gentrification ....................................................................................... 29 2.1.1 The essence of gentrification ........................................................................................ 29 2.2 Epistemological problems of gentrification on a global scale ............................................. 37 2.2.1 A Global theory for gentrification? ............................................................................... 38 2.2.2 The epistemological gap between the North and South ................................................ 43 2.3 Gentrification studies in China ............................................................................................. 51 2.3.1 State and institutional transition .................................................................................... 51 2.3.2 The new middle class in a transitional society .............................................................. 54 2.3.3 The rules of residential relocation and the working class ............................................. 56 2.4 Conclusion: A meso-level approach to state-facilitated gentrification in Chengdu ............. 60 Chapter 3 Grounding Gentrification in the Large Chinese City ....................................... 62 3.1 Social transformation in large Chinese cities ....................................................................... 62 3.1.1 Socioeconomic restructuring in the new economy ....................................................... 63 3.1.2 Individual consumption and the middle-class formation in the transitional society ..... 72 3.2 The geography of gentrification in Chengdu, 2000-2010 .................................................... 78 3.2.1 The extent of gentrification ........................................................................................... 81 3.2.2 The spatial manifestation of gentrification ................................................................... 86 3.3 Correlations between gentrification and social and spatial transformation ......................... 89 3.3.1 Industrial transformation in inner city .......................................................................... 92 3.3.2 Housing privatisation .................................................................................................... 96 3.3.3 The new urbanism ......................................................................................................... 98 3.4 Conclusion: Gentrification as a mechanism for economic development and social change .................................................................................................................................................. 101 ! v!Chapter 4 Inner-city Urbanism and the Construction of Consumer Citizenship ............ 104 4.1 New landscapes and new modernity .................................................................................. 104 4.1.1 Anti-urbanism and de-aestheticisation ........................................................................ 106 4.1.2 Defining the image of a global city ............................................................................. 112 4.1.3 Towards cosmopolitanism? ........................................................................................ 119 4.2 Spatial commodification and housing consumption .......................................................... 126 4.2.1 Two levels of land markets ......................................................................................... 127 4.2.2 State intervention and spatial commodification .......................................................... 132 4.2.3 Towards a consumer-driven economy? ...................................................................... 138 4.3 Inner-city urbanism and Chinese consumer citizens .......................................................... 143 4.3.1 Socioeconomic attributes of gentrifiers ...................................................................... 143 4.3.2 Place-based identity of inner-city urbanites ................................................................ 147 4.3.3 Emancipation or reproduction? ................................................................................... 155 4.4 Conclusion: Elite-oriented place-making, middle-class consumer culture and the development of consumer citizenship .................................................................................. 159 Chapter 5 Residential Relocation and the Working Class in Gentrification .................. 163 5.1 The social and tenure diversity in the old inner-city neighbourhoods ............................... 163 5.2 Homeownership and upward social mobility? ................................................................... 165 5.2.1 Consumerism and pro-homeownership reform .......................................................... 167 5.2.2 Transforming to a propertied stratum? ....................................................................... 173 5.2.3 The uncertain game of upward mobility ..................................................................... 183 5.3 Cultural exclusion and social activism ............................................................................... 187 5.3.1 Property and cultural activism .................................................................................... 189 5.3.2 Claiming legitimacy for informality ........................................................................... 196 5.3.3 Mass mobilisation and consensus building ................................................................. 200 5.4 Marginalisation and outcasts .............................................................................................. 208 5.4.1 The silent migrants ...................................................................................................... 208 5.4.2 Losing spaces for livelihoods ...................................................................................... 211 5.4.3 Homeownership and urban identity ............................................................................ 214 5.5 Conclusion: Hegemonic power and socialisation and disenfranchisement ........................ 219 Chapter 6 Structural Inequalities in the post-Gentrification Housing Market ................ 223 6.1 Displacement and concentrated poverty in the inner city .................................................. 224 6.2 Speculation and the affordability of the city ...................................................................... 234 6.3 Tenure-based exclusion and segregation ............................................................................ 240 6.4 Conclusion: Unequal rights to the city in a consumer society ........................................... 251 Chapter 7 Conclusion ...................................................................................................... 255 7.1 Gentrification in Chengdu: New urbanism and a new social order ................................... 255 7.1.1 Elite-oriented place-making, consumer culture and consumer citizenship ................. 255 7.1.2 State hegemony, socialisation and disenfranchisement .............................................. 259 7.1.3 New urbanism and a new social order ........................................................................ 261 7.2 Returning to the concept and research practice .................................................................. 264 7.3 Reflections on state-facilitated gentrification ..................................................................... 267 7.4 Further research .................................................................................................................. 270 Bibliography ................................................................................................................... 273 Appendix A An overview of informants ........................................................................ 303 Appendix B Interview Guide .......................................................................................... 305 Appendix C Consent Form ............................................................................................. 309 ! vi! List of Tables  !Table 1.1 The profiles of three redevelopment projects in Chengdu ................................ 20 Table 1.2 The locations of informants .............................................................................. 23 Table 3.1 Employment structures by sector in Chengdu, 1990, 2000 and 2010 .............. 67 Table 3.2 Change in the employment structure of Chengdu by occupation 1990-2010 ... 71 Table 3.3 The socioeconomic profiles of inner-city sub-districts in Chengdu ................. 82 Table 3.4 Simple correlations for 2000, 2010 and 2000-2010 change against the gentrification index in Chengdu (N=31) ................................................................... 91 Table 4.1 Socioeconomic profile of the most and least gentrified sub-districts ............. 144 Table 5.1 The experiences of three working-class groups in gentrification ................... 219 Table 6.1 Changes in population by education and occupation in Chengdu 2000-2010 226 Table 6.2 Simple correlations for 2010 and the 2000-2010 change against the change in LQ for the working class in Chengdu (N = 31) ...................................................... 231 Table 6.3 Changes in households by tenure type in Chengdu 2000-2010 ...................... 242        ! vii!!List of Figures Figure 1.1 Conceptual map to state-facilitated gentrification in China. ............................. 8 Figure 1.2 The city of Chengdu in China. ........................................................................ 12 Figure 1.3 Administrative divisions of Chengdu metropolitan region. ............................ 12 Figure 1.4 The inner city and the five historical urban districts of Chengdu. .................. 13 Figure 1.5 Field sites: old danwei compound and displacees’ area (CJA and Workers’ Village). .................................................................................................................... 17 Figure 1.6 Field sites: commercialised historic neighbourhoods (WNA). ....................... 18 Figure 1.7 Field sites: on-site and off-site resettlement neighbourhoods (JW and Quanshui Renjia). ..................................................................................................... 18 Figure 1.8 Field sites: gentrified neighbourhoods (Times Riverside and Waterfront). .... 19 Figure 1.9 Locations of main field sites in Chengdu. ....................................................... 20 Figure 3.1 Employment trend by sector in Shanghai 1990-2010. .................................... 65 Figure 3.2 Employment trend by sector in Chengdu 1990-2010. ..................................... 65 Figure 3.3 Employment trends by occupation in Shanghai 1990-2010. ........................... 69 Figure 3.4 Employment trends by occupation in Chengdu 1990-2010. ........................... 69 Figure 3.5 Changing social status of the population in the inner city of Chengdu, 2000-2010........................................................................................................................... 87 Figure 4.1 The socialist and newly built landscape in Chengdu. .................................... 107 Figure 4.2 Workers’ neighbourhoods in Chengdu. ......................................................... 107 Figure 4.3 Chengdu in Qin Dynasty. .............................................................................. 109 Figure 4.4 The central plaza of Chengdu in 1985 (above) and 2016 (below). ............... 109 Figure 4.5 New residential community in Chengdu in the early 1990s. ......................... 111 Figure 4.6 Suburban gated communities (International Community Foothills) in Chengdu.................................................................................................................................. 117 Figure 4.7 Historic site (Wide and Narrow Alley) after commercial renovation. .......... 119 Figure 4.8 The International Finance Square in Chengdu. ............................................. 121 Figure 4.9 The newly built offices and condominiums in the new financial centre. ...... 122 Figure 4.10 The renovation process of the Daci Temple area, 2004, 2009 (above), 2014 (below). ................................................................................................................... 123 Figure 4.11 Percent increases in housing supply year-over-year in Chengdu. ............... 139 Figure 4.12 Gentrified landscape (Chengdunese Paradise and Times Riverside) in the 2000s. ...................................................................................................................... 148 Figure 4.13 High-end communities newly built in the current wave of gentrification. .. 153 Figure 4.14 The landscape of retail gentrification in old neighbourhoods in the inner city.................................................................................................................................. 156 Figure 5.1 The types of buildings in old neighbourhoods .............................................. 164 Figure 5.2 Redeveloped communities and resettlement communities ............................ 178 Figure 5.3 Wide and Narrow Alley before (2004) and after commercial redevelopment (2016). ..................................................................................................................... 192 Figure 5.4 Banners resisting nail-households in the CJA area (2013) and banners resisting demolition in the WNA area (2004). ...................................................................... 203 Figure 5.5 Agricultural market in Caojia Alley before demolition. ............................... 212 Figure 6.1 Change in the working-class location quotient in Chengdu, 2000-2010. ...... 229 ! viii!Figure 6.2 Trend in land transactions, land tax revenue and unsold housing stock in Chengdu. ................................................................................................................. 235 Figure 6.3 Housing price change in Chengdu, 1997-2014. ............................................ 237 Figure 6.4 Average disposable income per capita in Chengdu. ...................................... 238 Figure 6.5 Price-to-income ratio in Chengdu. ................................................................ 239 Figure 6.6 The distribution of households by tenure type in Chengdu in 2000, 2010 .... 246           !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ix!!List of Abbreviations  BLRC Bureau of Land and Resource of Chengdu BURHC Bureau of Urban-Rural Housing of Chengdu CBS Chengdu Bureau of Statistics CBRC China Banking Regulatory Commission CCP Chinese Communist Party  CDB China Development Bank CIA China Index Academy MHURD Ministry of Housing and Urban-Rural Development MLR Ministry of Land and Resources  MS Ministry of Supervision NBSPRC National Bureau of Statistics of the People's Republic of China PBC People's Bank of China PCOC Population Census Office of Chengdu  PCOS Population Census Office of Shanghai PCOSC Population Census Office of the State Council  PCOSP Population Census Office of Sichuan Province !!!!!!!!!!!!!!!!!!!!!!!!!!!! x!!List of Characters    Banqian % Canquan zhihuan   Chaiqian  Chaiqianhu  Chaiqianfu % Chaierdai  Danwei   Datong  Dengjia zhihuan   Dingxiang zhengshou  Dingzihu ' Fangwu chaiqian  Fangwu zhengshou  Hukou   Hutong  Jiedao  Jiedao banshichu  Jumin zizhi gaizao  Lilong & Moni banqian xieyi %# Moni chaiqian xieyi %# Penghuqu  Qunzhong ! Renmin  Shuzi $ Tongzilou  Xiaokang  Xiaokang shehui  Youcanjieceng ( Zai 	 Zengshou  Zuojia buchang "	 !!!!!! xi!Acknowledgements I am deeply indebted to my supervisor, Professor David Ley, who has been supportive of the research and worked actively to help me in pursuing the research goals. I have benefited from his insights into gentrification and the way he builds knowledge for Asian cities, which guided me to understand and develop ideas for the urban and social change in China. Professor David Ley is patient and extraordinarily generous. The work would not have been possible without his assistance in the writing process. His critical reading and guidance in the refinement of ideas have enabled me to achieve higher quality. I cannot express enough thanks to my committee for reviewing the entire manuscript and for the excellent advice: Professor Amy Hanser, Professor David Edgington, Professor Elvin Wyly, Professor George C.S. Lin, Professor Hyun Bang Shin, and Professor Jim Glassman. I offer my sincere appreciation for the learning opportunities provided by my committee.   The completion of this project owes a good deal to a large number of interviewees, both urban dwellers and specialists from Chengdu. I thank my informants for sharing experiences and opinions with me. I have enjoyed and learnt from our conversations. I am also grateful to the community and institutional members who introduced me to the interviewees and aided me in collecting materials during my fieldwork in Chengdu. In particular, I feel immense gratitude to Professor Wei Zhao from the Southwest Transportation University, Benying Xu from the Chengdu Institute of Planning and Design, Bing Yang from the China Urban Planning and Design Institute and Ming Zhou from the Southwest University for Nationalities, for establishing the connection between me and local communities and institutions and organising focus groups and discussions.   The research would not be accomplished without the financial support of the Doctoral Scholarship of the UBC-China Scholarship Council and the MITACS Globalink Research Award.   My debt of gratitude is to my families and friends who keep me in the pursuit of an academic career. My parents, who are always proud of me, made me brave when the times got tough. Benying Xu, the partner in my life, has protected my dream with his love. Professor Heping Li, who first encouraged me to study abroad, provides unending trust and support to me. My heartfelt thanks. Finally, I would like to thank my friends in St. John’s College, who motivated me with passion in research, offered valuable opinions by personal experience and filled my life with laughter in the past four years. ! 1!Chapter 1 Introduction 1.1 The globalisation of gentrification?   Approximately four years ago when my doctoral program was laid out, gentrification studies in the Global South were only beginning. More than a few authors have hesitated to examine the globalised properties of the gentrification process. The concept adds powerfully to the criticism of class-related urban change and social inequality. Nevertheless, the term also appears to be distant and alien to the political and societal settings in those countries. Within only four years, however, gentrification studies have witnessed a significant amount of labour migration from the Global South, especially driven by the urban opportunities of the metropolitan areas. Nevertheless, confusion about the essence of this expansive concept has given authors wide discretion in defining and explaining the process, which in turn attracted critics concerning either the overgeneralisation or the provincialism of knowledge production under the framework of gentrification.  Recently, three of the leading figures, Lees, Shin and Lopez-Morales, have contributed to a substantial advance in studies on global gentrification. The authors ran two workshops, followed by two books (Lees et al., 2015, 2016) and two special issues of journals (López-Morales et al., 2016; Shin et al., 2016) on gentrification from the perspective of comparative study. This global discussion of gentrification has formally opened a debate on the conceptual and methodological issues for gentrification to be explored on a global scale.  1.2 The scene in China The scene in China exemplifies the way that the gentrification process can be varied with context. Since the implementation of the policy of “economic openness,” the Chinese government has shifted its political ideology from a grassroots to a middle-class political system. Deng Xiaoping offered a promise to the Chinese by proposing his social strategy under the slogan of “get rich first and achieve common prosperity later” in post-! 2!reform China (Fan, 2006). To give substance to the Chinese dream, from that point on, the Chinese state has maintained a continuous commitment to urban modernisation, upon which the rounds of urban redevelopment programmes have unfolded in large Chinese cities. During the last three decades, the landscape in China’s large cities has been greatly transformed, characterised conspicuously by a proliferation of master-planned gated communities, whether sprawling suburban luxury communities or more compact central redevelopment. These newly built neighbourhoods are homogenous, approaching a form of “Chinese modern,” with light coloured buildings, signs of Western architecture, a central green garden and playgrounds and barriers (see Pow, 2009). Apartments in these neighbourhoods are often self-owned, representing a private lifestyle in contrast with the tensely communal living in the work-unit compound (danwei,) in the socialist era that compressed one’s working, social and family activities.  Urban redevelopment and modernisation accompany an awakening of the consumer consciousness of China’s new rich, distancing them from the communal society in the socialist era. The Chinese new rich are the advance force and also the beneficiaries of the expanding market economy and new emerging industries in the tide of economic transition. Members of this group often fall under the media spotlight for their purchasing power, providing a showcase of the social and cultural transformation of contemporary Chinese society. Rather than differentiating themselves from modernist values, China’s new rich, who could be gentrifiers, are apparently in a race to pursue the new Chinese modernity through purchasing commodity housing and resettling in the newly built gated communities. In the past two decades, China’s cities are witnessing a soaring homeownership rate evolving from the public housing base of socialist China. A household finance survey from a Chinese university estimated that the homeownership rate in the urban areas of China reached 85.39% in 2012 (Gan et al., 2013).  Massive urban redevelopment has caused the destructive geographical reconfiguration of society, where the residents previously living in the redeveloped neighbourhoods have to face different arrangements on residential relocation, which in Chinese is called chaiqian (%). According to various compensation rules formulated by governments at both central and local levels, the circumstances of affected residents in residential relocation are not necessarily identical. A public tenant currently residing in ! 3!collective housing provided by his/her employer might exchange his/her old property for a high-rise apartment of improved quality. For this public tenant, the new apartment may introduce greater or less financial gains, while unfolding a new lifestyle in the modernised city.  However, as an immediate result of housing demolition and dislocation, a rural-urban migrant might be forced out of his/her current residence. For this migrant, urban redevelopment is an overt process of eviction by a new marketised regime of urban construction devoting itself to the overriding enterprise of economic development and urban modernisation. In Beijing, Hsing (2010) noted that there were more than 500, 000      “evicted households whose homes had been demolished (chaiqianhu, )” from 1990 to 2004 (p. 72). However, it must be noted, rather than a substantial retreat of the working class from the city, the large Chinese cities have continued to absorb low-paid migrant workers via rapid urbanisation and industrialisation. Based on the National Bureau of Statistics, the total number of rural-urban migrant workers in China reached 136 million in 2007 and accounted for 46.5% of total urban employment (Cai et al., 2009). Increasingly, low-quality settlements in urban peripheries have accommodated these migrants, who are seeking job opportunities in the city. Additionally, in inner cities it is easy to observe hybrid urbanism, comprising imperial Chinese dwellings and religious buildings, austere work-unit compounds, and the spectacular, modernist high-rise towers (see also Lin, 2007; Ma & Wu, 2005). The income inequalities measured by the Gini coefficient in China increased from less than 0.3 in mid-1980 to more than 0.45 in the latter half of the 2000s (Baffes et al., 2008; Li et al., 2013). The egalitarianism of socialist society has passed out of existence, and China has become a society with one of the widest gaps between rich and poor.  1.3 Research questions  Briefly, the above scenes sketch out the landscape and social change in urban redevelopment in post-socialist China. Instead of post-industrial transformation, gentrification in China appears in a time of profound urban modernisation and post-socialist social change, featuring in the elevation of personal consumption attending ! 4!economic liberation. The central and local states undoubtedly play a bold role throughout the whole process. Not merely, a complex scenario for middle- and working- class dynamics has been woven into the physical fabric of change and social geographical reconfiguration. These scenes do not only suggest contingencies in the transition to the eventual appearance of gentrification in Western society. They imply the fundamentally different logic of urban and social change and the distinctive meanings of the gentrification process for those involved in it. Specifically, the situations require a gentrification explanation in China that scrutinises state actions and unpacks social class dynamics in the sequential process of socio-spatial upgrading and displacement. Starting with this preliminary understanding, four questions guide this research. The former three questions aim to generalise an understanding of gentrification in a Chinese city. The last question returns to reflections on gentrification research on a global scale.  • Why and how do state actions in spatial production interact with the social dynamics of the Chinese middle class, leading to the process of socio-spatial upgrading in the inner city of Chengdu?  • Why and how do state actions in residential relocation and compensation penalise or alternatively, satisfy the current residents, while ending in the process of displacement in Chengdu?  • To what extent are the two processes of redevelopment and displacement expressive of a mode of gentrification in Chengdu, or do they negate the existence of gentrification in the city?  • If it is the former, how does gentrification in Chengdu speak to research practices on a global scale?  For answering these questions, three concepts must be clarified at the outset. The study examines state-facilitated gentrification in the inner cities of China’s metropolises. Before approaching the methodology and field sites, three components of this research subject need to be defined. First, geographically, the study is confined to inner-city gentrification. Lees et al. (2016) proposed that the emergence of multiple investment sites for urbanisation mean that the inner-city centrality has lost its specificity in gentrification studies. I do not absolutely reject a conceptual extension of the residence beyond the ! 5!inner city. However, based on my understanding, class reconfiguration and displacement is a more prominent urban trend in built-up areas characterised by a long-lasting, stable social structure and institutional establishments. In the turbulent rural-urban interface characterising many metropolitan regions in the Global South, capital reinvestment in former villages often involves substantial changes regarding, for example, administrative rescaling, citizenship and subjectivity, industrial and employment structures, and tenure and stakeholder structures. The meaning of gentrification could be obscured in light of these tremendous changes (see Wu, 2016). This study thus maintains consistency with convention by focusing on inner-city neighbourhoods.  Further, this study addresses state-facilitated gentrification, in contrast with individually-driven gentrification. It refers to a process involving state intervention through either initiating, sponsoring, regularising or directly leading projects. The means of intervention can be diverse, with variations across programmes and cities. In China, state interventionism in urban redevelopment is rooted in public land ownership and the large proportion of publicly or collectively owned housing in the inner city. In contrast to individual-driven gentrification, the state mainly exerts its influence through two forms; both will be examined in this thesis. First, it exerts influence in the form of a bureaucratic entity: state actors at the central and local level direct project implementation by generating policies, spatial plans, and institutional change. Second, in a more abstract form, the nation-state system can also impact the pathway to gentrification via a cultural and ideological influence on society (e.g., the cultural ideology of consumption, consensus building around demolition and relocation). This type of impact has been mostly expressed through state-market-society relations in China. The nation-state, as defined by Giddens (1985), is a “bordered power-container”; among other processes, the development of the nation-state “involves processes of urban transformation and the internal pacification of states” (p.120).  Finally, it is also necessary to clarify in this research how the concept of class has been treated. In 2005, Butler asked for a middle-range theorisation for gentrification that would examine the gentrifiers’ construction of local identity as a way of responding to global trends. This proposal is associated with Butler’s approach to the argument that class now has a weak sense, insofar as gentrification is a process where individuals seek ! 6!habitus and “elective belonging” so as to link themselves with the global economy and urbanism (Savage et al. 2005, p. 207). In contrast to Butler, this study insists that gentrification studies should lay particular emphasis on relational and political class reconfiguration and conflicts in place. The concept of class, although loosely defined, retains the power to depict shared habitus, changing social relations and inequalities. Class in contemporary China is considered to be dynamic and emergent in this study; it is captured here by referring to social groups that exhibit the same tendency to acquire certain resources and forms of power (see also Tomba, 2004). The research purpose does not serve to theorise class. Rather than trivialising class relations in Chinese society as established and clearly defined, I aim to illustrate the dynamics of social classification and the tendency towards deepened social divisions in contemporary Chinese cities. I follow Bourdieu’s viewpoint that “the question with which all sociology ought to begin” is “that of the existence and mode of existence of collectives” (Bourdieu, 1991, p. 250). The boundaries of a class can only be understood by considering “social practices” and not through “theoretical conjecture” (Weininger, 2005, p. 85). Through the lens of gentrification, this research will stress the significant role of cultural capital in class formation and reclassification.  1.4 A meso-level approach to gentrification  The study contributes to a meso-level approach to explore gentrification in its context and outside of the “usual gentrification suspects” (Lees et al., 2015, p.1). The current literature has enriched the global nature of gentrification from either a macro- or a micro- approach. Reviewed in the next chapter, the former starts with global forces, such as the urban imaginary of transnational urbanism (e.g., Atkinson & Bridge, 2005; Herrera et al., 2007; Rubino, 2005) and the omnipresence of transnational capitalists (typically, Smith, 2002; Hackworth & Smith, 2001), but fails to clarify their relations with endogenous forces and factors. By contrast, a majority of current work has been more likely to conclude with findings of discrete differences among even individual neighbourhoods and projects, which render theoretical inference unsatisfactory. The differences emerge from, for example, the diversity of the modes of gentrification (e.g., ! 7!Lemanski, 2014; Ghertner, 2014), the contingencies of the process (e.g., Yip & Tran, 2015; Sýkora, 2005; Kovács et al., 2012; Maloutas, 2007, 2014), and the uncertainty around social outcomes for current residents affected by the process, based in particular urban conditions in the Global South (e.g., Doshi, 2005; Weinstein & Ren, 2009; Islam & Sakizlioglu, 2015; Badyina & Golubchikov, 2005).   Bringing the concept of gentrification to alternative regions outside Europe and North America, Lees et al. (2015; 2016) proactively suggested a relational comparative approach to gentrification processes globally, drawing on the scholarship of new comparative urbanism (see also Lees, 2012). The new comparativism is essentially a cosmopolitan epistemology of cities in the world, aiming at “a global scope of urban studies” (Robinson, 2014, p.1). It firmly avoids region-based theoretical production, but also opposes any type of hegemony of theoretical production based on certain sources while defining other sources as “exceptions, mere case studies, ‘facts’ or ‘data,’ to illuminate existing theories” (Robinson, 2011, p.17). Instead, the authors advocate theory building based on a concept sufficiently abstract to allow constant theoretical breakthroughs through comparative studies (Robinson, 2011; Roy, 2011). Also, scholarship should emphasise urbanism shaping and adjusting in the wider national political-economic system and the global and local processes that impinge the site (Mcfarlane, 2010; Nijman, 2007). Based on this understanding, Lees et al. (2016) deem the concept of capital switching from productive industries to the secondary circuit of accumulation “in forms of fixed capital and, more specifically, the built environment and urban space” (p. 45) to provide an underlying economics of gentrification shared between different spatial contexts. Starting with this concept, this relational comparative approach has an essential purpose to build a global theory for gentrification processes. However, contextualised local embeddedness will always enrich the theory, for example, the state interventionism in the real estate market in Eastern Asia (see also Shin et al., 2016).  A question to ask of this comparative approach is to what extent the social class dynamics in gentrification, which is a central concern of research, may be set aside when gentrification is framed into a market process. In addition, given the intense debate that surrounds the essential issue of ontological differences between gentrification and processes with similar outcomes in the Southern cities, I suggest that currently, we need! 8!!! Figure 1.1 Conceptual map to state-facilitated gentrification in China. ! 9!more effort in learning about variations within contexts before global regularities of gentrification are reached. This study thus proposes a meso-level approach to show the way by which context can shift the explanation of gentrification. Gentrification in this study is understood as a process of urban and social change, driven by ideologies, forces and agents shaped in multiple scales and exerting influence in the neighbourhood. It is fuelled by urban redevelopment projects but is not equivalent to an urban project prompted in a given place and time. Gentrification develops through two processes of socio-spatial upgrading and working-class displacement. Based on this understanding, both the two processes will be examined through their structural tendencies and grounded processes in localities, which will highlight complexities and contingencies. Analysis at the two scales in conjunction fulfils a meso-level examination of the existence of gentrification in China and the explanation of the regularities within the process. Figure 1.1 presents a conceptual map for this research. 1.5 The state, institutional change and social change in gentrification in China   The study brings a policy/institutional perspective to analyse the ideology and practices of the state in leading gentrification. Institutions in this thesis refer to ideas, norms, and rules that regularise, formally and informally, activities around spatial production and residential relocation. These include not only changing economic regulations but also cultural ideas in place-making. The perspective is required because, in the transitional economy of China, not just a land market process, but institutional rearrangements concerning spatial production in a quasi-land market have guided state-facilitated urban redevelopment (except for China, see also Ghertner, 2014, 2015; Lemanski, 2014). The majority of the existing literature defines state-facilitated urban redevelopment as being stimulated by China’s economic transition, and by land marketisation and housing commodification in particular (He, 2007; He & Wu, 2005; Shin, 2009; Zhang & Fang, 2004). Authors have explored the continuous state intervention over the land market, embodied in interactions between state and market actors. The changing role of the central and local state in a growing market economy is ! 10!also a major issue at stake (He, 2007; He & Wu, 2009; Hsing, 2006; Shin, 2009; Zhu, 2004). Nevertheless, deeming urban redevelopment to be simply one type of urban projects, the authors did not grant sociocultural agents, in particular consumers, an active role in those projects. Meanwhile, working-class residents affected by urban redevelopment were studied as those who born the burden of relocation (Wu, 2004b; 2016), or in other cases, who intensely resisted the result or were disempowered from social participation (Hsing, 2010; Shin, 2013; Wu, 2016). In the gentrification literature, however, the relative play of consumption and production perspectives is from the very beginning a central debate in gentrification explanations. Through the lens of gentrification, landscape change signifies wider societal change, such as the advent of new urbanites to the city and new cultural trends behind place making (Ley, 1996; Zukin, 1982).  Contributing to the gentrification literature on China, this study is intended to approach equally the role of state and society in gentrification. A central assumption throughout the thesis is that the state and society relations, rather than either the state or the society acting in isolation, speak to the characteristics of the gentrification process in China. The aspects of institutional change, which give substance to the ideas and practices of the central and local governments, will be associated with the middle-class and working-class dynamics in a gentrification process. Mainly, in the process of socio-spatial upgrading, the cultural ideology of the local state in landscape making and the commodification of inner-city spaces will be linked to the subject construction of middle-class consumer citizens. In the displacement process, housing strategies for low-income residents will be related to the socialisation of a segment of the working class as well as the marginality of the others (Figure 1.1).  Based on this analytical framework, the study thus also contributes to an elaboration on the complex social process in gentrification. I challenge existing literature on both the Northern and Southern cities that oversimplifies class dynamics of gentrification as an interplay between gentrifiers who move in and indigenous residents who are displaced. Drawing on Lefebvre (1991), space is socially produced and carries on the ideology, values, and meanings of the producers. It impacts the behaviours and identities of individuals in place as well as the distribution of power across societies over the right to ! 11!space. After reconsidering the place and class relationship in gentrification, Butler (2007) suggests that class studies in gentrification should be broadened beyond the middle and working class struggle over living space to encompass the shaping of class through living space choices. This viewpoint reminds us that varied empirical processes and theoretical possibilities exist considering the pattern of place and social change involved in gentrification, particularly when contexts are changing. We are thus encouraged to probe into the subtle and all too often divergent experiences of residents in gentrification, which essentially reflect the multiple relations between place and social change.  1.6 The case study city  Chengdu is selected as a case study in this thesis. Chengdu city is a central city of western China and the capital city of Sichuan Province (Figure 1.2). It is commonly regarded as a second-tier city after Shanghai, Beijing and Guangzhou. Since 2000, the Chinese government has gradually transferred development from the eastern and coastal areas to the inland and western areas, and Chengdu has been at the forefront of national strategies for developing the western region of China.1 Recently, Chengdu’s economy has grown sufficiently to make it one of the fastest-developing cities in China. During the last two decades, two waves of leadership in Chengdu have successively advocated strategies for urban redevelopment, with great ambitions to compete with the first-tier cities of China and merge into the global economy. Apart from the fast pace of economic development, through a series of master plans and strategic plans, the city of Chengdu has been granted by the central government status as a historical, tourist and liveable city in China and as an administrative, service and transportation hub linking western China to the world. These city branding plans generate important stimuli motivating the local government to launch city projects or to plan capital accumulation through the built environment to a greater degree than in cities where developmental strategies have  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 The strategy is called “The Great Development of the West” or “Open-up the West” (Goodman, 2004; Yeung, 2004; Fan, 2006). Observing the situation of uneven development between the western and inland regions and the eastern and coastal regions of China, the State Council advocated the strategy in 1999. The 11th Five-Year Plan of the Great Development of the Western Region of China drafted by the Commission of National Development and Reform (2006) listed a specific agenda regarding to resources supply, environmental preservation, infrastructural construction, industrial restructuring, educational development, economic and institutional reform and so on.  ! 12! !Figure 1.2 The city of Chengdu in China. Source: Drawn by the author.     Figure 1.3 Administrative divisions of Chengdu metropolitan region. Source: Drawn by the author based on a materials provided by Chengdu Institute of Planning and Design.  ! 13!concentrated on industrial development. The city thus provides an appropriate lens for illuminating the characteristics of gentrification in Southern cities experiencing radical urban transformation. The administrative area of Chengdu is a metropolitan region with 12121 km2 and a population of 14 million in 2010 (Figure 1.3). The spatial structure in Chengdu city-region features a “circle tier” pattern, where the city core is surrounded by a large number of small cities, and counties as well as rural areas. Current administrative divisions include the central city (five traditionally established urban districts), four municipal districts surrounding the central city (administratively rescaled during the last three decades), six counties and four county-level cities (Figure 1.3). The first circle tier comprises the central city and part of the four municipal districts; the second circle tier roughly consists of the remaining four districts and part of three counties; the third circle tier comprises most of the agricultural area and eco-region. The “Chengdu-Chongqing Economic Region Plan” approved by the State Council in 2011 proposed construction of a new node in the southern area of the current central city to be called Tianfu New District. As drawn up in the Chengdu Master Plan (2011-2020), Chengdu region includes one central city, one sub-central city, 14 new cities, 34 core towns, 156 towns, and 2000-3000 rural new communities (with a minimum of 50 rural households for each community).  !!!!!!!!!! !Figure 1.4 The inner city and the five historical urban districts of Chengdu. Source: Drawn by the author.  ! 14! This study focuses on the central city of Chengdu, including five main urban districts established before 1990. The maximum extent of the inner city is defined as the areas within the Second Ring Road (Figure 1.4). In 2010, while the central city covered 465 km2 land and contained 5.03 million people, the inner city was 60 km2, with a population of 1.97 million (PCOC, 2010; Chengdu Bureau of Statistics, 2011). Chengdu city is historically a monocentric city, where the downtown is located close to the geometric centre of the metropolitan region. The inner city of Chengdu is commonly recognised as consisting of the areas within the First Ring Road, which is essentially the original city site of Chengdu before the establishment of the People’s Republic of China. However, the modern city of Chengdu has extended from the downtown to, currently, the Fourth Ring Road. In 2011, the Tianfu New District was established on the south side of the city, which has expanded outward to the Fourth Ring Road. This study thus defines the area within the Second Ring Road as the research scope for inner-city gentrification. 1.7 Methodology This study will take advantages of mixed methods, including statistical and spatial analysis, institutional analysis, and ethnographic studies, to investigate gentrification from a structural perspective, the historical perspective, and as a grounded process within the neighbourhood. In addition, rather than single case studies, the ethnographic studies on social groups engaged in gentrification are based on extensive fieldwork in Chengdu. This type of fieldwork will help to identify the common grounds across cases in the specific historical period of post-socialist transition. The triangulation of the three methods and data collection shape a portrayal of gentrification in Chengdu city and from 2000 to 2010, in particular. First, statistics and spatial analysis are used to identify the structural tendencies caused by gentrification at the city level. The structural tendencies include the two successive processes of social upgrading in the inner-city neighbourhoods and the resulting formation of patterns of social inequalities. Specifically, descriptive statistics and mapping show the geography of gentrification and working-class displacement in the ! 15!inner city. The analyses generate information on the socioeconomic characteristics of neighbourhoods before and after gentrification and demonstrate gentrification’s effects on changes in the housing market in the city. Statistical analysis establishes correlations between gentrification and displacement with other independent variables. The quantitative study in this thesis was accomplished using two public databases: the population census at the city and sub-district level and annual real estate statistics at the city level.  The real estate dataset was accessed through the National Bureau of Statistics. The Chinese population census is generated for four administrative divisions that are ranked in descending order: the province, municipality, urban district and urban sub-district. The last two divisions do not contain rural areas and are often densely inhabited. This study draws on data mainly at the sub-district level, which is commonly known as Street level (jiedao, ) in China. The geographical area of a sub-district is larger than a neighbourhood but smaller than an urban district. For the data of 2010, the National Bureau of Statistics keeps the census at the first three levels open-source, but the district government, which is below the municipal government, has managed the compilation and release of the census at the sub-district level. I thus collected the data at the sub-district level by permission of the district governments in the city during the fieldwork.  Second, the study draws on the method of institutional analysis to trace the trajectory of spatial change across waves of state-facilitated urban redevelopment programmes. Institutional analysis offers an important lens for a historical review of the trajectory of urban change. This method helps researchers to probe into the complexities of urban processes in a transitional economy. As Steinmo (2008) has clarified, historical institutionalism stands at the meso-level of the approaches to social science. A meso-approach such as historical institutionalism does not assume that real-world outcomes can be functionally explicable by the overarching structure because contexts exert influence on decisions and outcomes. Institutional analysis thus stresses not so much the who and why of urban redevelopment but how projects unfold in practice. The analysis is instrumental to disentangling the interactions between actors during the process of urban redevelopment, with an emphasis on the path of state intervention in economic, cultural and societal spheres. Data for the institutional analysis were drawn from the textual ! 16!production of the cityscape by the media, policy documents, archives, and official interviews. Finally, ethnographic studies generate information on the decision-making of local actors and the experiences, circumstances and subjective identities of gentrifiers and current residents affected by gentrification at the neighbourhood level. This work does not limit the ethnographic studies to individual cases, nor does it aim to tell a single story of neighbourhood-based gentrification. We recognise that differences can be found not only among regions but also among cases in terms of the driving forces, policy-making and physical modes of gentrification; current residents can experience gentrification and displacement in varied ways. Thus, taking a meso-level horizon in investigating gentrification, the study first selected three exemplary case studies (i.e., the Caojia Alley (CJA), Jinniu Wanda (JW) and Wide and Narrow Alley (WNA) cases); next, it conducted extensive fieldwork in Chengdu.  1.7.1 Fieldwork in Chengdu  The fieldwork was carried out through two stages. The first round of the fieldwork was during December 2, 2014 to April 1, 2015. My purpose was to concentrate on the two important cases of CJA and JW redevelopment. Based on a thorough understanding of the operation of the two projects, I was then enabled to decide on additional candidates of field sites that would be included in the investigation. Investigating the two cases, I interviewed organisational members and relocated residents and collected policy and spatial planning documents. Also, two displacees’ neighbourhoods were identified, where I interviewed migrant tenants who were forced out by the CJA redevelopment. In the later half of the fieldwork, I started a preliminary investigation of the WNA project, wherein I interviewed retail gentrifiers.  The second round of the fieldwork ran through November 2, 2015 to January 31, 2016. The investigation at this stage went beyond the above two cases. Specifically, the WNA project was added as a crucial case, because this case study supplements important findings on property and cultural activism that will be stressed in this study. Interviews were also conducted in three new neighbourhoods where a gentrification process was fully represented and two new off-site resettlement communities. The identification of residential gentrifiers’ neighbourhoods was assisted by the spatial mapping of the ! 17!geography of gentrification in this thesis, which pinpoints the sub-districts that have been most gentrified. In addition, field observations were conducted with residents in an extended geography of redeveloped neighbourhoods and resettlement communities in the city to strengthen the findings gained through the case studies.  The selection of the field sites in this work was based on three rationales: first, all of the redevelopment projects must be located in the central city (not including resettlement communities located in peripheral areas). Second, these projects together represent different modes of inner-city redevelopment and different methods of compensation to the original residents to allow a relatively complete picture of urban processes and social effects to be constructed. Third, given that gentrification is a process that results in the spatial dispersion of the original residents, cases were selected to provide full coverage of the social groups engaged in gentrification. Finally, the redevelopment projects under investigation include those involving mainly public/collective properties built by governments and work units and those with the majority of pre-revolutionary historic buildings. The field sites contain two on-site resettlement communities and five off-site resettlement communities, two displacees’ communities and five gentrified neighbourhoods (two for commercial and three for residential gentrification) (Figure 1.5-1.8).2  A complete profile for interviewees’ neighbourhoods is listed in Appendix A.     Figure 1.5 Field sites: old danwei compound and displacees’ area (CJA and Workers’ Village). Source: Photos taken by the author.  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!2!In this thesis, resettlement communities mean community where residents were resettled by the government after the older neighbourhoods were demolished, and displacees’ communities refer to places where residents not eligible for government-sponsored resettlement moved to independently.!! 18!      Figure 1.6 Field sites: commercialised historic neighbourhoods (WNA). Source: Photos taken by the author.    !! !Figure 1.7 Field sites: on-site and off-site resettlement neighbourhoods (JW and Quanshui Renjia). Source: Photos taken by the author. ! 19!   Figure 1.8 Field sites: gentrified neighbourhoods (Times Riverside and Waterfront). Source: Photos taken by the author.  The three key redevelopment projects are mapped in Figure 1.9. Table 1.1 lists the basic information about the three projects. The first two cases, the Caojia Alley (CJA) and Jinniu Wanda (JW) projects, concern residential redevelopment. The CJA redevelopment project is located on the northeast side of the commercial and business centre of Chengdu. Up to four thousand households (14,000 individuals) previously lived in a 13.2 ha area. Except for the rural-urban migrants, the residents are former and current employees of a state-owned construction group called the Huaxi Group and their families. Within the CJA neighbourhood, there were 2654 public housing and 880 subsidised units owned by the Huaxi Group and the Jinniu District Housing Department, while another 128 units were commercial apartments.  The CJA redevelopment project is one of the pilot projects of the North Chengdu Redevelopment Programme established by Chengdu municipal government in 2012. The redevelopment project was begun in 2012 and finished at the end of 2016. The ! 20!Table 1.1 The profiles of three redevelopment projects in Chengdu  Project name Completion year Area (ha) Redevelopment type Households receiving compensation Total Off-site relocation On-site relocation Cash  Caojia Alley 2016 13.2 Residential redevelopment of danwei/public housing 3756 600 2400 700 Jinniu Wanda 2012 13.6 Residential redevelopment of danwei/public housing 3162 162 2100 900 Wide and Narrow Alleys 2008 6.66 Commercial redevelopment of historic buildings 891 891  Note: The total households eligible for claiming compensation include not only owner occupants and public tenants but also absentee owners. They do not contain private tenants who were rejected for compensation.            Figure 1.9 Locations of main field sites in Chengdu. Source: Drawn by the author based on materials provided by Chengdu Urban Planning Institute.  ! 21!government and a financing company led the project as the primary actors in property expropriation and land consolidation; later a private corporation named Evergrande won in the land auction of CJA area and was in charge of real estate development. Impacted by the on-site resettlement policy, among the 3756 households in the old CJA neighbourhood who are eligible for compensation, approximately 2400 opted for the state’s offer of on-site relocation, while about 600 households were relocated off-site in three communities, which are outside the Third Ring Road of the city. More than 700 households opted for monetary compensation (O2). Aside from the original site of the old neighbourhood, three off-site resettlement communities and one displacee area for the original CJA residents were identified in the map. Note that households eligible for claiming compensation include not only owner occupants and public tenants but also absentee owners. This means the 3756 households include residents who were actually living outside the redeveloped neighbourhoods when the project was established. And the precise number of private tenants who were rejected for compensation was not recorded.  The second area, Jinniu Wanda (JW) is situated along the north side of the First Ring Road of Chengdu. The 13.6 ha area once accommodated 3162 households. Similar to the CJA neighbourhood, the old JW neighbourhood also consisted of work unit (danwei) compounds, but the danwei compounds were almost all small-scale, affiliated with public sectors employers instead of large state-owned enterprises. Also, owing to its proximity to wholesale clothing and houseware markets, the proportion of migrant workers among original residents in the JW neighbourhood was larger than the CJA neighbourhood. The JW project was from the very beginning endorsed by the government but operated by the Dalian Wanda group, which is commonly deemed to be China's largest private property company. The project started in 2009 and completed in 2012. In this project, 2262 households opted for in-kind compensation, among which 2100 households chose on-site resettlement, whereas only 900 households have opted for monetary compensation or off-site relocation (Jinniu Yearbook, 2012). After redevelopment, the place is separated into three sections. One section is reused for resettling existing residents while the other two are developed respectively to residential and commercial properties for new consumers and investors. Compared with the latter two sections, the resettlement neighbourhood is much more densely constructed, ! 22!containing eight high-rise buildings (more than 2500 apartments) on a 2.2 ha land parcel. The plot ratio reaches to 7.77. The Wide and Narrow Alleys (WNA) was previously a hutong (alley, or narrow street) neighbourhood adjacent to the west side of the CBD of Chengdu. The 6.6 ha area contained three hutongs and courtyard dwellings that were constructed in late imperial China and the pre-revolutionary period. Initially, the WNA area was constructed as a station of Manchu military power in Chengdu in the Qin Dynasty. The area was managed by the military power beyond the jurisdiction of Chengdu administration. Thus, the WNA is commonly known as Shaocheng (small city), which means a small city nested independently within the Dacheng (big city) of Chengdu. Before redevelopment, the place was inhabited by local residents, who were in possession of the buildings in the pre-revolutionary days, and public tenants, who moved in after the buildings were confiscated by the government and shifted to public housing. From 2003 to 2008, the WNA area experienced commercial redevelopment. Most of the traditional buildings were demolished and reconstructed using traditional construction techniques. The area is now a renowned tourist place in Chengdu. The project caused wholesale off-site relocation of local inhabitants. All of the 891 indigenous residents were relocated off-site during the process of commercialisation.   I approached a majority of the institutional members via the recommendations of different intermediaries. My relationship with the intermediaries came from pre-established connections with the Chengdu Institute of Planning and Design, the China Construction Southwest Design & Research Institute and the School of Planning and Architecture in Southwest Transportation University in Chengdu. However, despite these personal connections, normally, the process of establishing a connection was roundabout: first an acquaintance or friend introduced me to an intermediary. The intermediary would then give me the phone number of the informant whom I had identified (e.g., a member of a government sector, an expert in urban planning and architecture or a residential committee member). The advantage of this sampling method was that the intermediaries helped in building the trust of the informants, which then smoothed the conversations. However, on occasion the intermediary redirected me to another informant. For the ! 23!sampling of resident informants, I combined the methods of recommendation by residential committees and snowball sampling.   Three major groups participated in the study: organisational members (6 city officials, 2 danwei managers and 1 manager of the state-owned development company in the CJA project and 6 urban planners and architects who managed the operation of the JW and CJA projects); relocated residents who originally lived in the old neighbourhoods and have either been relocated into new communities or are still waiting for relocation (20 off-site, 20 on-site, 10 activists; 20 displacees); gentrifiers (12 residential gentrifiers from three different neighbourhoods in the inner city and 10 commercial gentrifiers from two renovated historic areas). Activist were residents with an intention to stay put and once intensely resisted the implementation of housing demolition and removal. In addition to interviews, two focus groups were organised with residents and experts, respectively. Table 1.2 enumerates the groups and numbers of interviewees from the field sites. All informants’ information is listed in Appendix A. In the following text, each interviewee is identified by a classification code – O for organisational members, R for relocated residents, G for gentrifiers – followed by a chronological interview number.   Table 1.2 The locations of informants  Informant group Field site The number of informants On-site relocated residents Caojia Alley  4 Jinniu Wanda 16 Off-site relocated residents  Qingxi Yazhu  1 Quanshui Renjia  10 Donghong Guangxia 4 Jinxiu Dongfang 5 Activists Caojia Alley 1 Jinxin Jiayuan 4 Wide and Narrow Alley  5 Displaced migrant tenants Workers' Village 18 Balizhuang 2 Retail gentrifiers  Wide and Narrow Alley  8 Tangba Street 2 Residential gentrifiers Times Riverside  4 Chengdunese Paradise 4 Waterfront 4  Primary data were collected through semi-structured interviews. Focus groups of residents and participant observation of their social and community life are also used to reinforce my connection to the cultural environments and circumstances of the residents. ! 24!The length of interviews with relocated residents and gentrifiers ranged from one to two hours. For organisational members, the discussion often extended to over three hours. Also, I had the chance to conduct in-depth communication with one activist seven times. In a displacees’ district, I was involved in the daily life of a few migrant tenants through being the home tutor of the tenants’ children. This kind of relationship enabled me to connect with the other displacees far sooner than I had expected. Still, my involvement with the few residents deepened substantially my understanding of the status of the migrants in the city. Questions addressed to the groupings of participants are given in Appendix B. The semi-structured interviews with organisational members participating in the decision-making process mainly covered four areas of information: basic information about urban redevelopment programmes and projects in the case study; questions about redevelopment policies at both the central and the local levels and about policy implementation; information on their participation in decision-making and social governance; and their subjective understanding of the projects. In-depth life history interviews were conducted with relocated residents to generate data on the association between changing housing conditions with the life opportunities, lifestyles and self-identification of the relocated residents in the long run. Through the interviews, residents were encouraged to share their work and housing experiences and critical events in their lives. For the most part, individual experiences mirrored the years of turbulent change from the danwei society to a market society and from an urban-rural division to the opening of the city to migrants. Then, the informants were asked about their interactions with other parties in the process of urban redevelopment and, finally, about their subjective perceptions of the projects and the other parties’ behaviour. The semi-structured interviews with gentrifiers concentrated more on the interviewees’ current occupation and income, the reasons for their housing choice, the lifestyle and consumption practices in the inner-city neighbourhoods, and self-identification, especially whether they distinguished themselves from other middle-class residents of the city. For better understanding, we should note that although generally divided into the middle class and the working class, the residents who participated in this study could be grouped based on new divisions. According to different working-class experiences, for ! 25!example, there are categorisations based on tenure (e.g., private tenants, private homeowners) and on the type of compensation (e.g., on-site resettled residents, off-site resettled residents and residents who received monetary compensation). Additionally, for gentrifiers, the study compared the conditions and attitudes of both young, cultural pioneers in historical sites and residential gentrifiers of high-end housing in the inner city.  1.8 Thesis overview Starting with a concept that contains vagueness in its definition, Chapter 2 primarily takes a historical approach to reviewing the evolution of gentrification and to clarifying its essential properties. The clarification consolidates the ontological basis of gentrification as abstraction in this research. Then, two bodies of work, one leaning towards global theory building and the other towards the revelation of regional differences, are summarised. This review is intended to illustrate the dilemma of knowledge production for this concept, which justifies the meso-level approach taken by this study. Following the global debate, Chinese work referencing gentrification is reviewed. Approaching this body of work, the present study claims an institutional perspective to connect state action and social process in gentrification in China.  Following the conceptual statement and methodological approach, Chapter 3 specifies the socio-spatial upgrading in a central city of China and identifies its formation within the post-socialist context. The chapter starts with background on socioeconomic restructuring and social-cultural change in post-reform China. In contrast to an advanced post-industrial context, the background reveals a distinct trajectory of societal change caused by the expansion of low-income service industries and the emergence of an individualistic consumer culture among China’s new rich. Then, this chapter maps the geography of socio-spatial upgrading in the Chinese city of Chengdu in the 2000s to confirm the extent and the spatial manifestation of gentrification. From this contextual and geographical base, the chapter finally correlates a social upgrading index with a series of independent variables, aiming to establish a dependent relationship between the gentrification process and broader societal and spatial transformation in the post-socialist era. This initial analysis concretises the structural tendencies of socio-spatial upgrading in ! 26!Chengdu and outlines the contextualised characteristics of gentrification in the post-socialist society, differing from conventional explanations in advanced post-industrial society.   Departing from the structural tendencies, the next two chapters approach the two grounded empirical processes that complete the cycle of gentrification. Chapter 4 examines the process of inner-city upgrading, considering the agency of both political-economic actors and middle-class consumers in the city. The role of the state is emphasised, particularly state-capital-society relations, thereby linking state intervention with societal change. Taking a historical perspective and using institutional analysis, the chapter examines two aspects of the urban redevelopment practices conducted by political-economic actorsthe creation of new landscape as well as land capitalisation and housing privatisation. These two practices then direct the study towards the implications for social dynamics the construction of Chinese modernity and the elevation of housing consumptionthat underlie the genesis of gentrification in China. Finally, by analysing the cultural and identity change of the middle-class gentrifiers, this chapter will explain the causes and patterns of the gentrifying process by combining both production and consumption forces.  Chapter 5 then turns to the experience of the working class in gentrification. Through ethnographic studies, it investigates working-class circumstances during the process of residential relocation induced by inner-city redevelopment, highlighting the influence of a hegemonic state. It stresses the different perceptions and compensation outcomes among different groups of the affected residents. It also uncovers the interactions between the residents, local officials and developers in a social governance process that allows consensus building. The divergent social outcomes complicate the meanings of housing demolition and relocation for the working class in China. Moreover, they reveal the uniqueness of state-society relations in the post-reform era of China. Nevertheless, they also raise a question on the conceptual extension of gentrification, which is committed to uncovering social inequalities in the urban process.  Based on this consideration, the study extends the examination of gentrification effects to the structural level of social inequities in Chapter 6. By tracing the spatial distribution of the working class in the city of Chengdu from 2000 to 2010, the chapter ! 27!proves the existence of working-class displacement and a potential for concentrated poverty in the inner city as a result of gentrification. Analysis of real estate datasets at the city level and census data at the sub-district level discloses that gentrification leads to housing speculation on the one hand while undermining housing affordability for the working class on the other. By mapping the changing geography of households by tenure in the 2000s, this chapter criticises tenure-based residential segregation and the exclusion of the working class, which are both outcomes aggravated by gentrification. The patterns of social inequalities will strengthen the relationship between gentrification and social injustice in China. The concluding chapter brings together the key findings of the studies at the structural level and grounded level. Joining the broad contexts with socio-spatial restructuring and activities in localities, the purpose is to delineate a gentrification process that appears in Chengdu at the middle ground. Meanwhile, it stresses the unique state-society relation and the way it articulates the peculiarities of the gentrification process in Chengdu. Based on the experience of studying China, this research returns to the original concept, identifying characteristics that reinforce the soundness of gentrification theory while also necessitating this type of meso-level knowledge production. Moreover, specific implications for studies on state-facilitated gentrification are summarised. Finally, the thesis present four directions for further research to enrich the understanding of Chinese urbanisation and gentrification in discussions of the Global South.          ! 28!Chapter 2 Gentrification Studies on a Global Scale  Gentrification studies have been developed in European and North American cities since the post-war era (the 1960s), a period noted by authors as characterised by critical social movements. It continued through to the liberal stage (the 1970s), the high water mark of the welfare state. Then, the body of literature transitioned to the neoliberal 1980s, marked by a “roll-back” of the welfare state (Peck & Tickell, 2002, p. 388), and finally, to the contemporary stage of globalisation, with a feast for the competitive market (Ley, 2012). The advent of a global context and geographical variations in both the cause and effect of the gentrification process require a rethinking of the conceptualisation and research paradigm with reference to classical gentrification. This chapter accepts that challenge and attempts to find an approach to understanding the gentrification process in China. It draws on three bodies of literature. The first section consolidates an ontological basis by tracing the course of (re) conceptualisation in mainly the Western literature. Next, I will elucidate the epistemological difficulties of gentrification research globally by examining both proposals for global theory and recognitions of the context-dependent nature of the process. Finally, particular attention will be paid to the characters of gentrification and urban redevelopment explained in the literature on China. However, although studying gentrification initially on a global scale, I do not assume that the urban process is currently emerging in every metropolitan region worldwide. What I investigate is how the paradigm of knowledge production should be reconsidered when that urban process encompasses the multiple realities of cities and regions. I do not presume any North-South dichotomy in theory production. These terms are used only as a geographical division, denoting places that currently yield diverse experiences and insights outside the heartland of gentrification scholarship. As Connell (2007) stated, “To use concepts such as ‘periphery’ is just the beginning of analysis, not the end” (p. 213). Context, rather than the territories per se, directly contributes to the multiplicities of gentrification in this thesis.  ! 29!2.1 Ontological basis of gentrification  Why are people concerned with gentrification? This primary but also important question is particularly pressing given the need to justify studies addressing a hard-to-define process. Essentially, the question concerns the ontological basis of the concept; thus, it is used to guide the direction of theoretical formulation and to specify the rationale and soundness of conceptual deployment. However, I have also borne in mind that the meaning of a concept is constantly polished by scholars through empirical studies and theory production, likely resulting in reconceptualisations. This section reviews how theorists have directed the meaning of gentrification to identify a viable ontological basis. 2.1.1 The essence of gentrification  Ruth Glass (1964) initially established the concept of gentrification in London to describe a process observed in a neighbourhood wherein middle-class residents replace the working-class; this shift is associated with landscape change and the price appreciation of residential buildings. The process referenced by Glass, however, applies particularly to the meaning of the lifestyle change of the middle class. A change in the middle-class way of life causes the process of gentrification. Glass granted a distinctive meaning to gentrification by emphasising its middle-class origin. It results from a middle-class view about central cities, rather than from any other process to physically and socially upgrade neighbourhoods (Brown-Saracino, 2010, p. 15). Thus, classical gentrification is at first gradual, one building at a time, one family at a time. The concept draws a clear but also narrow boundary and is clearly distinct from the process of urban regeneration. Peter Marcuse’s working definition emphasises the twin processes of gentrification and working-class displacement. Moreover, Marcuse (1985) clarified the meaning of displacement to precisely capture the changing social structure of the neighbourhood “to a degree differing substantially from the general level of change in the community or region as a whole” (p.199). It should include not only the direct displacement (physical or economic) of the households living in the area but also exclusionary displacement via the shrinkage of the housing market available to those households whose socioeconomic status prevents them from living in the now gentrified neighbourhood.  ! 30!Gentrification occurs when new residents - who disproportionately are young, white, professional, technical, and managerial workers with higher education and income levels - replace older residents - who disproportionately are low-income, working-class and poor, minority and ethnic group members, and elderly from older and previously deteriorated inner-city housing in a spatially concentrated manner, that is, to a degree differing substantially from the general level of change in the community or region as a whole (Marcuse, 1985, p. 199). Thus far, gentrification is still considered a neighbourhood process focused on the nexus between two classes on the issue of urban living. Placing gentrification within the period of post-industrial society and post-modernism has greatly deepened explanations of its relevance and meaning. Increasingly, the essential meaning of gentrification has been expanded from a neighbourhood phenomenon concentrated on individual lifestyle choice to include the cultural politics of changing urbanism in line with socioeconomic transformation and cultural change in post-industrial society. Gentrification authors have framed the relationship between gentrification and the consumer culture of a new emerging “service class” in the tide of structural change in occupations in post-industrial society (see also Hamnett, 1989; Ley, 1996, Savage et al., 1992). Other authors have examined the political ideology underlying the cultural characteristics of post-modernism and its expression in landscape styles and aesthetics (Ley, 1987a, Mills, 1993; Harvey, 1990; Zukin, 1998). In addition, the literature has revealed the relationship between gentrification and the appearance of other types of social groupings in the wake of post-industrial society and post-modernism, such as racial minorities and full- and part-time female workers (Butler, 1997; Butler & Hamnett, 1994; Bondi, 1999; Warde, 1991).  Specifically, focusing on the Canadian contexts, Ley (1980) solidified the emergence of gentrification within the liberal ideology of the newly emerging middle class, who once dominated the Liberal political party, in post-industrial society. The champions of social and cultural liberalisation promoted progressive political reform from the 1960s to the 1970s in opposition to the dominance of economic values and formalistic culture under Fordist mass production (Ley, 1996). The ideology of liberalisation invokes claims for a new urbanism that includes landscape aestheticisation, cultural diversity, neighbourhood enhancement and historic preservation, in the advanced industrial city. By the 1980s, however, the progressive reform in pursuit of social and ! 31!cultural care had faded into the middle class’ claim for a class interest in urban amenity, but at the cost of social justice for the poor. Drawing on the concept of cultural capital, Zukin (1982) and Ley (2003) also treated gentrification, particularly in neighbourhoods with high historic and cultural value, as bearing on the convergence of cultural production and capital accumulation.  The post-industrial explanation for gentrification has greatly reinforced its structural reach by linking the process with macro social, economic and cultural forces. Gentrification is an iconic process accounting for the characteristics of urban and social change in post-industrial cities. Theoretically, the body of the explanation is still grounded in the central relationship between place and social class; the cultural connotation of the gentrified landscape is still deemed to be the most important intermediary linking multiple forces together. The gentrified landscape, as an expression and representation of the emergence of a distinctive urbanism, reflects broad processes of socioeconomic change, capital (re)investment in style, and the displacement of those who are ill-suited to the trend. For instance, as suggested by Brown-Saracino (2010, p. 14),        “Zukin and Glass hold gentrifiers' cultural valuation of the central city as an important component of our concept of gentrification.” Gentrification processes reflect “a major component of the urban imaginary” (Ley, 2003, p. 2527), in which the so-called creative class (Florida, 2002) and other powerful decision-makers influence the path of urbanism towards a desirable middle-class lifestyle. As Mills (1993) stated, inner-city gentrification contains essentially cultural meanings generated through symbols, identity creation and social order representation “through which the economic process (of production) and social process (of consumption and social change) operate and are experienced” (p. 165).  The production-side explanation of gentrification departs from the grounded view of the place and class relations in gentrification. Smith (1982) started with the economic location of inner-city terrain and aimed for a structural explanation of gentrification. Rooted in the capitalist system of production, the historical geography of gentrification before the 1980s, for Smith, is part of the broad process of urban restructuring towards long-term uneven development. At the urban scale, uneven development is stimulated by the ground rent structure in those post-war cities that experienced suburbanisation. In these cities, a “land value valley” emerges in the inner-city areas surrounding an urban ! 32!centre (1982, p. 146; 1996, p. 58). When an economic crisis happened and the redirection of capital investments to the built environment was necessary, the inner-city areas would absorb new investment searching for new developmental potential and, most saliently, exploiting rent gaps. The advantage of development stems from the high ground rent of the central city due to the established, favourable features of downtown locations. A high ground rent accordingly implies a high likelihood that ground rent will be capitalised in nearby devalorised locations and, finally, that the rent gap will be filled. Gentrification is expected to be resurgent in a long-standing back-to-the-city movement in which development capital reverses the declining trend of inner cities. This is a trend associated with the essential nature of capitalism towards uneven development and the economic cycle of disinvestment and reinvestment to exploit rent gaps (Smith, 1982).  According to Smith and Williams’ (1986) theoretical aspiration, the neighbourhood-based, explicitly described class process is out-dated. Gentrification is not necessarily confined to residential or commercial gentrification, and the process does not require a clear-cut definition of the middle- and working-class involved in it. Gentrification might not even need a strict definition: “Rather than risking constraining our understanding of the developing process by imposing definitional order, we should strive to consider the broad range of processes that contribute to this restructuring and understand the links between seemingly separated processes” (Smith & Williams, 1986, p. 3). Moreover, the ultimate purpose for explaining gentrification is to illuminate the capitalist logic behind it. “Gentrification, and the redevelopment process of which it is a part, is a systematic occurrence of late capitalist urban development,” although a working definition remains:     “the process by which working class residential neighbourhoods are rehabilitated by middle-class homebuyers, landlords, and professional developers” (Smith, 1982, p. 152, 139).  A basic concern about the uneven development assumption for gentrification is that it starts from the overall decline of the inner city; this goes beyond Glass’ stress on those uniquely gentrified places as defined by middle-class bias towards lifestyle and living places. The assumption better explains the inevitable economic cycle of disinvestment and reinvestment in inner-city areas, which dates back to Hoyt’s model of land value change through time, but cannot explain why some dilapidated places (and some cities) ! 33!are subject to gentrification while others are not. Nor does it address why there are particular periods for the emergence and resurgence of gentrification and periods when gentrification cools down (Ley, 1987b; Hamnett, 1991). In other words, Smith has offered evidence on why inner-city land is likely to be upgraded due to its high productive potential but has not directly outlined the social process behind this, which is the core distinction of gentrification as an abstract concept. Through this omission, it is somehow assumed that reinvestment is equivalent to the result of social upgrading and that the meaning of gentrification is constructed by the (productive) meaning of the inner city. As Betancur (2014), focusing on the context of Latin America, has stated:   If we defined gentrification only as rent capture, any development involving rent could be classified as such. I am applying the term here to the reuse of low-income neighbourhoods by higher income populations. (2014, p. 11) Of course, the other side of the gentrification coin should not be ignored: the social inequality implied in gentrification. Over this period, relatively few studies started from the point of view of the working class, and they instead examined causes based on the changing middle class. Since approximately 2000, following the resurgence of policy-driven gentrification in some European and North American cities, class conflicts and the working class have increasingly returned to this focal point to defend the conceptual significance of gentrification. In part, this change has occurred because policy discourses such as the urban renaissance, the alleviation of social exclusion, and social mixing have tended to soften public opinion and destabilise somewhat the politically critical position held by earlier notions of gentrification (Slater, 2006; Lees, 2008). This change is also due to the retreat of the middle class as direct initiators of gentrification in favour of a shift to state-led, policy-driven urban regeneration programmes.  The tipping point was identified by Tom Slater in the middle 2000s; he criticised the missing critical voice in gentrification and endeavoured to renew concerns about the displacement of the working class (Slater et al., 2004; Slater, 2006).  He and other scholars emphasised that gentrification is anchored in those fields addressing class issues (Slater et al., 2004; Butler, 2007; Wyly & Hammel, 1999). After reviewing debates on the location (centre, suburb or rural areas) and categories (commercial land-use or residential land-use) of gentrification, Slater et al. (2004) finally asked for more rigour in ! 34!Glass’s definition and concurred with one study where the authors stress the meaning of class in gentrification:  Rather, the most important aspect of her work that we should register is her critical emphasis on class transformation. Whether gentrification is urban, suburban, or rural, new-build or the renovation of existing stock, it refers, as its gentri-suffixes attest, to nothing more or less than the class dimensions of neighbourhood change, in short, not simply changes in the housing stock, but changes in housing class (Lambert & Boddy, 2002, p. 144). One of the salient points in reinstating gentrification within the field of class relations and conflict is that it is concerned with both the middle and the working classes. Scholars adopting the class issue in gentrification have thus contributed a great deal of the work on displacement of the working class (e.g., Atkinson, 2000, 2002; Wyly & Hammel, 2004, Wyly et al., 2010, Allen, 2008; Watt, 2008; Butler et al., 2013). These authors have also intensively criticised studies that tend to neutralise gentrification, which argue that urban regeneration does create better urban environments and local services for not only the middle class but also for the working class; it also increases local private and tax revenues that can be used in broad-scale urban services (Byrne, 2003; Freeman, 2006; Freeman & Braconi, 2006). This position has attracted considerable scepticism, for example from Newman and Wyly (2006). Others have turned to revisiting the concept and methodology of displacement and to recovering the hidden inequalities in policy-driven gentrification (Davidson, 2008, 2009; Lees, 2008; Shaw & Hagemans, 2015).  Based on the background of the globalisation of production and the neoliberal transition of the state, Smith (2002) finally contended that gentrification is an emergent urban development strategy on a global scale. While gentrification provides new theoretical and political significance for the representation of urban and social change based on neoliberal ideology, nevertheless, the global extension of the geography of gentrification scholarship has increasingly called for a reconceptualisation of the core concept. In a series of studies on global gentrification, researchers have attempted to decontextualise and refine the abstraction of gentrification to adapt to its variation in time and space (see Atkinson, 2003; Clark, 2005; Lees et al., 2015, 2016; Shin et al., 2016).  According to these works, the focus of an abstracted gentrification ultimately returns to a class-related urban process. Moreover, they increasingly emphasise criticism based ! 35!on the social injustice caused by gentrification. Typical is Clark’s essay where he clarified the relationship between particularity and conceptualisation by arguing that it is necessary to erase those “non-necessary” relationships too frequently used in the abstraction of gentrification, which render the concept chaotic (2005, p. 258). Clark noted two aspects of the broad defining characteristics of gentrification, with the aim of better understanding its essence: Gentrification is a process involving a change in the population of land-users such that the new users are of a higher socioeconomic status than the previous users, together with an associated reinvestment of fixed capital. (Clark, 2005, p. 258).  This definition was later used by Lees et al. (2015) as a viable abstraction for gentrification on a global scale. They argued that the generality across global examples of gentrification lies in the common trend of the secondary circuit of capital accumulation through the built environment or real estate industry to transform urban land use, catering to wealthier inhabitants. Accordingly, they suggest simplifying the abstraction of gentrification to a process of class-related urban change and social cleansing to maintain the essence of the concept and to open gentrification processes to multiple series of comparative studies. Lees et al. (2015, 2016) underlined social cleansing as the condition that distinguishes gentrification from urban regeneration. They support the significant meaning of gentrification wherein the ultimate goal of the conversation is social justice for all (Lees et al. 2016), as earlier stressed by Atkinson (2003), who stated, “We shall focus on urban inequality where the phenomenon of gentrification thrived” (p. 2347).  We have endeavoured to define gentrification. It is not because a fixed definition is necessary, but the geographical extension of its conceptual use requires a convincing rationale. Why gentrification rather than something else? The question is essentially about an ontological match: the meaning of gentrification being used to generalise an endogenous process expressing comparable outcomes to gentrification. From this brief review, it can be seen that two contributions have created conceptual meaning: first, the origin of the conceptualisation and second, the significance added by scholars from different theoretical perspectives. Some of these meanings of gentrification, specifically those that link gentrification with socioeconomic background, can be chosen by people outside of the geographical heartland of gentrification in Anglo-America as part of the ! 36!ontology of the concept. This essential formulation precludes the concept from being stretched in particular contexts, such as connections to the post-industrial transition or capital switching from the suburbs to the inner city. Such contexts now become contingent rather than essential relations.  However, with the advent of the globalisation of gentrification scholarship, a certain level of agreement has been increasingly achieved. First, most of the agreement about the conceptualisation centres around the variable forces and causes. On this point, Glass’s definition from the early 1960s, which identified the singular factor of middle-class residency, might prove inevitably outdated. Then, the historical geography of gentrification invariably solidifies the basic meaning of gentrification as focusing on social class change and inequalities in place. This meaning is the base for distinguishing the theoretical orientation of gentrification from that of other processes based on similar empirical grounds.   Thus, I will argue that essentially a shift in class power from lower strata currently located in a place to higher social strata is at the base of the ontology of gentrification. This statement responds to the contention raised by Southern scholars concerned with the existing provincialism of knowledge production based on gentrification. Thomas Maloutas (2012), for example, has contributed an intense critique of Anglo-American hegemony in gentrification research. His main argument contends that the conceptualisation and theorisation of gentrification have been dependent on contextual causality generated in the Anglo-American metropolis. However, by simplifying the idea of gentrification, authors tended to apply the provincial concept to a broader range of contexts. This practice has not only reduced the rigour of theory building but has also resulted in the contextual extension of gentrification from the Western core to the periphery. According to this understanding, I assert that the concept holds its decontextualised nucleus. Neither the economics of land reinvestment nor the secondary circuit of capital accumulation nor an iconic urban process of neoliberalism and globalisation, but only the social class dynamics in place and the associated inequalities express the theoretical strength of gentrification when it is deployed globally. Gentrification is broadly affiliated with categories of urban change, social class and urban ! 37!politics. The one and only essential postulate behind the concept is that place change is not neutral but is political.   The meaning of gentrification not only consolidates its theoretical strength but also delimits its conceptual boundaries. It is recognised that, implicitly and explicitly, the existing literature may have treated urban renewal and gentrification as interchangeable concepts. The shaky base of conceptual understanding renders gentrification in the literature either an overarching conception or easily attacked given its lack of academic rigour. Gentrification scholarship is one thread of theoretical production empirically grounded in urban renewal; I will distinguish it from other urban theories in terms of its perspective and approach to knowledge production. The above statement aims to deliver not so much a definitional base as a theoretical direction for gentrification studies. I agree that a working definition, which refers to the conditions used by authors to define a process that is considered to be gentrification, can either be based on the simplest abstraction or can be regarded as contingent and varied. However, regardless of whether gentrification is caused by political-economic or sociocultural forces, gentrification studies must ultimately return to the generalised concept of social (class) change in relation to place change. Not only the forces and ideology behind gentrification must be uncovered but also, crucially, the relationship between place and the trajectory and results of societal (class) change and restructuring in the city must be generalised through a link to the relevant forces and ideology. Otherwise, gentrification studies will lose their distinctive knowledge contribution in comparison with urban regeneration and similar areas of study. Accordingly, it would then be difficult to justify an epistemological advantage from employing the concept of gentrification to study the urban restructuring of the Global South.  2.2 Epistemological problems of gentrification on a global scale  Accompanying sweeping urban transformation, gentrification has gradually come to the notice of scholars from a wide range of metropolitan regions in the Global South. While proponents of globalisation and economic structuralism endeavour to construct a global theory for the process, others, particularly those from the Southern Hemisphere, ! 38!have offered abundant evidence of territorial particularities. In this stalemate, many scholars display deep scepticism regarding the rationale of gentrification as a global agenda of research. This section examines the debate on globalisation and regional heterogeneities, with the aim of elucidating the epistemological difficulties met by gentrification studies in the South. For anyone who prepares to step into the arena of gentrification, in either the North or the South, these epistemological problems require prudence in the types of knowledge produced for subsequent abstraction. 2.2.1 A Global theory for gentrification?     In 2002, Smith argued that gentrification had now gone global, spreading from the Global North to the Global South and showing remarkable growth in metropolitan areas; economic globalisation and the neoliberal state restructuring are the new forces driving gentrification on a global scale. The globalism of economic production has produced two mutually intensifying impacts on the role of cities. First, capital and jobs flow among the centralised metropolitan areas, which constitute the nodes of the global economy. Second, the national state has retreated from city management and welfare state policies, as the metropolitan production system has become increasingly disconnected from the national economy. However, the national state has returned, having recast its economic role as a neoliberal state. The liberal urban policies facilitating social services for the local population to sustain social reproduction have been weakened and dismantled in favour of neoliberal urban policies aiming at reinforcing the productive potential of the city. The city tends to be seen as a “centrepiece of productive investment” (Smith, 2002, pp. 434-443). The neoliberal turn has driven changes in Western cities so that they are no longer centres of “progressive reform and policy innovation” but are reflections of a changing ideology (Peck, 2006, p. 683), that is, an emerging urban entrepreneurism (Harvey, 1989).  Urban regeneration and gentrification are deemed to be typically neoliberal urban strategies. The prevalence of neoliberal political ideology is also creating a common global trend of gentrification in relation to the role of the state. Hackworth and Smith defined this phenomenon as the third wave of gentrification that began after the 1987 recession in European and North American cities; this wave involves far more aggressive action on the part of real estate industries and greater state stimulation of the free market ! 39!(Smith & Defilippis, 1999; Hackworth & Smith, 2001). The neoliberal urban strategies lure the middle class back to the city (Smith, 2002; Lees et al., 2008) and strengthen the tax base and labour market potential; they absorb the entry of high-end service industries; they increase foreign direct investment in industrial and real estate development; and they stimulate city consumption with urban spectacles (Harvey, 1989; 1990). Foremost, by creating a promising business climate and comprehensive urban facilities, urban regeneration is likely to upgrade the bond rating of the metropolis as assessed by international bond-rating agencies. Thus, the city holds potential advantages in its ability to attract more investors in government bonds for public service delivery (Hackworth, 2007; Lees et al., 2008).  Considering these economic motives, Smith (2002) treated the current wave of gentrification as an urban strategy of states. To date, gentrification has transcended a project-based urban practice to be “generalised as a central feature of this new (neoliberal) urbanism” (p. 430). Notably, according to Smith, merging into the global economic system will spark far more significant and creative transformations in the new emerging metropolises through social production instead of reproduction, because the transformation of these cities is relatively less fettered by the Keynesian welfare system.  The central platform of the new wave of gentrification is thus anticipated to be the new metropolises rather than cities of the advanced capitalism. From a social and cultural perspective, another set of global postulates is also arising against the backdrop of cosmopolitanism and transnational urbanism. Classical explanations link the class dynamics in gentrification with the cultural politics of initially a young professional middle class, related to an urban lifestyle and rooted in ideas of autonomous self-expression (Ley, 1996, 2003; Hamnett, 2003; Harvey, 2005; Zukin, 1982). Transitioning to the new context of cultural globalisation, the cosmopolitan population and culture promote an urban imagery that is harnessed by those local ruling elites making decisions in favour of gentrification; meanwhile, gentrification is now based on so-called neoliberal ideas. In this global context, the connection between cultural ideology, urbanism and middle-class mobility implied in gentrification could be reinterpreted to indicate that the state, transnational capitalists and a new social group composed of a cosmopolitan population have been the main agents initiating the process. ! 40!Mega-projects are launched by the state, which aims to deliver a global image for its cities to attract expatriates and domestic professional and managers (Atkinson & Bridge, 2005). Accordingly, the cosmopolitan population also reshapes the neighbourhood landscape and socio-culture in order to establish a sense of place (Butler, 2003, 2007) and to reconstruct class identities (Lees, 2003; Butler & Lees, 2006).  Compared with political-economic restructuring, global urbanism is a less controversial concern in empirical studies on the Global South, although these studies do ask for caution in the formulation and implementation of so-called neoliberal ideas and policies in nations (Lees, 2012). Even in metropolitan regions possessing limited competitive advantages in the global economy, governments at the national and sub-national levels could, perhaps consciously, take a world-city landscape as an imperative to sell the city to transnational investors and consumers dreaming of a cosmopolitan lifestyle. The advancement of urban living environments and recreational facilities is a typical example of the neoliberal urban strategies, for example as manifested in waterfront redevelopments (Davidson & Lees, 2005; Wong, 2006) and tourist place-making (Herrera et al., 2007; Rubino, 2005). Gentrification triggered by international mega-events is also noticeable. Metropolises compete to host international events in an attempt to capture the potential to reinforce their urban reputation and attract inward investment opportunities. In addition to economic incentives, the grant of a specific mega-event is regarded as a symbol of national status. Moreover, mega-events are instrumental in forging and consolidating social unity and national identity in a nation/city (Shin, 2009a). However, the consequent downside can sharply contrast with the city’s prosperity. During the preparation for the Beijing Olympiad, 1.5 million families suffered from residential demolition from 2000 to 2008, approximately twice as many as were impacted between 1991 and 1999 (Shin, 2009a, p. 131). Recently, Lees et al. (2016) proposed a relational comparative approach to global gentrification. Based on the scholarship of comparative urbanism, the relational comparative approach sees gentrification processes as now being interconnected between different cities due to the globalisation of capital, but with contextualised local embeddedness. Theoretical production could thus be based on the empirical knowledge of both differences and similarities. This would allow different experiences from different ! 41!cities to generate questions about experiences elsewhere so as to constantly reinvent theoretical production. In other words, it would initiate the transnational examination of gentrification processes.  The future depicted is an exciting one. It transcends the existing approaches in the gentrification literature of the Global South that have either selectively used established explanatory frames for alternative contexts or have revealed endogenous divergence but fallen short of comparability and generalisability. Nevertheless, rather than addressing urban experiences in the South, the authors have begun by refining the “global regularities” of gentrification to establish the foundation for theory building upon various global urban experiences rather than focusing on any single locality (Lees et al., 2015, p. 6). Thus, the purpose has shifted now to identify the interconnections and commonalities of gentrification across regions, as demanded by the new comparative urbanism as a basis for the relational comparative approach.  How best to theorise “the city” and “the urban” has been raised before in this work (Walton, 1976). The challenge for any future scholarship on comparative urbanism is to move away from understanding cities as discrete, self-enclosed, and analytically separate objects. In other words, the future of comparative studies of cities might rest on pursuing a relational comparative approach to urban studies, one that acknowledges both the territorial and the relational geographies of cities. (Ward, 2008, pp. 407-408) Seeking global regularities, these authors asserted that the globalisation of gentrification has emerged to accompany the globalisation of capital, or more precisely, capital shifting from industries to the secondary circuit of accumulation “in forms of fixed capital and, more specifically, the built environment and urban space” (Lees et al., 2016, p. 45). Class struggle in gentrification, in this case, is treated as an abiding feature of capitalist society. Centring on this proposition, two grounds are claimed for establishing connections between contemporary world cities: planetary urbanisation and planetary rent gaps. First, the authors have attempted to expand the scale of gentrification. Since footloose capital has supported the emergence of multiple reinvestment sites under urbanisation through various types of state or capitalist initiatives, the inner city as a central context has lost its specificity in gentrification studies. A wider scale of capital reinvestment and societal restructuring is now encompassed in gentrification. The examples enumerated include, for example, the ! 42!establishment of a technical corridor in India between Mysore and Bangalore that involves the establishment of new townships (Goldman, 2011), mega-regeneration projects that promote “metropolitan-scale” gentrification (Shin & Kim, 2015), a state-led process for economic restructuring that involves housing formalisation in favelas (Queiroz Ribeiro, 2013), plus a general increase of suburbanisation on previously rural land. Recently, Lees et al. (2016) proposed a thesis for the new economics of gentrification to describe the process of reinvestment, exploitation and accumulation. This thesis advanced Smith’s rent gap theory to accommodate the existing divergences in the land reinvestment process in the South. Based on the Southern experience, for instance, the authors found that multiple actors could capture the capitalised ground rent and participate in a share of the rent gap generated by redevelopment. Additionally, besides economic actors, state and authoritative instruments could impact how rent gaps were created, such as through place stigmatisation based on state-led, discursive practices.  In the end, the comparative approach, in effect, tends towards the establishment of a global theory of gentrification. Moreover, the global theory develops Smith’s structural explanation, in which gentrification is invariably induced by the cycle of uneven development intrinsic to capitalism and promoted by champions of capital accumulation through rent gap exploitation. It is surprising that the global regularities found by the authors have been solely based on the presumed situation of global capitalism in real estate industries, while social and cultural agencies, locally and internationally, are secondary issues.  The above themes, in their aim to frame a global theory, remain at a conceptual stage. The literature does draw attention to powerful globalisation processes that could prompt reform strategies and urban projects at the local level and subsequently add fuel to the presence and permeation of gentrification. However, from both the political-economic and social-cultural perspectives, the analytical frameworks fall short of capturing the entire course of gentrification. As a result, these studies are insufficient in treating the relationship between transnational and local forces as well as globalisation processes and historical trajectories of change in gentrification. For instance, placing only the economic activities of reinvestment at the centre of the conceptual framework for theorisation, those central concepts to gentrification that I identified earlier—class ! 43!dynamics and social injustice—are reduced to an assumed outcome. Further, concerning the concept of global urbanism as a universal force for gentrification, a question remains as to how a gentrifying process can be sustained given the social and cultural settings in cities with a less developed post-industrial economy.  In addition, the structuralist interpretations—initially based on the nature of capitalism, later shifting to neoliberalism, and currently encompassing global capitalism—could be fallacious arguments when transplanted to the South. Although reinforced by neoliberal policies and involving the exploitation of land parcels, the gentrification process by no means fully represents neoliberal transformation or the regulation of a capitalist system in the land market. Accordingly, the existence of so-called neoliberal ideas and policies and land reinvestment do not require the outcomes of class dynamics and struggles similar to those of advanced capitalist society. One needs to be cautious in using the variations evidenced in empirical studies to enrich the twin themes of capitalism/neoliberalism and class struggle. The integrated package of neoliberalism, global capitalism (as manifested in the real estate industries) and gentrification (centring on the subject of class dynamics and struggles in place) does not exist everywhere.  So, why are researchers “looking for gentrification” in China? It can be sketched out as a transitive relationship: since gentrification is understood as a neoliberal urban process, and neoliberalism is understood to be a global system, gentrification must also be occurring globally. Moreover, “just as capital and culture have become quintessentially global, class and politics are also global” (Smith, 2008, p.25). (Ren, 2015, p.220) 2.2.2 The epistemological gap between the North and South  Treating context as the source of variation in the geography of gentrification, many Southern scholars do recognise signs of gentrification on the surface but are unsatisfied with both of the well-established explanations from the post-industrial cities and currently ascribed to globalisation. These authors have uncovered a variety of contextualised differences between regions, requiring new approaches and explanations. This subsection sorts out four sources of pressure on the epistemological gap between gentrification and endogenous processes outside of “the usual gentrification suspects” (Lees et al., 2015, p. 1). The review in this thesis does not offer an elaboration on ! 44!regional variations in gentrification. This would in fact be impossible, as currently, the literature offers vague conceptual understandings and a distinct emphasis in its explanations. This review instead draws together literature that has challenged existing explanations and has added insights to new approaches explaining varied contexts.  State power   The role of the state is a common source used by scholars to justify their alternative interpretations. To a large extent, these interpretations are determined by the condition of Southern cities, where capitalist markets and regulatory systems over properties are far from established. In these cases, state actors may have functionally embarked on an urban trend or managed or intervened in assorted systems of change related to urban renewal, thereby fashioning a mode of spatial and social change. Shin et al. (2016) argued that generally in the Global East, the states have been associated with capitalists and engaged in real estate development as a means of urbanisation, industrialisation and state building. Ghertner (2014) has challenged an agnosticism of extra-economic forces in gentrification pre-assumed in the conventional understanding. In particular, Ghertner’s extra-economic force means state violence.  Harris (2008) compared gentrification in London and Mumbai, both cases being promoted by state imagery of a global city landscape. In Mumbai, think tanks have directly modelled the urban strategies (Mumbai First) of London (London First) to create a receptive location for global business. Politicians and commercial elites in both areas continuously restrain the resistance of union power. However, the gentrification in Mumbai and London resulted in different social outcomes as a result of distinct political and sociocultural conditions, such as different scales of displacement, degrees of class struggle and ways of treating historical structures. Harris (2008) thus encouraged a critical approach to understanding “generalised” gentrification and, importantly, the need to revisit previously “parochial assumptions, practices and language” (p. 2423).  In other Southern metropolises, state weakness can be a pivotal constraint on gentrification. This effect implies that the intention and capability of the national and sub-national state to promote urban transformation and engage with concrete globalisation processes can impact whether the gentrification process is present. For instance, in Cairo, private capital and artists have been arriving in downtown areas with buildings of high ! 45!historical value. However, booming middle-class suburbanisation, urban policies in favour of the suburban real estate market, and the so-called “neglectful” behaviour of the state in governing the urban core and making policy concessions for revitalisation have limited the efficacy of gentrification (Elshahed, 2015). The lack of administrative power at the local level is also found in Vietnam, in particular when compared with the one-party state in the post-reform era of China. This lack has greatly reduced the scale of the redevelopment of even state-owned housing (Yip & Tran, 2016).  Systemic transition A major body of the current literature treats the flexible methods behind land and housing transformation in the Southern cities as their starting point in questioning classical explanations. The literature enriches the trajectories of urban and social change in Southern cities, which are path-dependent on institutional environments and depart from a simple linear economic process. Charlotte Lemanski (2014) introduced the case of a bottom-up process of housing marketisation through a conceptual comparison of gentrification in the North with “downward raiding” in the state-subsidised settlements in South Africa. Downward raiding is a process in which subsidised, low-income homeowners sell housing, often illegally, to businessmen and the wealthier class, who might use these low-value properties to house employees. Technically, the resale process means integrating the subsidised properties into the market system. Ultimately, the resale ends in the displacement and exclusion of low-income sellers, as place and property upgrading begin and the wealthy enter. By comparing the gentrification process with downward raiding, Lemanski proposed a “hybrid gentrification” concept for South Africa, which suggests a blend of the features of gentrification derived from the distant North with the local contexts, in particular “the significance of agency, the role of context and process and the moral/economic status of housing” (p. 2955). By noting the importance of the moral status of housing, Lemanski challenged the “over-reliance on theorising gentrification in terms of the capitalist market” (p. 2956), wherein housing is deemed to have a singular economic attribute. The author also stressed the welfare nature of housing and, most importantly, the unintended results of the permanent exclusion of non-welfare recipients and market buyers limited by affordability, namely, vendors, from homeownership.  ! 46!Due to public land ownership in the post-socialist city of Moscow, private developers must depend on their connections with the city administration to facilitate the progress of gentrification by mediating public land recycling and allocation along with policy changes addressing building conservation in city neighbourhoods. The selection of developers is often based on a closed competition without transparency. This gentrification, in Badyina and Golubchikov’s (2005, p.127) terms, represents an aspect of   “authoritative neoliberalism.” The authors provide the ambiguous conclusion that gentrification manifests “global/local dualism.” The case of Moscow illuminates some of the characteristics of gentrification in post-socialist cities, with a legacy of state planning and intervention in land marketing and housing provision (see also Yip & Tran, 2015). In post-communist European countries, Sýkora (2005) reveals that the large share of housing privatisation in the inner city and the continuance of rental price controls by the local government have prevented the eviction of low-income residents in the inner city. In the same vein, Kovács et al. (2012) showed that the majority of renovation in Budapest’s CBD areas is occurring in traditionally high-status historic areas, where most properties are privately owned and spontaneously renovated by in situ property owners or purchased and renovated by expatriates and well-paid professionals. Only the working-class neighbourhoods that contain dilapidated, public housing are redeveloped through state-facilitated regeneration with a new-build pattern. Gentrification thus has only occurred within the newly built areas.  D. Asher Ghertner is an author on India who feels that gentrification research fails to capture the fundamental difference in the privatisation of space in the South. Ghertner’s (2014, 2015) work has mostly concentrated on the process of land expropriation and development in Indian cities. The author read conventional explanations of gentrification as based on a series of assumptions that may be incompatible with conditions in Indian cities. These assumptions are as follows (2014, p. 1555): 1) “gentrification theory presumes a reinvestment of capital in already once-capitalised urban spaces,” while in Indian cases, the land is often experiencing privatisation for the first time; 2) the concept was born “in advanced capitalist countries with well-established private property regimes,” while land redevelopment serves to establish a private system in Indian cities and represents a thorough transformation; 3) conventional explanations “assume that the ! 47!land in question is converted for a ‘higher and better use’,” which may run counter to the realities in India; and 4) “gentrification is agnostic on the question of extra-economic force,” but state force is the main tool of displacement in India. The urban conditions in India thus led Ghertner (2015) to question whether the abstraction, simplification and conceptual extension of gentrification could cause the overgeneralisation of Indian cases, which would relegate other theoretical inquiries to mere by-products. In response, Ghertner regarded three concepts as being more appropriate to explain the land process of urban renewal in India: urban revolution, land enclosure, and accumulation by dispossession.  An ideal example of a complicated governance process is the redevelopment of informal settlements in Indian cities. Doshi (2015) presented five factors to describe the process of negotiating informality and formalisation through land enclosure in Indian cities. The factors involve not only a developmentalist government and the formation of a new state tending towards the middle class but also an urban poor with their claims and subjectivities who are mobilised to game government policies and cooperate with NGOs to win advantages in the transformation. Based on these empirical materials, Doshi (2015) suggested that gentrification should be “complementary, rather than interchangeable” with other categories of theoretical production and that it should not be an all-inclusive concept “subsuming processes to outcomes” (p.101). This viewpoint supports the stance taken by this study on the relationship between gentrification and other theories from studies on urban transformation in Southern cities. Social and cultural factors The current literature has either de-emphasised the agency of social and cultural actors in less advanced economies or has treated them as an intricate factor likely to impede gentrification. Betancur (2014) found that while a neoliberal regime shift is a common force underlying gentrification in Latin American countries, the reality of slow employment restructuring could set back the presence of gentrification. Commercial and business districts do not yet serve as the command and control centre of the global economy. The new middle class has not reached a volume equal to that in metropolitan cities in developed post-industrial societies. Inner-city settlements continue to absorb a large share of low-paid workers. In this case, manifold limitations render gentrification ! 48!barely sustainable, while local governments must be “active agents turning potential into actual gentrification” (p. 4). Betancur (2014) thus asked for flexible explanations of gentrification via a combination of universal enabling conditions (somehow following Beauregard) and contingent factors.  Other sets of social and cultural factors can also make the gentrification process less than decisive in urban transition. For example, unique property conditions and residential patterns in the inner city might make a location too costly to be redeveloped (Maloutas, 2007, 2014). In Indian cities, caste communities often circumvent the influence of gentrification by creating extremely diverse tenure types and different communities of interest that impose themselves between the political regime and its constituents (Ghertner, 2015). In cities with less vigorous middle-class suburbanisation and those that have never experienced the real decline of central cities, such as Hanoi, scholars might also question whether it is accurate to attribute middle-class settlement in the inner city to a gentrification process (Yip & Tran, 2015).   Class conflicts The final set of studies reveals the ambiguity of social outcomes from gentrification among affected residents. Both advantages and disadvantages exist for residents after gentrification. These divergences partly result from the influences of state force and systemic transition as mentioned above, which bring multiple factors beyond the market into the formation of class-related inequalities in the South. From my reading, the moral base of gentrification is indeed part of its essential nature, distinguishing this conceptualisation from other terms. The working-class experience could be the most debatable component of gentrification studies in the Southern cities because of its complexity and because evidence has been generated that opposes the assumption of injustice in working-class displacement. Meanwhile, new methods should be deployed to solidify evidence for the unquestioned injustice caused by gentrification, beyond uncritical measurement of working-class displacement.  A few studies have shown the scale of class turnover that is far more substantial than that seen from the usual suspects of gentrification. In Seoul, among the 850,000 residents affected by the New Town programme, approximately only 20% of local residents remained in the redeveloped areas (Kyung & Kim, 2011, p. 14). From November 2004 to ! 49!February 2005, approximately 90,000 housing units were demolished in Mumbai (Weinstein & Ren, 2009, p. 423). Between 1995 and 2005, the number of displaced households in Shanghai reached approximately 750,000 (Iossifova, 2009, p.102). Nevertheless, other work has presented de facto contingencies and divergences concerning the social outcomes of gentrification. For instance, Ley and Teo (2014) present a neutral attitude to urban redevelopment in Hong Kong among affected residents, influenced by expectations of monetary compensation or relocation and by an enthusiastic property culture. In Mumbai, Weinstein and Ren (2009) found that during the processes of renewal and relocation residents legally own their dwelling and are resettled on-site, whereas residents who rent and those without complete housing ownership receive no compensation. This finding implies that the existing social divisions are incorporated into reconfigured social relationships and are contributing to the establishment of a new property system in the process of spatial commodification. Similar situations can also be identified in Turkey and Moscow (Islam & Sakizlioglu, 2015; Badyina & Golubchikov, 2005), where tenure groups are differentiated in terms of their compensation, and some individuals are able to gain bargaining power. Islam and Sakizlioglu (2015) stressed that the divergence was fundamentally determined by scattered, ambiguous property rights, which in turn reinforced the discretionary power of state actors.  These works have successfully highlighted the complexity of gentrification. The internal intricacy of varied types of urban and social change has fragmented the power of globalisation. However, meanwhile, the current literature has also conveyed the difficulties facing current gentrification scholars in developing explanations. Causation, meaning the pathway giving rise to gentrification, is constructed in the historically specific and territorial contexts in which gentrification occurs. However, the implementation of urban strategies and projects can be improvised in the Southern cities based on so-called practical knowledge. As a result, easily observed micro-level contingencies and complexities in field studies have generated an impression of a chaotic nature in this process. Finally, I add to the list of difficulties the less emphasised but important issue about the uncertain responses of affected residents in a gentrification project. Social reactions could be at odds with class conflicts and social inequalities ! 50!treated by the literature of gentrification as a main concern Given these epistemological difficulties, the studies presented above often lack a systematic analytical frame and a complete explanation and instead tend to pick up on the peculiarities relevant to the process in a particular locality. Methodologically, these peculiarities are drawn from specific urban projects or cases and based on an individual lens of observation. Because of these variations, these authors tend to resort to micro-level approaches, such as the grounded approach (Islam & Sakizlioglu, 2015) and processual analysis (Doshi, 2015). Ultimately, as the proposals of global theory stop at the conceptual phase without unrolling concrete processes, the fragmentation of the process into individual cases also penalises any potential for conceptualisation or theory building.   The two bodies of work addressing the global nature of gentrification finally highlight a dilemma of the approach to gentrification studies on a global scale. The literature of the North and South as presented in this chapter does agree on the context-dependent nature of gentrification. Meanwhile, empirical studies on the South shatter any illusion of homogeneity in the process and incapacitate a global theory that embraces regional differences; these studies also increase the elusiveness of the concept by overloading it with untraceable variations. With greater scepticism, a basic question ultimately emerges, concerning whether there is sufficient regularity in gentrification to support the establishment of a coherent explanation, even given the complexities and contingencies in different fields.  Starting from this question, I suggest a meso-level approach to theory building for gentrification. It is sceptical both of structural explanation as well as the revelation of unsorted differences, instead aiming to generate middle-range regularities that cohere to context. Foremost, this type of study negates any inevitability in the effects of globalisation or economic determinism on gentrification while also rejecting the amplification of multiple differences. This approach aims to generate regularities at the middle level, where the theoretical range can be defined by contextual homogeneity. Then, in addition to describing the underlying forces, the process of urban and social change is also elaborated upon instead of providing a general description of social ! 51!upgrading and displacement. Finally, it must be noted that a context-sensitive explanation is not necessarily inward-looking. Agents and practices play out at different geographical scales; spatial restructuring has also materialised the processes and ideology of globalisation. The approach in this study is expected to produce well-articulated explanations for types of gentrification in light of both locally situated and cross-boundary knowledge of urban change. 2.3 Gentrification studies in China Chinese urbanisation provides an opportunity for meso-level research in the Global South. However, plenty of literature on urban redevelopment and a rising focus on gentrification have proclaimed the unique features of the gentrification process in Chinese cities. These features aid this research to specify the perspective from which the city of Chengdu will be explained. In what follows I sort out the current literature into three categories. While the first two deal with the cause of gentrification from the political-economic and social-cultural perspective respectively, the third is concerned with the process of residential relocation experienced by affected residents. The study also makes use of studies on urban redevelopment, but the literature is carefully selected and used for avoiding misunderstanding. Most importantly, I am cautious about the multiple types of urban redevelopment, among which not all were accompanied by gentrification (e.g., urban redevelopment in non-working class neighbourhoods).  2.3.1 State and institutional transition  The well-known background for the emergence of urban redevelopment and gentrification in China is a fundamental transition from a socialist redistributive economy to an emergent market economy, and by land marketisation and housing commodification, in particular. As scholars have argued, market-oriented urban redevelopment in China is fuelled by the “re-emergence” (He, 2007, p. 193) and recognition of land values by the local state (see also He & Wu, 2005; Shin, 2009b; Zhang & Fang, 2004). From the 1980s to the early 1990s, urban reconstruction was characterised by the on-site rehousing of inner-city residents and was “still mainly considered to be a social welfare project to improve living conditions rather than as a profitable project” (Wu et al., 2007, p. 239). ! 52!Gentrification emerged in the wave of property-led urban redevelopment, driven by the commitment of the reformist state to housing commodification (He, 2007; He & Wu, 2005).  Based on this understanding, authors have paid sufficient attention to the interaction between state and the market in promoting such a process and, most importantly, the changing role of central and local governments in the transitional economy. Without reaching an agreement, current studies tend to grant the Chinese state as either being authoritarian in its intervention in the societal and market spheres (Shin et al, 2016) or with a neoliberal orientation in economic strategies while showing entrepreneurship among local governments especially (He & Wu, 2009; Shin, 2009). As He (2007) has concluded, the local state has actively mitigated the market risk for developers and created a market-friendly environment for investment, through issuing preferential policies for attracting investment, clarifying property rights, and organising demolition. Shin et al. (2016) have also argued that generally in the Global East, the states have been associated with capitalists and engaged in real estate development as a means of urbanisation, industrialisation and state building. Moreover, for the local state in China, Shin (2009) insisted on the critique that it is fully economistic, even though, for example, the government has attempted to increase on-site rehousing for residents who were affected by housing demolition. The author indicated that to provide more affordable housing on site, the local state has accordingly increased the density of development. Still, solely due to the on-site re-housing policy, residents have become increasingly receptive to urban redevelopment. The purpose remains profit and revenue maximisation, with less attention paid to the residents’ housing welfare.   Hsing (2006), instead, claimed that the local state’s intervention in the land market is more likely a process of “local state building” (p. 587), resulting from interactions between local states and local coalitions engaged in land manipulation. The competition and coalition between different segments of the state also involve a residual role as   “social protectors” (p.577). Within the inner-city brownfields, local governments have to be the intermediaries and contenders with various socialist land masters to launch land-use rights conversion and transaction. Also, there is a dependence of private developers on political connections with the government in accessing land-use rights. The private ! 53!developers who have no access to government agencies have to purchase land through official land leasing, where the price may amount to ten to eight times higher than a negotiated price (p.48). In a word, the socialist and market systems are mutually at work in determining the pathway of land marketisation and its attendant social results. The recent work of Wu (2016) contributes to a thorough rethinking of the justification of gentrification research in China, which is still attributed to the role of the state. The field study is located in a migrant settlement at the urban periphery of Shanghai, which is currently under redevelopment for new industrial and office spaces. In this work, the role of the state departs from the above work towards its effects in controlling the social power of rural-urban migrants in the city through regulating property rights and land development rights. Wu raises the question of how state domination fundamentally implies the distinctive “‘logic’ that leads to demolition and redevelopment” (pp. 635-636). In particular, Wu built the comparison between state-led urban redevelopment in China and Smith’s (2002) idea of gentrification as a global urban strategy against the backdrop of neoliberal state restructuring.  Wu argued that state-led urban redevelopment serves a much wider objective in, for example, economic restructuring, than the immediate purpose of land income. Through controlling self-housing building, state power underlies the degeneration of informal settlements, which makes demolition inescapable (see also Wu, 2009). Through the definition of property rights, the state has tightly managed illegal construction and retained the accountability of who possesses which kind of right to live in the city. The types of state dominance are evidence taken by Wu to challenge Smith’s ideas in gentrification as a global urban strategy, which directs attention instead to the liberal market and the emergence of neoliberal urbanism. Urban redevelopment in China is more a scene within “urban transformation beyond gentrification” (Wu, 2016, p.652), containing the logic of not only upgrading, but also “conversion, and formalisation” (p. 655).   From the political-economic perceptive, the literature has consistently returned to the state level. States fulfil their leading role through manipulating institutional changes in, most importantly, the property right system, land development system, and housing system, so as to further intervene into both the economic and social spheres in urban ! 54!redevelopment. However, as current work has clearly expanded on the relationship between the state and capitalists in the process of land marketisation, it remains unsatisfactory as a complete explanatory framing for gentrification. Wu’s (2016, 2009) work is proactive in this vein, associating state dominance with a substantially unequal power distribution penalising rural-urban migrants living in an informal economic and residential status. However, we should expect a complete picture depicting the range of state-society relations in the process, including both the middle-class consumers and the different working-class groups. Moreover, compared with migrant settlements at the periphery, dilapidated neighbourhoods in the inner city could have a more diversified social composition and distinctive state-society relations stemming from the socialist heritage of the danwei society.  2.3.2 The new middle class in a transitional society  Compared with these political-economic actors, the middle class in China is not granted sufficient agency in the gentrification process by the current literature. Several questions about the motivations of gentrifiers are underlined, which have pressured authors to develop consumption-side explanations. Primarily, the phases and trajectories of employment restructuring induce uncertainty about the source and number of gentrifiers. Zhang et al. (2014) once mentioned the important background information that the large cities in China have not yet become advanced service economies. Contemporary employment restructuring in China is manifested in chiefly the rapid growth of tertiary industries. Ren (2015), from another perspective, emphasised the expansion of China’s new and sizeable middle class and the increasing density of the urban centre in post-reform cities; market-oriented economic reform and urbanisation both exert their force on these trends. It is thus questionable whether the middle-class settlements in the inner city are a result of demographic trends rather than of a gentrification process.  In particular, the growth of gentrifiers’ socioeconomic status is closely connected with China’s economic restructuring from the planned economy to the market economy (He, 2010). Comparing the change of occupational groups in the metropolitan areas and central areas of Shanghai from 1990 to 2000, He (2010) argued that the absolute numbers of managers, senior officers, and professionals in both metropolitan and central areas ! 55!have actually decreased. This phenomenon was due to the reform of state-owned and collective-owned enterprises, and government and party organisations from 1990 to 2000, which resulted in a substantial decrease in not only low-skilled but also high-skilled workers, managerial staff and officers in public sectors (p. 349). Groups of clerical employees and sales and customer service workers, on the contrary, have significantly increased in the metropolitan and central areas of Shanghai. Finally, potential gentrifiers in Shanghai from 1990 to 2000 as defined by He contained managers, senior officers, and professionals and clerical workers. Most of them worked in the public sector (He, 2010, p. 351). In the 2000s, however, an increasing number of residents living in redeveloped neighbourhoods were working in the private sector (He, 2010; He & Wu, 2007).  In addition, is there a demonstrable consumer culture shaping the collective motive of the Chinese middle class in settling in the central cities? Ren (2015) questioned what motives might drive “the amorphous, disjointed, and unstable” middle class” (Zhang, 2010, p.3) in China towards “class conquest of the city” (Smith, 2008, p. 25). This query can be reinforced in the consideration of Wang and Lau (2009)’s investigation in Shanghai. The investigation classified a professional middle class in Shanghai’s inner city into three categories: advanced professionals and managers from overseas, Chinese managers who may possess sufficient economic capital but relatively lower educational capital than the first group, and office workers (p. 60). It found that the professionals and managers may not have been homogenous in their lifestyle. Professionals who use urban amenities frequently may live in the suburbs. Those who live in the inner city may use urban facilities very rarely, but optimise their concern about commuting costs.  Based on this context, scholars in China barely treat the burgeoning Chinese middle-class consumers as a primary driving force of gentrification. Again, the state is called back. He (2007) once mentioned that housing privatisation has evoked the consciousness of homeownership among Chinese citizens. The spreading high-end communities and commercial facilities have directed the affluent citizens’ purchasing behaviours (p. 187). In addition, Wang and Lau’s survey emphasises that the consumption preference of the middle class and upper-middle class can be channelled by the products of real estate companies. For example, the inner-city professionals could indicate explicitly the name ! 56!of the neighbourhoods or the real estate companies when they are asked for their ideal gentrified locations (Wang & Lau, 2009, p.62).   From a sociocultural perspective, systematic analysis is needed of the relationship between the formation of the middle-class and inner-city urbanism in the post-socialist cities of China. Observing the connection between place-making and the formation of middle-class identity, the authors did not pinpoint the special meanings of place-making in the inner-city neighbourhoods, culturally, economically and politically. In addition, noting the role of gentrifiers usually comes after discussing that of the state and developers; rarely has the literature directly analysed the relationship between production and the consumption of spaces.  2.3.3 The rules of residential relocation and the working class  In China, the process of residential relocation in state-facilitated urban redevelopment is commonly known as chaiqian (), which combines the two terms of demolition and removal. Keeping consistent with existing literature, residential relocation in this study covers a wider range of meanings than displacement. Displacement is treated as a principal defining character of gentrification, while another character is social upgrading in place. Displacement means the removal of residents out of place and the disempowerment of them to stay put. Nevertheless, according to the use in the literature on China, residential relocation may cover different situations of the resettlement of a sizable population. Omitting the spectre of displacement, relocation may also include residential mobility based on individual choices, and thereby precludes the condition of disempowerment. For example, this situation could be found in urban reconstruction before the 2000s, for which high-ranking and high-income residents in danweis might have retreated from the inner city to pursue a high quality of life (see Wu, 2004b).  The process of chaiqian is regulated by a set of urban policies of property acquisition, remedial measures for housing losses among affected residents and residential relocation. Also, decision-making on resettling or compensating current residents retains room for individual bargaining with the local government, which easily renders the process pragmatic and long-winded (Dowall, 1993; Wu, 2004b). In this case, groupings of current residents may be confronted with different situations and outcomes ! 57!in and after urban redevelopment and residential relocation (see Shin & Li, 2013; Lin, 2015; Shin, 2016). For example, residents being subject to the socialist welfare housing system may be different from those already housed by the market system. Legal residents do also encounter different housing arrangements from illegal inhabitants in the city (see Shin, 2016; Shin & Li, 2013).  Between 1995 and 2005, the number of displaced households in Shanghai reached approximately 750,000 (Iossifova, 2009, p. 102). Except for sizeable numbers of displacees, adverse impacts on the relocated residents are also stressed, including unfair or obscure compensation practices, hardship imposed by inadequate facilities, long commutes after relocation to the suburbs, and the risk of reduced rental income and unemployment (Wu, 2004bc; He, 2010, 2012; Shin, 2016). Drawing on Harvey, Shin (2016) stresses that China’s land capitalisation scheme is preconditioned by a process involving accumulation by dispossession to achieve primitive accumulation that has “entailed taking land, say, enclosing it, and expelling a resident population to create a landless proletariat and then releasing the land into the privatised mainstream of capital accumulation” (Harvey, 2005, p. 149). Nevertheless, Wu (2004b) also identified both passive and active residential relocations. An active relocation is that due to open-market housing purchase, often by a highly-educated population. The first category of passive relocation is work-unit reallocation. Danwei will purchase the new commodity housing and sell to the employees at a discount. The other passive relocation is faced by residents in the most dilapidated municipal housing in the inner city. The relocation usually occurs because of land use development in the inner city for either infrastructure or real estate projects. Still, there is a disparity among residents in work-units during the relocation process. The work-unit cadres are more likely to choose active relocation than work-unit allocation, while work-unit employees may only rely on allocation. However, it is important to bear in mind that Wu’s survey involves different types of neighbourhood redevelopment and residents with a relatively mixed socioeconomic status, not purely low-income communities targeted by gentrification.  A survey in Shanghai indicated a high degree of satisfaction with their new dwellings from the displaced residents, which could be counter-intuitive (Li & Song, ! 58!2009). The survey involves displaced residents, voluntary movers, stayers and migrants in their dwellings in 2006, who previously lived in the inner city and then relocated to the inner suburbs. There is nearly no difference in satisfaction between the displaced residents and the voluntary movers; both are more satisfied than other groups (pp. 1095-1102). Li and Song have explained this situation by arguing that it is perhaps due to the considerable political and economic strength of municipal Shanghai, which had guaranteed more reasonable compensation and residential relocation (p.1104). Another reason leading to the result could be the details of the investigation, which reflected on satisfaction for physical conditions of dwellings and neighbourhoods, while disregarding employment, broken social networks, the living arrangements of households and the like.  Besides government policies, factors including the tactics of social governance, the rising social participation as well as the bargaining game between individuals and local governments can all impact the gains and losses of community members. Highlighted in recent work by Lin (2015), the rise of incidents of social activism has alerted the Chinese government, making it accord the highest priority to social stability. With case study in Guangdong province, Lin (2015) also reveals profit concession to current residents in both redevelopment projects of work-units and villages in the city, which functions as      “an incentive to encourage them to engage in urban renewal” (p. 865). He (2012) indicated that in the new wave of gentrification since the middle 2000s, residents are tending to pay more attention to not only the right to stay put and reasonable compensation but also to the right to participation. In a historical area in Guangzhou facing commercial redevelopment since 2006, 180 out of 1950 households persistently resisted property demolition and eviction in the subsequent years. In 2011, the resistance achieved particular success when the government agreed to expand the preserved areas and to introduce a pilot plan for self-help redevelopment. Empirically drawing on the same case, Shin (2016) also noted that the compensation methods were varied based on increasing social pressure on the municipal government in Guangzhou. For instance, for public tenants, what initially began as on-site relocation later became an off-site property exchange (5-6 km away from the current site) without tenure change. The relocation method finally ended by encouraging the purchase of price-controlled housing, due to the imperative to win consensus among residents while maintaining social stability.  ! 59!The above studies thus prompt a fine-grained analysis at the different experiences of the working class in the process of residential relocation, notably how they have been understood as “positive” and “negative” for the working class. Moreover, they require a closer look at the policy/decision-making in providing social remedies for the various groups of low-income residents, which should be determined by the ideology of the Chinese state. In addition, the debates do raise questions concerning the examination and explanation of the social injustice of gentrification. The findings vary methodologically in accessing the social outcomes of urban redevelopment and gentrification; these different results might not necessarily erase the exclusionary character of gentrification.  Based on the three aspects of research in the current literature, we should be aware of three central features of gentrification in Chinese cities. First, the Chinese central and local states hold an important role throughout the whole process of spatial (re) production and residential relocation. Second and related, thorough economic and cultural institutional transition codify the processes of redevelopment and residential relocation in the inner city, led by the new market-oriented urban regime. Lastly, the redevelopment and residential relocation led by political-economic agents should have accommodated with intricate social conditions and social dynamics. Substantial socioeconomic and sociocultural transformation determines the attributes of the middle class in China, which renders the sources and motives of gentrifiers more or less indistinct. Still, impacted by reform strategies of housing and other factors, decision-making on residential relocation can produce variable outcomes among diverse social groupings of the affected residents. However, I also underlined that the current studies fall short of a complete explanation of the role of the state played in the social sphere. Also, both the agency of the middle class in socio-spatial upgrading and the experience of the working class in displacement need more systematic analysis.  This research thus suggests an institutional perspective in scrutinising the changing ideology of the state and the way redevelopment and residential relocation are organised in the transitional economy. The institutional analytical perspective departs from conventional production-side explanations for gentrification, which draw on land economics in a capitalist system. Meanwhile, it aims to unravel the social meanings of ! 60!different aspects of institutional change so as to associate state action with social dynamics. According to the three schools of new institutionalism (i.e., rational choice, sociological, and historical institutionalism), institutions capture both formal rules and informal norms (e.g., values, customs and conventions) (Hall & Taylor, 1996; North, 1990, 1991). Institutions are social constructs: they entail values and ideas of how things should be established and arranged, and they illustrate the power relations around participation in decision-making (Hall & Taylor, 1996, p. 940). Over time, institutions have varied effects, creating incentives for some behaviours while restraining others, celebrating some ideas while eclipsing others, empowering some groups while disempowering others. Studies on the aspects of institutional changes are thus anticipated to link the ideas and practices of state actors with the results of behavioural change and asymmetric power distribution among societies.  It must also be acknowledged that in the field of institutional analysis, market transition and social relationship changes have become well-established themes in post-reform China, for example, in the debate over either the persistence of state actors or the empowerment of market actors (typically, Bian & Logan, 1996; Nee, 1989; Nee & Opper, 2012; Parish & Michelson, 1996). Also, a large number of works have explored the characteristics of China’s land, housing and hukou institutional reforms and their influence on social power, though few of these studies have been contextualised for the case of urban redevelopment, let alone gentrification (e.g., Chen, 2008, 2009; Ho, 2001; Hsing 2010; Lin, 2009; Lin & Ho, 2005; Logan et al., 1999; Wang & Murie, 2000; Zhu, 2000; Zhu, 2004).  2.4 Conclusion: A meso-level approach to state-facilitated gentrification in Chengdu  Thus far, the research has brought together three sets of literature to establish a firm ontological and epistemological foundation for this studythe (re) conceptualisation of gentrification, gentrification studies on a global scale, and gentrification and urban redevelopment in China. The most fundamental issue facing gentrification scholarship today, as I have argued, is not the chaos or simplicity of grounded processes, but the conceptual boundary and theoretical strength of gentrification as pertaining to its ! 61!ontology. Confronting the basic problem, this thesis has discerned the connotations of the concept evolving together with the historical geography of gentrification and has ultimately anchored the ontological basis for gentrification to reflect the shift of class power from lower to higher strata in neighbourhood sites. The chapter then moved to the more complicated issue of the approach to theorising gentrification. It questions the competence of explanations found in the global literature that aim at either building global theory or uncovering contextualised differences and suggests a meso-level approach to studies. Finally, with the Chinese literature, the thesis pinpoints three general features of gentrification in Chinese cities, which necessitate an institutional perspective from which the relations between state and society in gentrification will be examined.  The primary contribution of the study to the three bodies of literature lies in a meso-level study on the gentrification process in a Chinese city. It will systematically associate broader social contexts with the spatial manifestations, causes, patterns as well as social outcomes of the process. As Chapter 3 grounds gentrification in the context of post-socialist societal transition in China, the following chapters deal with the socio-spatial upgrading (Chapter 4) and displacement process (Chapter 5, 6), respectively. While Chapter 3 and Chapter 6 map and explain the structural tendencies of socio-spatial upgrading and working-class displacement in the gentrification process, Chapter 5 and 6 investigates the agency of multiple actors and spatial and social processes in neighbourhoods, which will highlight complexities and contingencies.  So the study provides an institutional explanation for the role of the state and its relations with societal sub-groups present in gentrification. In the process of spatial (re) production (Chapter 4), it decodes the cultural messages expressed in the landscape making of the newly built neighbourhoods in the inner city and investigates the policy intervention in spatial commodification and housing consumption. The two aspects of institutional change will then be linked to the social dynamics of the middle-class gentrifiers. In the displacement process (Chapter 5), I explicate the various arrangements of residential relocation and compensation for working-class residents, which are preconditioned by the large systems of property rights and housing supply for low-income residents. It is these arrangements that account for the different experiences and attitudes of different working-class groups.  ! 62!!Chapter 3 Grounding Gentrification in the Large Chinese City  To concretise an understanding of the gentrification process within the context of social transformation in contemporary China, this chapter measures the geography of gentrification in a Chinese city and identifies the dependencies in the process. Two aspects of the social transformation under study are introduced at the very beginning to lay a foundation for comparison with Western-style post-industrial society. This background distinguishes the peculiarity of the socioeconomic restructuring trend in cities within a newly developed economy and the emergence of the new rich and explosion of individual consumption in a transitional society. Then, the study measures and maps the geography of gentrification in the inner city of Chengdu, China, from 2000 to 2010. Finally, this chapter tries to navigate the thicket of correlations to explicate the gentrification process in Chengdu. An index that captures the extent of gentrification is correlated with various attributes and changes in social and physical structures in a locality throughout the 2000s. The statistical analysis indicates three sets of dependencies that condition the presence and shape the patterns of gentrification in the city. These connections will guide an explanation of the causes of this process in the following chapters.   3.1 Social transformation in large Chinese cities   Although there have been various approaches to explanation, the transformation to a post-industrial society is generally treated by gentrification scholars as an important backdrop to gentrification. In the 1960s and 1970s, following the arrival of urban pioneers, who tended to renovate inner-city landscapes and lifestyles (see Ley, 1994; Zukin, 1982), the expansion of high-status service classes, specifically the professional, technical and managerial workers, created a pool of potential consumers of inner-city housing and lifestyles. The backgrounds of post-industrial transformation and middle-class consumer culture demand re-examination in contemporary large cities in China. Two aspects of the background are highlighted: socioeconomic restructuring in the new economy that is characterised by the primary stage of developing knowledge-based ! 63!services; and the rise of the new rich and the reintroduction of individual consumption into personal life, which prompts the formation of middle-class cultural distinctions in a post-socialist society.  3.1.1 Socioeconomic restructuring in the new economy The rise of major cities in China as new destinations of transnational service corporations has been well documented. According to Edgington and Haga (1998), an obvious expansion in the number of Japan-based service companies was found in Beijing and Shanghai from 1985 to 1995, although the numbers in Guangzhou and Shenzhen did not increase proportionally. In an attempt to associate capitalist economic development with the urbanisation process worldwide, Scott (2011) collected materials on the emerging third wave cities following the movement of capitalism towards “a global cognitive–cultural economy” (p. 295). These cities have been well regarded by global capitalists for their business potential. The Chinese cities of Shanghai, Beijing, Chengdu and Chongqing (the latter two are inland cities in western China) are on the list and marked as cities in transformation from a “marginal” status previously to those amongst the new global high fliers. Nevertheless, the transformation towards a quaternary economy is incomplete, and these cities are still dominated by tertiary industries.  Based on the population censuses of 1990, 2000 and 2010, this study traces labour restructuring in the city of Chengdu.3 The employment structure of Shanghai, which is commonly recognised as the most advanced city economy in China, is examined and compared with that of Chengdu. For both cities, datasets on the industrial and occupational populations are generated for the main urban districts.4 In China, the administrative definition of a metropolitan region includes traditional urban districts as well as newly established districts, prefectural-level cities and counties. By confining the analysis to the main urban districts, this study emphasises employment restructuring in intensively urbanised areas within a city-region. Although they are incorporated into the jurisdiction of urban districts, some peripheral districts or counties still retain a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!3 The census data for the occupational and industrial population in 1990 were based on the total population, whereas the 2000 and 2010 census data were both based on a 10% sampling of the total population. In this study, the 2000 and 2010 data have been multiplied by ten to allow estimation for the true populations in the two years.  4 The main urban districts of Chengdu include Jinjiang, Qingyang, Jinniu, Wuhou and Chenghua Districts. The main urban districts of Shanghai include Huangpu, Luwan, Xuhui, Changning, Jingan, Putuo, Zhabei, Hongkou and Yangpu Districts.  ! 64!significant share of agricultural employment. In addition, within a city-region, manufacturing may have been transferred from the main urban districts to peripheral districts. Although there is no unified definition, the main urban districts refer to the commonly recognised traditional districts established at least before the 1990s. The employment categorisation varies with each census. In particular, market reform in China has separated certain sectors from public sectors and transformed them into private sectors. For example, real estate management was combined with public services and residential services in 1990 but was classified separately in 2000 and 2010. Meanwhile, new sectors such as leasing services and business services were established in 2010, and these may combine separate sectors listed in prior census data. In this case, the data of the employed population are re-sorted into a unified classification, thereby allowing comparison between censuses. In particular, the service sector is specialised into distributive services, personal services, producer services, social services and public administration. Distributive services include transport, storage, and postal services, information and communication, and wholesale and retail trade. Personal services combine the industries of accommodation and catering, culture, sports, and entertainment, and residential, repair and other services. Producer services aggregate finance, insurance and real estate (FIRE) services and business and professional services. Social and public services contain sectors in education, health care, social insurance and welfare along with public administration and social organisations.5 From 1990 to 2010, the main urban districts of both Shanghai and Chengdu transitioned from manufacturing to service industries (Figure 3.1, 3.2). The number of those employed in manufacturing sectors declined from 54.44% of the employed population in Shanghai in 1990 to 17.40% in 2010, while the change in Chengdu was smaller, from 35.35% to 15.40%. In contrast to the changeover seen in manufacturing, service labour increased in all sectors except for social services in the two cities. By 2010, distributive services captured the largest percentage among the four types of the service sector, amounting to more than one-third of the total employment in both cities. Although producer services achieved the highest growth rate among all service industries across the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!5!The four categories are based on the classification of the Organisation of Economic Cooperation and Development; the minor categories within the four sections are based on the International Standard of Industrial Classification.!! 65! Figure 3.1 Employment trend by sector in Shanghai 1990-2010. Sources: Computed based on Population Census Office of Shanghai [PCOS], 1990, 2010, and Population Census Office of the State Council [PCOSC], 2000.   Figure 3.2 Employment trend by sector in Chengdu 1990-2010. Sources: Computed based on Population Census Office of Sichuan Province [PCOSP], 1990; PCOSC, 2000; Population Census Office of Chengdu [PCOC], 2010. 54.44 32.31 17.40 16.15 25.07 34.30 4.87 10.08 13.42 3.47 8.80 16.38 1.25 5.11 8.60 11.86 12.02 11.37 0 10 20 30 40 50 60 1990 2000 2010 Percentage (%) Manufacturing Distributive services Personal services  Producer services FIRE Social services  35.35 26.41 15.40 15.02 25.24 36.84 5.55 9.20 12.97 3.67 6.89 11.52 0.91 3.41 6.89 12.89 12.45 10.17 0 5 10 15 20 25 30 35 40 1990 2000 2010 Percentage (%) Manufacturing Distributive services Personal services  Producer services FIRE Social services  ! 66! twenty years, the total labour force in these high-skilled industries accounted for less than one-fifth of total employment in both of the cities in 2010. In Shanghai (16.38%), the number employed in producer services exceeded that of personal services (13.42%) and social services (11.37%) in 2010, while in that same year, producer service employment (11.52%) remained slightly lower than personal service employment (12.97%) in Chengdu. Compared with Hamnett’s (1986, 2003) observation of Greater London in 1998, when financial and business service workers reached one-third of all employees, their counterparts in the two large cities in China were still much fewer in number in 2010. According to Sassen (1991), the FIRE group represented 29.8% of the total employment in New York City in 1977 and 28% of that in London in 1971 (p. 132). Hence, despite showing strong momentum, the knowledge-based service industrial jobs remain of modest size in large Chinese cities.  Turning specifically to the five main urban districts of Chengdu, the sum of employment almost doubled from 1990 to 2010 (Table 3.1). However, the ratio between the employed population and the population aged 15 and over actually declined, from 71.4% in 1990, to 57.7% in 2000, and to 54.9% in 2010. In 2010, the five traditional urban districts of Chengdu were inhabited by more than 4.5 million in the population aged 15 years and over, of which the employed population reached 2.5 million. The increasing unemployment rate needs to be explained by referring to labour market reform during China’s transitional period. Based on the labour allocation system designed in the planned economy, local governments attempted to guarantee full employment through the straightforward assignment of students to state sectors after graduation. The labour system-generated redundant labour and low labour mobility, which ultimately reduced firm productivity (Cai et al., 2008). In accordance with enterprise reform, labour market reform enabled managers to gain greater autonomy in both labour recruitment and wage setting. Increasingly, system-generated privilege has been dismantled, and potential employees must compete in a flexible market (Meng, 2000; Tomba, 2002; White, 1987, 1988; Yueh, 2004). The reforms brought extensive layoffs in departments and enterprises that were subject to restructuring and increased the risk of unemployment in the city. Since the 1990s, and in particular after the issue of the labour law in 1994, the! 67!Table 3.1 Employment structures by sector in Chengdu, 1990, 2000 and 2010   1990 2000  2010  Change 1990-2010   No. % No. % No. % No. % Farming, forestry, animal husbandry and fishery 210235 15.86 130820 8.15 19080 0.76 -191155 -90.92 Mining and quarrying 4939 0.37 2600 0.16 8280 0.33 3341 67.65 Manufacturing 468594 35.35 424120 26.41 384360 15.4 -84234 -17.98 Utilities 34103 2.57 64580 4.02 41090 1.65 6987 20.49 Construction 107853 8.14 104620 6.52 248560 9.96 140707 130.46 Distributive services 199141 15.02 405290 25.24 919460 36.84 720319 361.71       Transport, storage, and postal services;          information and communication  62885 4.74 103720 6.46 245040 9.82 182155 289.66       Wholesale and retail  136256 10.28 301570 18.78 674420 27.02 538164 394.97 Personal services  73522 5.55 147800 9.2 323780 12.97 250258 340.39        Accommodation and catering 21832 1.65 83920 5.23 162980 6.53 141148 646.52        Culture, sports, and entertainment 12113 0.91 21670 1.35 35750 1.43 23637 195.14        Residential, repair and other services 39577 2.99 42210 2.63 125050 5.01 85473 215.97 Producer services 48697 3.67 110570 6.89 287530 11.52 238833 490.45        Finance and insurance 9607 0.72 33550 2.09 84500 3.39 74893 779.57        Real estate 2423 0.18 21270 1.32 87360 3.5 84937 3505.45        Business and professional services 36667 2.77 55750 3.47 115670 4.63 79003 215.46 Social services  170871 12.89 199900 12.45 253880 10.17 83009 48.58       Education 75895 5.73 79860 4.97 106680 4.27 30785 40.56       Health care, social insurance 35103 2.65 47070 2.93 59710 2.39 24607 70.1       Public administration and social organisations 59873 4.52 72970 4.54 87490 3.51 27617 46.13 Other services 5535 0.42 3950 0.25 9690 0.39 4155 75.07 All others 2010 0.15 8940 0.56 0 0 -2010 -100 Total employed 1325500 100 1605790 100 2495710 100 1170210 88.28 Population aged 15 and over 1855747  2782110  4543200  2687453  Note: The census data for the industrial population in 1990 were based on a total population enumeration, whereas the 2000 and 2010 population were estimated by multiplying the census data by ten. Sources: Computed based on PCOSP, 1990; Population Census Office of the State Council, 2000; PCOC, 2010.  ! 68!unemployment rate has been annually increasing. Based on the population accepting the Subsidised Minimum Living Standard, at least 22 million households and 6% of urban citizens may have suffered from underemployment in 2004 (Cai et al., 2008).  The two decades under study witnessed roughly a reverse between the development of the primary and secondary versus the tertiary and quaternary sectors in the city. Whereas in total the labour force in the four categories of services expanded from 37.14% in 1990 to 71.51% of total employment in 2010, the total number of those employed in primary and secondary industries shrank from 62.3% to 28.1%. The change indicates that within these two decades, an additional 1,292,419 people either transferred their job out of manufacturing or newly arrived to the city and joined various services. This number is much larger than the 275,389 workers who left manufacturing and agricultural jobs. This asymmetric change implies a tremendous influx of labour into the city as a consequence of urbanisation. In 2010, service jobs reached more than 1.7 million in the city out of an employed population of 2.5 million. Producer service workers in Chengdu in 2010 represented 4.9 times the number in 1990, which is faster growth than the other three categories of service industry in the city. This multiple is equivalent to an added labour force of 238,833 persons. Within producer services, the development of real estate management has been the most rapid, increasing by more than 35 times the number of jobs in 1990, followed by still remarkable growth in the financial and insurance industries (7.8 times) (Table 3.1). However, together these two sectors represented only 6.89% of the entire labour force in Chengdu (approximately 170,000 employees), compared with 8.6% in Shanghai (279,220 employees) in 2010 (see Figure 3.1). It is noteworthy that although lower in scale, producer services in Chengdu grew faster (4.9 times) than the same sector in Shanghai (2.7 times). Considering the industries within services, personal services have rapidly expanded, increasing by 3.4 times from 1991 to 2010. However, the proportion of public service workers has consistently decreased from 12.89% of the entire employed population in 1990 to 10.17% in 2010 (Figure 3.1). This result is consistent with He’s (2010) findings for Shanghai, which attributed the result to the economic reform that downsized public sectors and state-owned and collective-owned enterprises during the 2000s. The labour force generated by distributive services represents the largest portion of local employment in    ! 69!Figure 3.3 Employment trends by occupation in Shanghai 1990-2010. Sources: Computed based on PCOS, 1990, 2010, and PCOSC, 2000.   Figure 3.4 Employment trends by occupation in Chengdu 1990-2010. Sources: Computed based on PCOSP, 1990; PCOSC, 2000; PCOC, 2010.    4.67 4.45 8.53 18.24 19.22 22.47 8.51 18.43 18.38 17.89 28.76 35.93 50.24 28.91 14.46 0 10 20 30 40 50 60 1990 2000 2010 Percentage (%) Managers Professionals Clerical workers Sales and service workers Production workers 4.13 4.54 3.89 18.40 17.72 18.90 7.34 11.78 13.03 16.95 29.97 42.01 37.39 27.41 20.81 0 5 10 15 20 25 30 35 40 45 1990 2000 2010 Percentage (%) Managers Professionals Clerical workers Sales and service workers Production workers ! 70!Chengdu (36.84%), and a majority comes from wholesale and retail industries.  The occupational structure depicts more particularities in the socioeconomic classification in China, derived from not only the status of industrial transformation but also from the economic transition. The share of professionals and managers in urban society in both Chengdu and Shanghai remains far more consistent. Specifically, in Shanghai, when the share of low-paid production workers shrank from 50.24% of total employment in 1990 to 14.46% in 2010, together the ratio of managers and professionals increased from only 22.91% to 31.01% (Figure 3.3). In comparison, the share of production workers in Chengdu declined only 17% (from 37.39% to 20.81%), whereas the proportion of managers and professionals was maintained at approximately 22% throughout the two decades (Figure 3.4). In contrast, Chengdu witnessed notable growth in the proportion of low-paid sales and service workers, which grew from 16.95% in 1990 to 42.01% in 2010. Comparably, in Shanghai, low-end service workers (from 17.89% to 35.93%) also outnumbered managers and professionals in 2010. Table 3.2 shows that in absolute numbers, managers and professionals in 2010 expanded by 77.39% and 93.47% of the 1990 population in Chengdu, respectively, equal to 42,348 and 227,933 people. By 2010, the two highly ranked occupations reached 568,870 out of the 2.5 million workers in the main urban districts of Chengdu. Nevertheless, among low-paid workers, only agricultural workers experienced a reduction in numbers throughout the twenty years across all urban districts of Chengdu. Although its share among all occupations was decreasing, the population of production  workers was still greater in 2010 than in 1990. The number of sales and service workers more than tripled from 1990 to 2010, in line with the change in the number of clerical workers to over twice the number in 1990. In 2010, low-paid service workers accounted for the largest labour pool in the city, totalling over 1 million. The rapid growth in low-skilled sales and service workers may have partially derived from the increase in white-collar workers, who in turn have generated demand for personal services; a similar relationship has occurred in advanced post-industrial cities. However, in Chengdu, it is more likely to be a result of economic strategies to encourage commercialisation and consumerism and the influx of a large number of poorly educated migrants to the city from rural areas. ! 71! Table 3.2 Change in the employment structure of Chengdu by occupation 1990-2010  The evidence shows a lower growth rate for professionals and managers in Chinese cities from the 1990s to 2000s than was seen in advanced industrial economies in the 1970s. Ley (1980) revealed a growth rate of 30% for professionals and 65% for managers in British Columbia within only five years, from 1971 to 1975. He (2010) also showed that the increase of professionals and managers in terms of both number and ratio in Shanghai from 1990 to 2000 was much smaller than that for their Western counterparts in the 1970s and 1980s. Notwithstanding, it should also be noted that in 1990, the percentage of professionals and managers in Chengdu (22%) already exceeded the level in the province of British Columbia in 1975 (20%).  This situation represents the influence of enterprise reform on socioeconomic restructuring during China’s transitional period. Since the late 1980s, enterprise reform has encouraged the decentralisation and marketisation of enterprise operations. It has resulted in the privatisation or dismantling of small-scale, state- or collective-owned firms and the development of private companies in both urban and rural areas (White, 1993; Nee & Opper, 2012; Naughton, 1994). The corporate restructuring was accompanied by the job transfers of individuals in former state sectors and an enlargement of the labour force in private sectors. The number of employees in state-owned enterprises dropped by 49 million, from 113 million in 1995 to 64 million in 2004 (Cai et al., 2008, p.1 76). Fundamentally, the new labour market redistributes life   1990           2000 2010 Change 1990-2010 No. % No. % No. % No. % Administrators and Managers 54722 4.13 72950 4.54 97070 3.89 42348 77.39 Professionals and technicians 243867 18.4 284500 17.72 471800 18.9 227933 93.47 Clerical workers 97336 7.34 189120 11.78 325230 13.03 227894 234.13 Sales and service workers 224617 16.95 481240 29.97 1048360 42.01 823743 366.73 Agricultural workers 206353 15.57 130540 8.13 16670 0.67 -189683 -91.92 Production workers 495647 37.39 440070 27.41 519240 20.81 23593 4.76 Others ----- ----- 7370 0.46 17340 0.69 ----- ----- Total employed  1322542 100 1605790 100 2495710 100 1173168 88.71 Note: The census data for the industrial population in 1990 were based on a total population enumeration, whereas the 2000 and 2010 population were estimated by multiplying the census data by ten. Source: Computed based on PCOSP, 1990; PCOSC, 2000; PCOC, 2010. ! 72!opportunities among the employed population, whose position and salary have been more conditioned by personal skills, education, labour productivity and the economic performance of an enterprise (Cai et al., 2008; Tomba, 2002; Yueh, 2004). Based on this background, high-ranking employees affiliated with the old system might initially have experienced profound career change, being either filtered out of or developed as members of the new sectors or benefiting from participating in the dual system, for example. The turbulence on the job market, reflected in employment retrenchment and subsequent labour mobility, could increase or decrease the presence of highly educated workers in the major urban districts.  The stage of industrial transformation, accompanied by the unique history of economic system reform, is important contextual information, as it could generate contradictory socioeconomic dynamics with the manifestations of gentrification.  Undoubtedly, the large Chinese cities achieved overwhelming growth in producer service industries over the past two decades. Meanwhile, manufacturing closure lays the foundation for the working-class retreat from the inner city. However, high-skilled services expanded simultaneously with low-skilled services. Compared with the global cities in the West in the 1970s the knowledge-based service industries in cities are in a less dominant position. Socioeconomic realities may challenge the specific segment of high-end service classes that comprise the gentrifiers in China. Economic reform further complicates the mode of socioeconomic restructuring. Various socioeconomic groups have been experiencing a process of re-stratification along with a transition from a planned production system to a market-oriented system and the subsequent reconfiguration of job opportunities among individuals. The socioeconomic re-stratification in progress may generate inquiries into the social composition of the middle-class gentrifiers and the source of their collective motives in inner-city resettlement. These questions lead this study to stress the second context of gentrification, with regard to cultural transformation and middle-class formation in post-socialist society.  3.1.2 Individual consumption and the middle-class formation in the transitional society The cultural tide of post-modernism is another critical backdrop to gentrification in post-industrial cities. The classical gentrification model is bred in a liberal ideology of the ! 73!cultural innovation of urban lifestyles (Ley, 1980, 1996, 2012). Caulfield (1994), who has meanwhile questioned the use of the term gentrification, deemed middle-class resettlement in the older inner city as a critical social practice of middle-class agents. It contributes to the creation of an emancipatory city and aims to compete with the dominant lifestyle characterised by a standardised corporate form and suburban lifestyle. As reviewed in Chapter 2, Ley (1996) analysed urban pioneers and the new middle class’s engagement with historic preservation, neighbourhood development and progressive political reform from the 1960s to the 1970s, contrasting this trend with the economistic values and mass culture under Fordist mass production. Zukin (1982) scrutinised the emergence of loft living in Lower Manhattan in New York City, with an intricate thesis about the relationship between cultural production initiated by artists at the outset, followed by capital accumulation through re-creating built environments by subsequent developers and the state in a period of deindustrialisation.  The above literature highlights the cultural politics of pioneer-led gentrification at the early stage in European and North American cities. In China, pioneer-led cultural innovation has been limited. Instead, the state is the primary actor launching large-scale urban redevelopment and gentrification. Moreover, the state-facilitated urban redevelopment programme has been integral to the reinvention of urban modernity and a new national policy to explicitly move the economy from production to consumption and services. This background – the state-led reinvention of urban modernity and consumption promotion – has impacted the formation of consumer culture and laid the foundation for the emergence of a consumer class. This, in turn, could rewrite the social and cultural implications of one type of elite-oriented gentrification in China as compared with the pioneer-led gentrification in the West.    Unlike post-industrial contexts, the relationship between housing consumption and class formation and division in post-socialist China is pertinent to the sociocultural characteristics of the socialist society and the composition of the new rich, which has been impacted by economic reform. First, urban society under Mao’s China lacks a clear-cut middle-class division of lifestyle and identity upon which to specify the fragmentation and reformation of classes in the wake of market reform. A lifestyle characterised by communal consumption and collective living between cadres and workers in danwei ! 74!compounds established an egalitarian base where social divisions were rarely embodied in the distinctions of habitus. Zhang (2010) explained that the so-called egalitarian society was precisely manifested in homogenous living patterns, in low salaries and in an underdeveloped consumer society in socialist China. As a consequence, the socialist society in China presented no clear classification of lifestyles and political attitudes, although it did have an explicitly hierarchical system of occupations. Within this historical social context, contemporary class society in China thus features the creation and circulation of cultural distinctions among social groupings.  Second, the pathway of economic reform has generated social differences among the new middle class in Chinese cities. The result can be traced back to Deng Xiaoping’s middle-class politics, which aimed at facilitating reform; this area has been largely underdeveloped by urban researchers compared to its economic counterpart. At the very beginning of reform, a basic national policy was proposed, which has to date served as a constant principle leading China’s social development objectives: to establish a moderately prosperous (Xiaokang, ) society. Together with another two important public policies at the central level, rural and urban household registration and the One Child policy, these national policies have deeply impacted the trajectory of social change in China. Deng’s description of the “moderately prosperous society” is in contrast to the affluent society of the developed countries and to Mao’s common prosperity (Datong, ) society. Deng was concerned that Chinese society, within the medium term (say by 2050) could by no means achieve the quality of life seen in advanced societies, but it could aim to be a middle-class society (Deng, 1983). This policy is thus a moderate revision of the bold “four modernisation” objectives offered by Mao (see also Li, 2003; Lu, 2010; Tomba, 2004). While Mao’s Datong society presented his idealist imaginings for a communist society, Deng was absolutely realistic in recognising the impossibility of achieving the objective of common prosperity based on the economic level of pre-reform China. Another difference between Datong and Xiaokang is that while Datong society advocated for an absolutely public notion of ownership and distribution, Xiaokang society recognised the necessity of individual wealth, private assets and stress on family values. According to Lu (2010), Deng considered this concept to be much more ! 75!acceptable to the masses, and it could be used to deliver market principles to the masses in accordance with economic reform.  Based on this background, “get rich first and achieve common prosperity later” was promoted by Deng in the 1980s as a realistic path to economic transition and Xiaokang society in the short term, with the ultimate goal of achieving common prosperity (Yan & Liu, 2002). Significantly, the strategy promotes the formation of specific groups in particular sectors and regions, who would benefit from the reformist policies. These favoured groups are the so-called “new rich” and “the newly emerging middle class” in China (Yan & Liu, 2002). This economic reform generated various sources, formally and informally, of affluence. As mentioned in the last section, the pre-reform elites can be separated by the direction of their social mobility within the processes of job change, such as either consolidating an elite social status or fading out in competition. Walder (2003) suggested that the opportunity structure for pre-reform elites varies for different transitional economies and is determined by, first, the divergent trajectories of political reform and then the systematic reform of the public ownership of assets. The peculiarity of reform in China lies in that hitherto, the Chinese authority has shown no interest in political reform and has maintained mixed types of ownership and relations of production. Different economic systems and workplaces impact the path of success, political ideology, wealth accumulation and lifestyle (Lu, 2002, p. 265), which creates social differences within the new middle class. The most salient case is that the first generation of the new rich in China is not composed of well-educated and skilled professionals or managers but instead of petty businessmen and private entrepreneurs who are relatively lower in educational attainment (Lu, 2002). Still, a danwei with stronger political and social ties can create better life opportunities for its employees. Managers in large danwei can possess substantially more wealth and social prestige than those at small danwei because of their differential power in acquiring resources. The “cadre-entrepreneur,” who occupies a high-end position in both private firms and public administrations, is particularly a social construction of the transitional economy (Nees, 1991, p. 269). As a result, the new middle class in China lacks internal recognition, in particular, recognition embodied in the uneven distribution of educational attainment and cultural character (or symbolic capital).  ! 76!In contemporary China, the conceptualisation of class itself—especially what is middle class—is under debate. The “new rich,” or those with wealth sufficient to allow disposable income, often have Party-state connections, so a liberal democratic concept of class does not reliably map on to the Chinese case. (Cartier, 2009, p. 373) Meanwhile, the concept of a moderately prosperous society places family happiness, self-interest and material well-being at the core of individual achievement, rather than public concerns, political participation or human development (Anagnost, 2008; Lu, 2010). Lu argued that in a society where policies release enriched opportunities and diversified channels for pursuing affluence, with “everyone caring about themselves, people want more than what and how they can achieve it, while not [being concerned with] what the others get and if it is fair” (Lu, 2010, p.114). Accordingly, post-reform social policies have transitioned from representing the liberation of productive forces to representing the liberation of consumption forces to expand the middle-class foundation of the Chinese Communist Party (CCP) (Tomba, 2004). Tomba (2004) showed that in the second half of the 1990s, salaries for professionals in the public sectors rose by 168%, which is 40% higher than the average. Tong (1998) revealed that average household incomes in Shanghai, Guangzhou and Beijing more than doubled in two years from 1993 to 1995, whereas in the late 1990s, it was expected that up to 0.3 million workers would become redundant in a year in Shanghai (p. 33). Concerning the cultural formation and social mobility of the middle classes, Zhang (2010) thus argued that in the reform era, the formation of a middle class is “emerging, fragmented and precarious” in China (p. 6). As one uniquely social characteristic entrenched in the transitional period, conspicuous consumption practices, rather than the latent means of production, become empirically necessary and viable for class analysis in China. Scholars have underlined a twofold influence of private housing consumption on middle-class formation: it mobilises both the cultural (re) production of middle-class distinctions and the spatialisation of class distinctions and privileges. Zhou and Chen (2010), for instance, argued that in conjunction with the collapse of socialist danwei institutions, private housing (and transportation) “not only represents consumer goods with which they can build their self-identity and win social recognition but also practice fields for moulding new notions of consumption” (p. 94). Employing a top-down perspective, Anagnost (2008), for instance, deemed the state promotion of consumption practices to be a type of “national project on ! 77!a cultural form” (p. 497). The project conforms to the CCP’s “social engineering” (Sigley, 2006. p. 495, quoted by Anagnost 2008, p. 498) with the aim to “expand the middle class by inciting aspiring individuals to adhere to new social norms of middle-class identity often defined around consumer practices” (p. 498).6  From the perspective of consumers, the newly built communities are arguably fields that produce new social and cultural norms. Living in a gated community as a homeowner has become an approach taken by consumers to not only embody their desire for the good life but also to invest in themselves by joining with people from the same strata and establishing self-conscious consumer citizenship as the ultimate declaration of social status (see Davis, 2006; Pow, 2009; Pow & Kong, 2007; Ren, 2013; Zhang, 2010). Further, the gated community turns out to be a decisive, practical field in which residents form new patterns of collective interests, mobilisation and consequential collective conflicts (Tomba, 2005; Zhang, 2010). In contrast to the socialist society that standardised the lifestyle of citizens across socioeconomic statuses, the cultural force driving the contemporary formation of middle-class distinctions in urban society is founded on the development of a private lifestyle and individual consumption. Moreover, the social differences existing among the new rich as generated in the transitional economy could arguably strengthen the role of cultural capital in class formation. Gentrification in China is thus occurring in a society wherein so-called middle-class distinctions are emerging and forming. Whereas classical pioneer-led gentrification in Western cities conveys a cultural claim by a new middle class standing opposite to suburbanites, state-facilitated gentrification in China, through the massive demolition and reconstruction of urban places and the accompanying social reorganisation in place, can be the very force driving class formation. The inference suggests that to explain gentrification in China, one should not simply consider the process as being an interaction between two established classes in place. Instead, the process includes a highly dynamic social change that, in particular, reflects the relationship between inner-city urbanism and class formation, replacement and displacement.  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!6 Social engineering is conceptualised as constructing xiaokang shehui in China. Xiaokang shehui means the “relatively well off” society (Anagnost, 2008, p. 502). The term originates from the description of an ideal society in the Confucian Book of Rites (Liji), first referenced by Deng Xiaoping in 1978 and then recovered by Jiang Zheming. Source: Li, S. E. (2003). Lun quanmian jianshe xiaokang shehui (On Building a well-off society in an all round way). Seek Truth From Facts, 1,13-16. ! 78!3.2 The geography of gentrification in Chengdu, 2000-2010    The choice of indices to measure gentrification has varied with the literature and has been impacted by the working definitions of gentrification employed by scholars. Analysts can select specific indices and measuring methods based on their theoretical perspective and research purpose (Walks & Maarance, 2008). For instance, Ley (1996) defined the socioeconomic upgrading of neighbourhoods as a decisive factor marking the existence of gentrification and used changes in occupational and educational status as two main measures (see also Davison & Lees, 2005). Following Smith’s (1996) production-side explanation of gentrification, authors adopted indicators representing reinvestment as an effective measure of gentrification. Wyly and Hammel’s (2004) work, for example, was assisted by household-level mortgage data in New York City. More generally, changes in land and building values have often been used to identify the existence of gentrification in a neighbourhood (Clark, 1988; Lopez-Morales, 2011; Badcock, 1989) Other studies have used mixed methods and indices to improve their accuracy in identifying gentrified neighbourhoods. For instance, Walks and Maaranen (2008) included indicators measuring both socioeconomic upgrading and capital investment. Lees (2003) combined ethnographic fieldwork and statistics when generating evidence for super-gentrification led by financiers in New York City. Wyly and Hammel’s (1999) method also drew on literature reviews, field research and a multivariate discriminant analysis of the socioeconomic characteristics of neighbourhoods. The discriminant model was particularly effective in supporting the authors’ studies of gentrification in an extended number of cities in the United States.  This study takes the position that the most solid conceptual foothold for gentrification is the upgrading of places in terms of social class. However, in practice the measurement of social class in China is still problematic. As mentioned earlier, due to the fundamental reform of the labour force allocation system into a competitive labour market, the unemployment rate throughout the city continued to soar from the mid-1990s; meanwhile, the employed population has been restructured across new emerging sectors based on a market-oriented production system and among old sectors based on a planned system (Cai et al., 2008). As a result, the percentage change in professionals and ! 79!managers in large Chinese cities is much lower than that seen in advanced economies since the 1970s. Labour market reform may cause underestimation of the degree of change in the socioeconomic population of a neighbourhood. Moreover, the internal social differentiation in the status of employees with similar occupations in companies or organisations in terms of, for example, educational level, income, values, etc., increases uncertainty in defining the property and size of the middle class in China (Goodman, 2014). A direct impact of these problems is that researchers have deployed diversified indicators and gained highly divergent estimations of the volume of the middle class in urban China. For example, a measurement according to indicators of consumption level estimated the middle class as representing 54% of the overall urban population in 2012 (Mckinsey, 2012), whereas a prediction based on occupation and income estimated the middle class as only 23% of the urban population in the mid-2000s (Lu, 2007). When an indicator capturing self-identification is added, the number continues to fall (Li, 2006; Zhou, 2004).  Considering the essence of gentrification, one main purpose of the measurement in this study is to compare the degree of social structural upgrading in place. Unlike the above literature, this work does not aim to accurately estimate the size of the Chinese middle class. In effect, this study confirms that the attributes of the so-called middle class are essentially unreliable at this stage, being subject to varied constructions and self-ascription. Further, the concerns mentioned above do not necessarily reject the use of the ratio change of professionals and managers as an important index of social upgrading in neighbourhoods.7  Thus, following Ley (1996), this study continues to adopt the two main indices of change in social class: the change in the location quotient for those local residents over 6 years of age with university degrees and the change in the location quotient for professionals and managers in the sub-district.8 However, it is worth remembering that, by creating a gentrification index based on these two indices, this study offers a relatively conservative estimation of the volume of gentrifiers.  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!7!In effect, income can be an optimum index of social upgrading in China. However, unfortunately, the census data do not offer any information about personal or household incomes. The other social survey datasets that cover income do not release geographical information on households.!8 The age group (i.e., residents over 6 years old) to which the ratio of university population is measured follows the standard of the Population Census of China. !! 80!Compared with the relative percentage change, change in the location quotient (LQ) can be more effective for measuring gentrification in China (see also He, 2010). The location quotient index expresses the comparative advantage of a place in containing a population, relative to the average level of a larger territory. The index of changing location quotient thus essentially reflects the difference in the concentration of a social class in a place throughout a period, while the relative percentage change re-states that change against the base at the start of the period for that location. The former can avoid the information bias caused by rapid demographic densification and the turbulence in the labour market throughout the entire city. With census data for 2000 and 2010, this study has created the location quotients for the population with a university degree and professionals and managers in each sub-district of the inner city, which is smaller than an urban district and larger than a neighbourhood. All of the location quotients reflecting changes in the inner-city sub-districts are compared with changes in the five main urban districts of Chengdu established since 1990. The location quotients for the two populations are calculated by dividing percentages for a sub-district by percentages for the main urban districts. The change in the location quotient is estimated by directly subtracting the value in 2000 from that in 2010. A gentrification index is then created based on the arithmetic mean of the two change indices for the location quotients from 2000 to 2010.  Based on the quantitative results of this index of gentrification, I then verified the gentrified locations through field observation of the visible landscape of renovation and reinvestment. Moreover, using the yearbooks that documented the sub-district changes, this study excludes factors that could have caused bias to the quantitative data. First, the study adjusted the educated population when the outmigration of universities and colleges in the sub-districts caused a sudden decrease in the population with a university degree during the census interval (Jianshelu, Wangjianglu).9 The portion of change caused by campus outmigration is omitted from the gentrification index. Then, I focused attention on sub-districts where most of the land parcels have been subject to commercial !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!9!Based on the Annual Report of Chengdu, I first noted the number of students at university in a sub-district in 2000 and the number of graduate students. I then subtracted the two numbers from the total population and the population with a university degree in this sub-district respectively. Finally, the two estimated values were used to calculate the ratio of the population with a university degree in the sub-district in 2000. This ratio was compared with the 2010 ratio in this sub-district.   !! 81!use (Yanshikou) and sub-districts where upgrading was due to a large proportion of new greenfield construction in the inner city (Shuangnan). These areas may necessitate data adjustments for certain variables of the two cases, which are anticipated to have a significant influence on the correlations analysed in the next section. Finally, concerning the gentrified neighbourhoods identified in this study, I also confirmed the reasons for upgrading from the archives to avoid being misled by the factors presented above.  3.2.1 The extent of gentrification Table 3.3 lists the population with a university degree and senior employees in 31 sub-districts in the inner city of Chengdu. In the last column, the composite change over location quotients indicates the magnitude of gentrification in these sub-districts in descending order. Two contextualised characteristics depict the distinctive urban condition of the inner city of Chengdu upon which gentrification occurred. First, during the 2000s, the inner city of Chengdu retained its prominence in housing the middle class. In both 2000 and 2010, the proportion of population with a university degree in the inner city was higher than it was in the entire urban district (LQ=1.06 in 2000 and 1.22 in 2010). Comparable results ar