The Meaning of “Private Governance” in Urban China: Researching Residents’ Preferences and Satisfaction

ABSTRACT The existing literature suggests that gated communities and master-planned estates represent a form of private governance. They are a response to the state’s retreat from providing public goods. In post-reform China, newer residential development projects also take the form of gated neighbourhoods. To what extent do these neighbourhoods represent a new form of “private governance”? Results from the perspective of residents reveal that the meaning of “private governance” refers to services delivered by private organisations rather than self-governance. Gated neighbourhoods in urban China are a response to the state’s intention to retreat from serviced land provision while retaining neighbourhood control.


Introduction
Since the 1990s, gated residential neighbourhoods have been widely developed in different regions, reflecting shifts towards neoliberal modes of urban governance (Blakely and Snyder 1997, Chen and Webster 2005, Glasze et al. 2005, Roitman 2005, Cséfalvay 2011, Pow 2011, and representing a form of private governance (McKenzie 1994). Existing research has defined private governance as providing public services through organisations that are privately held and governed by residents (Gordon 2004). In North America, the gated community as private governance has residents who elect the homeowners' association to run the community by their own rules (Blakely and Snyder 1997). In Sydney, Australia, the significant growth of masterplanned estates represents a growing phenomenon in favour of the market and the private sector (Cheshire et al. 2010). The use of private compacts in these gated estates recasts residents' everyday practices and reshapes residents' relationships with the public realm (McGuirk and Dowling 2007). Many studies, mostly conducted within the circumstances of a liberalised housing system, agree that private governance provides residents with efficiency and autonomy (Atkinson et al. 2005, Goodman et al. 2010.
In urban China, state provision and governance of socialist work-unit housing ended in 1998. As alternatives, newer residential development projects similar to gated communities and master-planned estates have proliferated as one of the dominant residential neighbourhood forms over the past decade. They have shown salient features of private governance, and thus have stimulated substantial public and scholarly debates (Huang 2005, Pow 2011, Breitung 2012, Hendrikx and Wissink 2017. Institutional changes, fundamentally the state's deepened market reforms relating to land and housing development, have encouraged the privatisation of housing and public goods provision. Especially during the ongoing suburbanisation process, local governments developed incentives to draw the private sector into developing residential land and turned suburbs into spaces of capital accumulation (Zhang 2010, Shen andWu 2017). Furthermore, central government enacted property management regulations and homeowners ' association regulations, respectively, in 2007 and 2009. These private organisations have been formalised to take charge of maintaining facilities, providing services, and representing homeowners' rights in the neighbourhood. Changed governance in housing privatisation has received as much attention as and possibly more than centralised governance in existing neighbourhoods (Lu et al. 2018, Wu 2018a. Apart from supply-side changes, demands for "private governance" among residents themselves have also been strengthened. The reasons are threefold. First, the privatisation of public goods at the neighbourhood level has stimulated residents' responsibilities in terms of managing their neighbourhoods and hence securing the value of their estates. Second, the marketisation of neighbourhood services has been catching up with residents' growing consumption in everyday practices. Third, the variegation of gated residential development projects has provided residents with a variety of options for pursuing better neighbourhood experiences. However, little research attention has been paid to residents' preferences for and satisfaction with "private governance". One reason is that the weakness of the private sector, for example, civic organisations, in China has restricted residents' choices (Read 2012, Fu andLin 2013); another reason is that housing privatisation is designed to secure greater household privacy under state hegemony for the emerging middle class, as indicated by Pow's (2009) research in Shanghai, rather than giving residents choices regarding neighbourhood governance. Under such circumstances, a more nuanced understanding of the meaning of "private governance" in urban China is worth considering.
Limited research has tended to examine factors that may influence residents' preference for private governance, with a few exceptions. During a study exploring levels of support for privatisation in Toronto, Walks (2008) identifies that the spatial factors of residential settlements, for instance, suburban locations, have the most profound impact on residents' preferences. When evaluating the choice between public space and private space in Phoenix, Kirby (2008, p. 92) argues that rather than being "wielded directly by representatives of the state", residents deliberately choose privately managed spaces for practical reasons, such as better neighbourhood services. Moreover, based on a study of Seoul's gated communities, Woo and Webster (2014) propose that a more scientific assessment of governmental efficiency is needed to clarify many assumptions arising from the debates. This paper intends to study "private governance" from the perspective of residents in the context of post-reform China. Exactly what factors have driven residents to prefer different models of management in gated residential neighbourhoods? And to what extent do these gated residential neighbourhoods represent a new form of "private governance" in urban China? This paper intends to answer these research questions by drawing on a large-scale questionnaire survey conducted in the city of Wenzhou in China. The rest of this paper is organised to review the existing debates related to private governance and the emergence of "private governance" in post-reform China, followed by an explanation of survey methods and analyses of empirical evidence in Wenzhou. Multinomial logistic models are used to identify the ways in which different factors affect residents' preferences and satisfaction, as well as differentiations among gated residential neighbourhoods. Further discussions of the meaning of "private governance" in urban China and a conclusion are also provided.

The Current Debate on Private Governance
Burgeoning research has concentrated on gated residential neighbourhoods, ranging from masterplanned estates in Australia (McGuirk and Dowling 2007), gated communities in North America (Low 2003, Kirby 2008 and Europe (Webster 2002, Raposo 2006, Cséfalvay 2011, Hirt 2012, enclaves for the global elite in Singapore (Pow 2011), to porous space in the Philippines (Fauveaud 2016). Scholars have defined these gated residential neighbourhoods as forms of private governance wherein homeowners create their own covenants and contract private firms for professional management (McKenzie 1994, Gordon 2004). Three perspectives have been raised for examining private governance. The first perspective uses Buchanan's (1965) club theory to explain the private provision of goods and services (Le Goix 2005, Le Goix andWebster 2008). With private governance, each gated residential neighbourhood becomes a club of consumers who pay membership fees for using club goods and services collectively. This perspective emphasises the efficiency of private governance because the exclusivity of "club goods" overcomes the problems of free riders and public interventions (Webster 2002). The development of club realms also helps local governments to raise tax revenues and to relieve congestion in the public realm. The "clubbisation" of governance has also been understood as a market-driven process by Charmes (2009); his study of Paris suburban communes suggests that the market provision of public goods and services reduces public expenditure on suburban governance.
The second perspective stresses private governance for shareholder democracy (McKenzie 2005). Self-governance organisations act as agents of private governance under agreements made among homeowners. Cséfalvay (2011) considers that homeowners in gated communities undertake self-governance to secure the value of private property and to cope with local governance failure. McGuirk and Downling (2007) have demonstrated that master-planned estates adopt a private governance structure to entitle neighbourhood communal properties and to enforce neighbourhood covenants. Through private governance, property shareholders are able to "vote with their feet", and hence to ensure "a supply of civic goods which apparently are better acquired there than elsewhere" in the urban area (Webster 2001, p. 153). Kirby (2008, p. 12) also refers to the core implicationprivate governance is to be "policed by those who inhabit it".
The third perspective is the theory of exit and voice option. Cséfalvay (2011) argues that none of the above perspectives can fully explain the complexity of private governance. He then raises the public choice perspective based on Hirschman's (1970) work on the exit option and voice option, highlighting residents' preferences regarding local government and private governance. Specifically, when residents become disaffected with local government, they either pay additional costs to move to private governance, i.e. taking the exit option, or alternatively they stay with local government and express dissatisfaction through their voices. The exit option regards residents as voluntarily choosing private governance over local government in accordance with their preferences and mobility, whereas the voice option reflects democratic political contexts within which residents can express their demands without migrating or paying extra costs. Cséfalvay (2011) suggests that this perspective explains why gated communities are not common in central Europe because the voice option is profoundly rooted in their decentralised political systems, while private governance, the exit option, is also particularly costly there. Moreover, Roitman (2005) points out that gated residential neighbourhoods have been proliferating rapidly in developing countries. This is because local states in developing countries are more likely to promote costly exit options for economic growth than to provide sufficient non-profitable public goods as voice options for urban residents. This perspective considers social spatial differentiations as resulting from the differences between private governance and local government. Residents with higher mobility and similar demands may turn to private governance, while residents with low socioeconomic status stick to local government.
In summary, private governance highlights the efficiency of service provision, together with the autonomy of self-governance, at the neighbourhood level. However, the characteristics of private governance vary under specific local contexts. There is no empirical evidence on the perspective of residents' preferences in situations where the effects of both the exit option and the voice option are restricted.

The Emergence of "Private Governance" in Post-Reform China
Since the deepening of housing privatisation in 1998, the state welfare housing system has been replaced by the market provision of private property ownership Murie 2000, Zhou andLogan 2008). Residential development projects which used to be financed and managed by the public sector in the work-unit era have begun to show features of "private governance" over the last decade. Prior research has defined newer gated residential development projectsphysically enclosed and developed since housing privatisation in 1998as gated neighbourhoods (Pow 2009, Zhang 2010. Residential forms such as urban villages and prior work-unit housing are not included. Attempts have been made to consider gated residential neighbourhoods as club realms for private provision of services (Wu and Webber 2004, Wu 2005, Hendrikx and Wissink 2017. This strand of literature suggests that property management companies establish a fee-paying scheme with residents for neighbourhood management. Apart from service provision, research on grassroots governance acknowledges that homeowners' associations bring homeowners together to protect the rights of private property (Read 2003, Tomba 2005. Therefore, gated residential development projects demonstrate important features of "private governance". In addition, "private governance" has been enhanced by the decentralisation of urban governance for local growth. During rapid processes of suburbanisation, local governments have been confronted with financial burdens in providing public goods and maintaining services. By supporting "private governance", local governments transfer these administrative costs at the neighbourhood level to the private sector, which subsequently charges the users. Recent studies stress that decentralising neighbourhood governance has become an essential measure for local governments to attract private capital to invest in suburban residential and infrastructure development (Fleischer 2010, Shen andWu 2012). Under these circumstances, gated residential development projectsmostly evident in suburban areashave been developed, financed and managed by the private sector.
Meanwhile, the consumption demands of residents for "private governance" have been fostered since the move from welfare housing to private properties. Scholars have tested residents' satisfaction with aspects of the neighbourhood built environment, advocating the importance of neighbourhood maintenance for increasing satisfaction (Fang 2006, Wu 2010, Li et al. 2012, Li and Wu 2013. Influenced by consumerist culture, residents have begun to pursue middle-class lifestyles (Davis 2000). In gated communities in Shanghai, aesthetic landscapes, privilege and the moral order of the good life, as symbols of middle-class lifestyles, are delivered by private governance (Pow 2009). Existing neighbourhood studies also imply that urban residents show an increasing awareness of self-governance through the establishment of private governance organisations (Chen and Webster 2005, Huang 2006, Zhang 2010. However, a series of studies on neighbourhood governance argues that the state has continued to play a dominant role in organising neighbourhoods in urban China (Hsing 2010, Lin 2011. Despite land and housing reforms activating the privatisation of property ownership, the state continues to be the ultimate owner of urban land and the regulator of urban affairs. Within this context, studies consider that gated residential development projects have primarily emerged to replace the retreat of the state from the provision of housing and public services, and secondly to open new avenues of revenue for local growth by suburbanisation and urban regeneration (Wu 2018b). In this scenario, "private governance" is limited. Huang (2005) argues that the homeowners' association is in fact a phantom agent launched by local governments to restore neighbourhood governance. Therefore, the formation of "private governance" is different in the context of urban China due to strong state intervention. First, according to the club theory, privatised services in gated communities are provided for efficiency and exclusivity. Since the state's retreat from the provision of neighbourhood services, market forces have been used to supplement inadequate public services and to boost property management industries in urban China. However, property management companies do not guarantee the efficiency of neighbourhood services provision because delivery is managed in accordance with property management fees, which are decided by local government.
Second, the method of shareholder democracy is not possible because residents do not have sufficient power to determine the system of neighbourhood governance in urban China. Although homeowners' associations have gradually emerged for managing neighbourhoods, they are set up with an acutely practical purpose to supervise property management companies. Homeowners' associations lack autonomy and financial independence and have poor representation of residents.
Third, in the exit and voice option theory, private governance is the result when residents are not satisfied with the local government and have hence taken an exit. Alternatively, residents choose to stay with local public services and voice their demands through local government. This exit and voice theory explains the transformation from government-led to self-governance; however, it is not sufficient to explain "private governance" in China. First, residents cannot freely choose the provision of public goods. The state has established policies, such as national planning codes, to standardise the provision of public goods at the neighbourhood level. Concurrently, local government has nominated state agencies to regulate private provision. The main purpose of introducing "private governance" is not to respond to residential demands, but rather to stimulate the housing market and generate local revenues from residential land leasing.
Based on our review of the above literature, we found that residents have been able to exert more influence in terms of shaping neighbourhoods, although the state remains a dominating role in their governance. Examining the perspective of residents regarding "private governance" and the determinative factors offers a new research approach, while simultaneously providing empirical evidence of the relation between residents and "private governance" under the condition of strong state intervention in post-reform China.

Data and Methods
This research is based on a large-scale questionnaire survey conducted in the Wenzhou municipal area, which includes the districts of Lucheng, Ouhai and Longwan, from March to May in 2013. Wenzhou is a densely populated city with more than nine million registered residents. The city is also well known for having the most active private economy in China, with the private sector taking more than 80% of local GDP (Wei et al. 2007). Since the termination of work-unit housing in 1999, residential developments have become a pillar industry for local economic growth, attracting tremendous investment from the local private sector. Official statistics from 2000 to 2010 indicate that the scale of average annual investment in new residential development projects is more than 10 billion Yuan, with the private sector contributing 81.7% in total; as a result, in 2010, 42.6% of residents in Wenzhou were private homeowners (Wenzhou Statistical Bureau 2010).
In the city of Wenzhou, gated residential neighbourhoods are commonly seen in suburban areas where large-scale new residential developments were planned. For their development, national planning codes require real estate developers to be responsible for providing residents with public goods, including green spaces with recreation facilities, grocery stores, educational institutions such as nurseries and kindergartens, security control centres and waste collection points. The amount and quality of public goods originating from private provision vary in gated residential neighbourhoods. For their management, private organisations as constituted by property management companies and homeowners' associations are in direct charge. Neighbourhood services are entirely privatised and delivered by professional property management companies at additional cost to residents. Homeowners' associations have been widely established to carry out homeowners' decisions in neighbourhood management. Meanwhile, the local government still has an administrative role at the neighbourhood level. Therefore, in each sampled gated residential neighbourhood, local government agencies, property management companies, and homeowners' associations all exist. Therefore, in this research, residents of sampled gated residential developments have experience of all three types of management.
Empirically, three forms of gated residential neighbourhoods, including market-led, mixed and state-led development projects are recognised in the research area. A market-led development project is exclusively for market buyers. They are often located in a suburban area. A mixed development project contains affordable housing, involving both the state and market for development. Its market mechanisms receive strong state intervention under affordable housing provision schemes. A state-led development project is initiated by the government to relocate residents whose properties have been demolished in massive redevelopment projects. Three distinct forms of gated residential neighbourhoods represent different state-market-society relations, and thus they have variegated management models.
Regarding the questionnaire survey, this study followed the Probability Proportional to Size sampling method. It first sampled 11 out of 559 gated residential neighbourhoods in the Wenzhou municipal area, and secondly randomly selected 94 households in every gated residential development identified, aiming at 1,034 valid questionnaires in total. Figure 1 shows the location of the sampled gated residential neighbourhoods. The two-step sampling ensured an equal probability of being surveyed for each household. Questionnaires were delivered to heads of household or their spouses. Sampled households were visited and had their questionnaires filled on site. The quality of responses was checked before being collected, so as to ensure the quality of the survey.
The attributes surveyed constitute three sets: (1) respondent's socio-economic profiles; (2) preferences for and satisfaction with different management models; and (3) everyday practices in the neighbourhood. Table 1 shows the socio-demographic status of sampled respondents in gated residential neighbourhoods. Over 93% of sampled households have private ownership of their properties in the neighbourhood. The majority of respondents hold an educational degree from college, university or above and work for the private sector. The last set of attributes measures whether residents use particular goods that are privately provided in gated residential neighbourhoods as frequencies of contacting property management companies and homeowners' associations. Additionally, neighbourhood factors include respondents' housing types, degree of facility provision in neighbourhood and neighbourhood forms. This set of data was collected by on-site observation and by reviewing planning documents for gated residential development projects.
Multinomial logistic regression analyses have been employed in this research, with the intention of revealing the factors that produce significant impacts to shape residents' preferences. Specifically, the dependent variable in the first multinomial regression model is residents' preference for neighbourhood management among property management companies, homeowners' associations and local government agencies. This model reveals the factors behind the preference for "private governance". The second multinomial regression model identifies differentiations among three gated residential neighbourhoods. It focuses on examining whether residents are segregated into different forms of gated residential neighbourhoods due to the different contents of management. The third model tests residents' satisfaction with property management companies and homeowners' associations, inspecting whether they emphasise the practical benefits of private management or its symbolic meaning of self-governance. The dependent variables of residents' satisfaction are measured by binary answers, namely "satisfied" and "not satisfied".
At the same time, qualitative analysis is used to offer supplementary explanations for the results of the regression models. This set of data is drawn from interviews conducted from December 2013 to March 2015 with residents of different forms of gated residential neighbourhoods and covers a variety of social groups in terms of age and occupation. Table 2 shows the results of multinomial logistic regression, testing the significance of different attributes in determining residents' preferences among property management companies, homeowners' associations and government agents in gated residential neighbourhoods. Survey data reveal that an overwhelming amount of residents prefer private organisations in the neighbourhood, with 81.2% of residents choosing property management companies and 8.6% of residents choosing homeowners' associations. Those who prefer government agencies take up the remaining 10.2% of the total amount, and are considered as the reference group in this analysis. Model 1  examines the effects of residents' socio-economic attributes and neighbourhood factors, while Model 2 includes attributes of residents' neighbourhood experiences for a full analysis. As a result, the first model is weak in explaining the differences of residents' preferences, while the second model has greatly improved model significance, Cox and Snell R 2 , and Nagelkerke R 2 . Overall, residents' preferences among property management companies, homeowners' associations and local government agents are more likely to be influenced by neighbourhood forms and everyday practices, regardless of their educational level, occupation and household income level. It is difficult to identify the social groups who support property management companies and homeowners' associations, thereby suggesting that socio-economic status hardly determines residents' preferences for "private governance". Housing quality and neighbourhood quality hardly have an influence on their preferences.

Findings: The Determinants of Residents' Preferences for "Private Governance"
Residents' everyday practices regarding "private governance" are interpreted by two kinds of factors: using the private provision of public goods in gated residential neighbourhoods and contacting the relevant private governance organisations. In this research, only three categories of public goods are found in all samples, namely recreation facilities, nurseries and grocery stores. Specific to the use of them in these gated residential neighbourhoods, 52.0% of residents report using recreation facilities, while 36.8% and 37.0% of residents, respectively, use nurseries and grocery stores. The results show that residents using recreation facilities in the neighbourhood are more likely to support property management companies. Additionally, the odds ratio of this factor is among the highest, confirming that good packages of facilities from private provision Note: *p < 0.05, **p < 0.01, ***p < 0.001. In model 1, Cox and Snell R 2 = 0.041, Nagelkerke R 2 = 0.058, p < 0.01. In model 2, Cox and Snell R 2 = 0.097, Nagelkerke R 2 = 0.137, p < 0.001. strongly contribute to residents' preference for professional management. In contrast, using grocery stores in the neighbourhood decreases residents' support for homeowners' associations. This is because many grocery stores in the neighbourhood are rented and open to all consumers, which is agreed by homeowners' associations. However, residents consider that this threatens homeowners' benefits as stakeholders of private properties. At the same time, it is reasonable to find that residents are more likely to support the organisations they contact more frequently, while frequent use of neighbourhood services leads to support for private governance organisations in general. Implications can be seen here. On the one hand, private providers focus on packaging gated residential neighbourhoods with recreational facilities rather than providing costly functions such as education and commerce for residents, leaving residential demands for the latter mostly outside neighbourhood provision. On the other hand, residents' preference for "private governance" is generated from neighbourhood experiences, particularly by using public goods in gated residential neighbourhoods and contacting private governance organisations.
Furthermore, living in market-led residential development projects tends to indicate a preference for property management companies, even when the attributes of socio-economic status, the built environment quality and neighbourhood living experiences are controlled. In suburban areas where governance is relatively weak and fragmented, the allocation of public resources from local government to meet residents' demands is hardly sufficient or effective. By adopting market mechanisms, market-led residential development projects provide high quality services, charging residents expensive management fees. Property management companies are likely to provide satisfying services for residents' everyday uses, aiming to develop a trustful relationship with homeowners, thereby to renew their contracts and raise service prices. This provokes the prevalence of consumerist culture. Three key words were frequently mentioned by residents in justifying the reasoning behind their preferences for property management companies: "contracted", "profession" and "trust", as shown in the interview with a resident: We are in fact consumers; the property management company is hired to solve neighbourhood problems for us, providing neighbourhood services we need; they arrive at once when we demand, and have nice attitudes to us. (resident of market-led residential development project, interviewed on 29 December 2013) To better understand the differences between the three forms of gated residential neighbourhoods, another multinomial regression model was conducted, with results shown in Table 3. Overall, both attributes of residents' socio-economic status and everyday experiences are strong in differentiating these gated residential neighbourhoods. It is evident that residents of market-led residential development projects are likely to have a high income and educational level. They also have a strong tendency to live in spacious flats or villas, which again confirms their wealth. On the other hand, the households of mixed residential development projects are not well established in terms of socio-economic status, for example, being young and being renters rather than property owners. Compared with state-led residential development projects that have newly emerged during urban redevelopment processes, the residents of mixed developments tend to have been settled in their neighbourhoods for a longer time.
When the socio-economic attributes of residents are controlled, the contents provided by different forms of gated residential neighbourhoods can be distinguished. The results confirm the previous analysis that, compared with state-led residential development projects, residents living in market-led developments are more likely to use neighbourhood facilities and neighbourhood services, and frequently turn to property management companies rather than homeowners' associations. In contrast, mixed residential development projects do not share any similar outcomes in terms of management content. In fact, residents of mixed developments have a higher likelihood of frequent contact with homeowners' associations.
Analyses so far suggest that, compared with relocated residents, market buyers are more concerned with the quality of private provision, while affordable housing residents who have relatively lower consumption tendencies turn to contacting homeowners' associations. It further appears that residents are separated by different gated residential neighbourhoods due to being attracted by the different management models that gated residential neighbourhoods adopt. First, through market-driven processes, market-led residential development projects provide neighbourhood services and facilities that meet residents' consumption demands. Dissimilar to market-led developments, mixed residential development projects offer affordability for residents and hence place less emphasis on costly provision. Residents in state-led developments were relocated through government redevelopment policies and did not voluntarily choose "private governance". The change in service provision and management has been imposed on them.
Second, mixed residential development projects with a longer established time are generally found close to the city centre, while market-led residential development projects mostly emerge in suburbs. Living near the well-established city centre allows residents convenient access to public facilities and a diversity of city services in their everyday practices, rather than relying on private provision in the gated residential neighbourhood. In contrast, as the local government retreats from providing serviced land for suburban residential development, the residents of suburban gated residential neighbourhoods have very few alternatives but to rely heavily on the private sector for services and management in everyday practices.
Furthermore, logistic regression analyses of the likelihood of being satisfied with property management companies and homeowners' associations were carried out; this is demonstrated in Table 4. The satisfaction with private governance organisations is generally low, as 45.2% of the sample respondents responded as being satisfied with the former, while the proportion of respondents expressing satisfaction with the latter can be identified as 17.6%. The much lower satisfaction with homeowners' associations suggests that they fail to adequately represent all homeowners and convey self-governance in the neighbourhood. This again confirms earlier analyses whereby services commodification is better performed than self-governance among residents in gated residential neighbourhoods.
Residents' satisfaction with property management companies and homeowners' associations appears to be influenced by different determinants in the different models. In analysing satisfaction with property management companies, attributes reveal that residents using public goods and contacting private governance organisations are all significant. When it comes to residents' socio- Table 3. Differentiations of the three forms of gated residential neighbourhoods (multinomial regression models: the reference group = state-led residential development projects). economic status, similar results are found: females and high-income earners are more likely to report satisfaction with property management companies. However, it is noteworthy that residents working for the public sector are less likely to feel satisfied with the private provision of neighbourhood services. They tend to consider that state agencies, rather than the private sector, hold the proper power in relation to neighbourhood governance, as shown in the interview with a resident in a mixed development: I believe it is government departments that have the duty of governing residents'. (resident of mixed residential development project, interviewed on 7 January 2014) Neighbourhood form is the next highlighted factor, while the remaining factors relating to housing quality and neighbourhood quality have little impact on independent variables. Compared to state-led residential development projects, market-led developments are more likely to satisfy residents in terms of property management, while mixed developments have poorer services provision which disappoints residents the most. It is noteworthy that on being relocated to gated residential neighbourhoods residents are obliged to pay property management companies a property services fee that they never had to pay in urban villages. The imposed service fees and the dramatic changes in their everyday practices lead relocated residents to become involved in conflicts with property management companies, resulting in the fact that this group of residents rarely favours property management companies, as shown in the interview with a resident in a state-led residential development: I do not communicate with staff of the property management company. There is nothing to communicate about with them, and I don't pay for neighbourhood service fees. (resident of state-led residential development project, interviewed on 8 January 2014) In contrast, as shown in Model 2, very few personal attributes and neighbourhood attributes have significance in impacting residents' satisfaction with homeowners' associations. At the same time, factors of neighbourhood experiences, such as frequency of contacting homeowners' associations and frequency of contacting property management companies, tend to have similar effects on residents' perceptions of homeowners' associations. Homeowners' associations fail in delivering satisfying work in response to residents' demands because the state has placed restrictions on the governance power of homeowners' associations, as shown in the interview with a resident of a mixed residential development project: The homeowners' association contacts me only if it is about to elect the board; however, I think the homeowners' association should represent residents in every aspect and bring forward residents' decisions; then it should hire a property management company to convey these decisions. (resident of mixed residential development project, interviewed on 12 January 2014)

Conclusion
Private governance has been encouraged as an innovative approach for local neighbourhoods to navigate urban changes since the end of the twentieth century (Imrie and Raco 1999). In gated communities in the US context, private governance underlines the role of homeowners' associations, representing the surging demand for self-governance (McKenzie 2005). In Australia, meanwhile, private governance in master-planned estates is largely supported for the production of suburban residential space (McGuirk and Dowling 2007). The development of gated residential neighbourhoods in different regions reflects the withdrawal of the welfare state in allocating public resources and the adoption of neoliberal urban governance. China is also developing new forms of gated residential neighbourhoods. Yet the underlying mechanism reveals different state-market relations. This paper has examined the preference for different forms of neighbourhood management and satisfaction from the perspectives of residents in the city of Wenzhou in China. Specifically, we explored whether residents prefer property management companies to government agencies for neighbourhood governance, and whether residents are satisfied by different management types. We have found that the preference for private management companies over governance by government agencies is evident across residents with different socioeconomic status. This preference is clearly affected by everyday practices in gated residential neighbourhoods; for instance, when there is frequent use of private provision or living in a market-led residential development project, it becomes stronger, similar to the findings of Walks (2008, p. 279) in the city of Toronto that suburban forms can lead to increased support for privatisation. Nevertheless, residents' satisfaction with homeowners' associations are generally low as for most residents gated residential neighbourhoods mean market provision of neighbourhood services rather than self-governance. Through interacting with property management companies, residents begin to understand their services and prefer them to other government agencies for daily maintenance, reflecting Kirby's (2008) argument that residents choose "private governance" for practical reasons.
Furthermore, this study shows that different forms of gated residential neighbourhoods recast everyday practices differently, and the variegated implications of neighbourhood governance are determined by the state-market relation. With control over socio-economic attributes, the market-led form has a stronger preference for property management companies or homeowners' associations due to emphasising the privatised provision of public goods and services. The local government introduces market instruments for neighbourhood services, with the aim to facilitate suburban land development and to retreat from providing serviced land. A state-led form is used by the local government for land acquisition in the urban redevelopment process. It is not a voluntary choice for residents regarding neighbourhood governance as they are relocated by the local government. The mixed form has strong state intervention and thus offers affordability rather than satisfying neighbourhood services to fulfil the local government's political tasks of providing affordable housing. The poor performance of private governance organisations under state control reflects the visible role of the state in governing the mixed form and the state-led form of gated residential development projects.
Evidence from Wenzhou shows that "private governance" in the Chinese context actually means the commodification of neighbourhood services, specifically delivered by property management companies, rather than establishing self-governance through homeowners' associations. The changed approach to neighbourhood governance helps local government to generate revenues from land and housing development during suburbanisation and regeneration processes, holding back demand for self-governance from neighbourhoods. Although employing different management models of neighbourhood governance, the state holds a key role in the process of neighbourhood governance transformation (Wu 2018b). Consequently, gated residential neighbourhoods in urban China are a response to the state's intention to retreat from providing serviced land, while retaining neighbourhood control through market instruments.

Disclosure statement
No potential conflict of interest was reported by the authors.