Approaching Neighborhood Democracy from a Longitudinal Perspective : An Eighteen-Year Case Study of a Homeowner Association in Beijing

Neighborhooddemocracywas introduced into urbanChina in the early 1990s as away tomanage the social con�icts associatedwith the housing reform. Based on a case study of Dragon Villas, Beijing, this paper explores the causes, processes, and consequences of neighborhood democracy at the microlevel from a longitudinal perspective. ree insights are particularly noteworthy. First, the decrease in rental revenue and occupancy rate and the arrival of Chinese owner-occupiers contributed to the emergence of neighborhood democracy in Dragon Villas. Second, the establishment of a homeowner association, far from ending in the conclusion of neighborhood democratization, was only a �rst step. Furthermore, con�icts between the developer and the homeowners, and among homeowners, played a crucial role in lengthening the process of neighborhood democratization. ird, democratic self-governance resulted in improved governance, a more diverse built form that articulates individuation through consumption, and changes that re�ect the importance of privacy and exclusivity.


Introduction
e rise of homeowner associations (HOAs) and neighborhood democracy in Chinese cities has been acknowledged as one of the most remarkable features of China's urban housing reform in the last two decades [1].Before the reform, housing provisions and management in Chinese cities were predominantly provided by work units and local governments as welfare bene�ts [2].However, the combination of insufficient state investments and very low housing rents meant that many cities suffered from housing shortages and a lack of basic repairs and maintenance to existing housing stock [3,4].China's housing reform in the 1990s aimed to address this problem through wholesale privatization and commodi�cation [5,6].Since then, the homeownership rate has skyrocketed from 10% in the early 1980s [7] to 30% in 1995 to more than 70% in 2000 [8], and gated communities have dominated the landscape of Chinese cities [9], bringing signi�cant changes to urban governance in the process [10].e central government officially introduced HOAs in 1994 when the Ministry of Construction (MOC) promulgated "Methods for the Management of Newly Built Urban Residential Neighborhoods, " although unauthorized HOAs were established in Shenzhen as early as 1991.e key consideration was that the traditional, top-down neighborhood governance systems (e.g., residents' committees) lost their legitimacy for governance because they no longer provided housing services [1].In a survey of homeowners in Beijing, Shanghai, and Guangzhou, 80.6% agreed or somewhat agreed that their neighborhoods should be managed by HOAs, whereas only 33% agreed or somewhat agreed that their neighborhoods should be managed by residents' committees [11].Another consideration was that due to inadequate legal protection and the working of local progrowth regimes [12][13][14], infringements upon homeowners' rights by local governments, developers, and management companies are commonplace.Consequently, homeowner activism has become a permanent �xture in Chinese cities in the last decade, despite the fact that homeowners are supposed to be the benefactors of reform [15].e newspapers in Beijing and Guangzhou reported 600 and 1226 incidents of homeowner activism, respectively, from 2003 to 2008 [16].While most incidents are legal, the use of violence is not unheard of.e central government thus hoped to introduce HOAs and democratic self-governance as an institutionalized means of resolving con�icts [17].Table 1 shows that a substantial number of neighborhoods in China's major cities are overseen by an HOA.is idea of democratic self-governance is radical and sensitive within a China that is still ruled by a Communist Party [17].HOAs also pose challenges for urban governance at a time when the state is seeking to reestablish control through large-scale community building projects [18,19].
e main contribution of this study to the literature is its approach to neighborhood democracy from a longitudinal perspective.Longitudinal research utilizes multiple waves of observations over an extended period of time, is concerned with the temporal sequence of events, and focuses on duration, time, trend, periodicity, continuity, and rate of change [20][21][22].Longitudinal research is thought to have two key strengths: identifying patterns of change and establishing the direction and magnitude of causal relationships [23].e aim of this paper is to demonstrate, through a case study, the value of longitudinal research in providing new information about neighborhood democracy in China.Existing studies are mostly cross-sectional in nature, observing neighborhood democracy at one speci�c point in time either through a quantitative, survey-based method or a qualitative, interview-based approach.e only two notable exceptions are Cai [15] and Fayong [24].However, in both cases, the authors address the broader phenomenon of homeowner activism rather than neighborhood democratic self-governance.
is study utilizes historical records to reconstruct past events and thus can be classi�ed as retrospective longitudinal research-an approach that has been considered as a practical alternative to the costly and inefficient panel-based longitudinal research [25].I selected Dragon Villas, a gated community in northeast Beijing, as the �eld site because of its long history and its presence of contacts.Evidence was primarily drawn from documents collected over eighteen years (1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012), including records of homeowners and tenants, meeting agendas, management expenditure statements, and internal memoranda of the management company.I also conducted twelve interviews between June and July 2012 with a diverse group of people that included the developer, the property manager, and several HOA members.
e paper begins with a discussion on relevant government regulations in Beijing that constitutes the macrobackground of the case study.is discussion is followed by three sections that describe the case study.In each section, I demonstrate how a longitudinal perspective can reveal new insights.e �rst section examines the drivers of neighborhood democracy.e second section investigates the process of neighborhood democratization.e third section discusses the impact of neighborhood democracy.e paper concludes with a re�ection on the strengths and weaknesses of a longitudinal approach.

Neighborhood Democracy Regulations in Beijing
Regulations over HOAs at both the national and local levels have changed substantially over the last decade.MOC's "Methods for the Management of Newly Built Urban Residential Neighborhoods" laid out the basic structure and power of HOAs.e key functions of HOAs were to represent the rights of the homeowners, to govern neighborhoods through draing and enforcing local covenants, to maintain common properties, and to monitor property management companies, but the right to dismiss property management companies was withheld.In 1995, the Beijing Municipal People's Government (BMPG) released the "Methods for the Management of Residential Neighborhoods in Beijing, " which introduced several changes and clari�cations to the national regulations.e most important change was the granting to the HOAs in Beijing the power to dismiss property management companies under �ve conditions: the management company is (1) unable to repair houses and public utilities in a timely manner; (2) unable to correct disordered management; (3) implementing the unauthorized increase of fees; (4) implementing unauthorized changes in the use of public facilities; (5) unable to perform other contractual obligations.is change signi�cantly increased the power of HOAs.In comparison, Hong Kong requires 50% support from all homeowners to dismiss their property manager [26].Accordingly, at this point, the Beijing HOAs had similar legal power to their counterparts in the United States, less than the foreclosure power [27].Another important addition under Beijing's policy was that developers were required to make a one-time payment to HOAs or property management companies as seed funding for management fees.e payment was calculated at 2% of the total construction cost.
In 1997, the BMPG released the "Notice on the Trial of Establishing Management Committees, " which introduced a �ve-step procedure for establishing HOAs in Beijing.First, when a residential neighborhood's occupancy rate reaches 50%, developers are to establish a preparatory committee, under the guidance of the district government and the local street office.Second, the preparatory committee should nominate candidates aer consulting with homeowners and resident committees.ird, the preparatory committee should convene a conference with all homeowners to elect an HOA committee.Fourth, the HOA committee is to elect a chairman.Fih, the preparatory committee is to submit all relevant documents to the district government for approval.e problem with this procedure was that developers had little interest in establishing HOAs, preferring to govern the neighborhoods themselves.Consequently, very few neighborhoods established HOAs.In response, the BMPG announced in 1998 that homeowners could request the developer to establish an HOA.If the developer did not respond aer one month, the homeowners could establish a preparatory committee without the developer.In addition, the BMPG further simpli�ed the procedure for establishing HOAs.Rather than following the �ve-step procedure as discussed earlier, preparatory committees could create HOAs directly and without a proper election, provided that all homeowners were noti�ed and that no more than half of the homeowners rejected the proposal.However, this arrangement also enabled developers to establish puppet HOAs.Nevertheless, these regulative changes indicated that the BMPG was generally supportive of HOAs.
Major changes to the regulations at the national level were made in 2003, when the State Council released a policy entitled "Property Management Regulation." is regulation changed the distribution of power as almost all power that was formally attributed to HOA committee was now transferred to the homeowner conferences (HOCs), which were annual meetings of all homeowners.Aer the changes, all major decisions made by HOA committees needed to be passed by HOCs.e purpose of this change was to prevent the abuse of power by a small group of homeowners or a puppet HOA controlled by the developer.However, the regulation also made it difficult for homeowners to exercise their rights.For example, dismissing property managers now required the approval of two-thirds of the homeowners present at the homeowner conference.
In 2010, the BMPG released the "Methods on Property Management, " which contained some changes and additions to the central government's "Property Management Regulation." e most signi�cant change was that developers were not allowed to collect management fees until an HOA was established in the neighborhood.is measure was designed to address the problem that developers were reluctant to establish HOAs.However, the regulation also stipulated that street offices and town/township governments were responsible for the coordination, guidance, and supervision of HOAs, indicating that the BMPG was seeking to increase government control over homeowners.

Drivers for Neighborhood Democracy
In the United States, HOAs are predominately formed by developers when a neighborhood is being built [28][29][30].e situation is somewhat different in China, where developers generally want to govern the neighborhood for as long as possible.It is the homeowners who must assume the initiative to democratize their neighborhood.us, it is important to understand the drivers behind neighborhood democracy.
Dragon Villas (DV) was �rst populated in 1994.Similar to other gated communities, the developer established a property management branch (PMB) to manage and serve the neighborhood.Several decisions made by the developer at this stage would later become causes of concern.Among these concerns were the level of management and utility fees and the lack of transparency and accountability regarding how the management fees were spent.However, from 1994 to 1997, these problems were dormant and did not lead to homeowner democratization for two reasons.
DV was originally developed as a foreign gated community with high-end houses designed for foreigners [31,32].In 1994, when the �rst phase of construction was completed, houses were mostly sold to Hong Kong investors, who then leased the houses to multinational corporations that needed to house their foreign employees.Foreign companies were reluctant to purchase properties because of the legal and bureaucratic complexities and uncertainties about owning properties in China.Because Beijing suffered from an acute lack of suitable accommodations for foreigners, rent increased to an unusually high level (Table 2).In comparison, management fees only amounted to approximately 3% of the rent.
However, conditions began to change in 1997 in two ways.First, the rental market for foreign housing slowed signi�cantly in 1997.On the demand side, the Asian �nancial crisis brought foreign investments to a standstill.At the same time, as con�dence in the Chinese property market grew, foreign companies began to purchase houses rather than renting.On the supply side, because of the construction boom in the early 1990s, there was an oversupply of high quality foreign housing.As a result of these factors, some DV absentee landlords were unable to �nd tenants, while others were forced to reduce their rents to as low as 1800 USD per month.In this new context, high management and utility fees became a problem for absentee landlords, especially those whose properties were vacant.
At the same time, as China's economic development continued at a fast pace, a growing number of local Chinese became interested in buying a home in DV.However, there were very few local owner-occupiers between 1994 and 1996 because the income of local Chinese was insufficient to purchase houses.As shown in Table 3, the proportion of owner-occupiers increased from 12.2% in 1995 to 54% in 1997.e Chinese owner-occupiers, many of them �rst time homeowners, were shocked by the high costs of management fees.For example, the monthly management fee for the largest type of house was 647 USD, and accordingly, those who had assumed a mortgage were particularly sensitive to pressures of the cost of living.
Because of these changes at the macrolevel, the demand for establishing an HOA began to surface in the spring of 1998.e key actors were homeowners (absentee landlords and owner-occupiers), and very few tenants participated.e low participation rate of tenants is predictable because of the low transaction costs for the tenant to exit, the tenant's weak attachment to the neighborhood, and the tenant's noninvestment orientation [33,34].To study the causes of concern that motivated homeowners, I aggregated data from meeting minutes, correspondences, and open letters in 1998 to obtain a list of concerns and their frequency rate (as a proxy to relative importance).I classi�ed all concerns into three groups: cost of living, exchange value, or use value.Cost of living (management and utility fees), which was by far the main source of anxiety among homeowners, accounted for 63% of the total concerns.Use value issues, including water quality, shuttle services, and security, were raised less frequently (23%), while exchange value issues (property price, occupancy rate, and rental revenue) accounted for only 14% of the total concerns.e homeowners regarded the monopolization of power by the developer as the root cause of the problems that plagued the neighborhood.us, it was fundamentally an issue of trust.e homeowners simply did not trust the developer in matters of governance, nor did they believe that the developer had their interests at heart.e DV experience shows that while material-interest con�icts provide the immediate driver for neighborhood democracy, broader factors are also at play, including the decrease in rental revenues and occupancy rates and the arrival of Chinese owner-occupiers.e driver of neighborhood democracy, and homeowner activism in general, is complex, and Purcell's [35] argument that people mobilize politically based on subjectively realized interests, rather than objective material interests, rings true.

Neighborhood Democratization
e regulations outline a straightforward process for establishing HOAs.In reality, bringing democratic selfgovernance to neighborhoods is not easy.Not only do the homeowners face uncooperative or even hostile developers, they also must deal with opposition from the local government [17,36].Microlevel factors also determine the success or failure of neighborhood democratization.In a comparative study of 23 neighborhoods in China, Read [12] posits that several microlevel factors, including the size of the neighborhood, the commonality of interests, and the intensity of grievances, in�uence the success of democratization.To this list, we add the availability of personal connections, strong leadership, appropriate tactics, and the availability of resources [37,38].Given the complex situation, there is surprisingly little research performed on the process of neighborhood democratization experienced on the ground.Accordingly, the aim of this section is to provide a demonstration of how to implement a new democratic self-governance system.e process of democratization began in March and April 1998, when a series of meetings was conducted between the developer and DV homeowners to discuss community issues.During the meetings, the issue of establishing an HOA, to be called Dragon Villas Management Committee (DVMC), was raised.e developer declared that it would help the homeowners to realize this goal, perhaps hoping to placate the disgruntled homeowners.e parties then began to work collaboratively on the preparation of the �rst homeowner meeting and election.
Con�icts soon emerged among the homeowners.Owneroccupiers argued that only long-term residents should be eligible for chairperson of DVMC, ostensibly for the sake of efficacy.e absentee landlords demanded that half of the seats of DVMC be reserved for their group because they worried that their interests would not be represented by owneroccupiers.Both measures failed to be adopted because of strong opposition.Disagreement between the developer, on the one side, and homeowners, on the other, further delayed the preparation process.e two sides disagreed on the number of votes the developer should have and the number of candidates the developer could nominate in the election.As both sides had equal numbers in the preparatory committee, the matter was deadlocked.ere was an attempt to invite the district government to resolve the differences, but the government opted to stay out of the con�ict.e deadlock was �nally resolved in June 1999 when a compromise was reached in favor of the developer, who obtained the right to compete in elections and the right to vote.e developer's victory was short lived, however, as none of its candidates won a seat in the election held in September 1999 due to a surprisingly strong turnout of homeowners.Two weeks later, DVMC was formally recognized by the district government.
As the homeowners quickly learned, the establishment of DVMC did little to change the way DV was governed.While DVMC was allowed to handle internal affairs, such as draing a dog regulation, the developer was still in control of neighborhood affairs as long as the neighborhood was managed by its subsidiary.e issues that most concern the homeowners, namely, management fees and accountability, remained unresolved.While new actions and strategies were required, at the same time, a new fault line emerged among the homeowners.e conservative faction favored a gradual approach through independent audit and negotiations with the developer.However, the radical faction preferred a swi dismissal of the PMB and invited a new management company that will follow DVMC's leadership.Both groups began to work on their objectives in an uncoordinated effort.While the conservative group was making arrangements for an audit, the radical homeowners busied themselves in persuading the manager of PMB to form his own company.When the manager was dismissed by the developer, however, even more problems erupted.Upon learning the news, the radical faction convened a meeting and promptly dismissed PMB, but they retained the same staff and appointed the manager as an advisor.All actions were within the legal power of DVMC at that time.e problem was that the developer refused to cooperate, and the ensuing chaotic situation began to threaten the daily lives of the homeowners.e radical homeowners had no choice but to acquiesce and allow the developer to resume control of the community.e conservative faction did not have much success either.e independent audit e�posed a multitude of �nancial problems, but negotiations with the developer were ineffective when few promises of reform made by the developer were kept.By June 1999, realizing that their efforts were unsuccessful, the homeowners tried the tendering approach (i.e., open tender by inviting bids from management companies).e advantage of this approach was that the method was formally endorsed by the government.During the tendering process, the government would supply professionals to participate on the evaluation committee.e homeowners believed that the involvement of the government would force the developer to recognize the tendering result.
However, this path was also marked with difficulties.While the developer continued to obstruct the process (e.g., by refusing to submit the management company assets list), internal division became the biggest obstacle at this stage.e division was evident in the heated discussion during the 2001 homeowner conference, and the key points made by the homeowners are summarized as follows.
Mr. F believed that the homeowners who paid management fees were taken advantage of by those who did not pay such fees.

Mr. O said, as a homeowner who refused to pay management fees, he felt that he was inadequately protected by law. He wished to see a competent DVMC and an independent property manager.
Mr. Z wanted to understand how arrears in management fee payments are being addressed.

Mr. P thought that it is important for the homeowners and DVMC to unite. He believed that arrears should be paid into the trust account rather than to the PMB.
Ms. M requested, from an absentee landlord's perspective, that the occupancy rate and the quality of management be improved.

Ms. C affirmed the creation of the trust account.
Mr. Q questioned the legality of the trust account.

Ms. W emphasized the importance of understanding and communication between DVMC and the homeowners.
Mr. S thought that DVMC members should be owner-occupiers and that DVMC should employ a secretary.

Ms. L explained the reasons for not paying the management fee and suggested an audit of PMB.
Owner of #35 thought that auditing was too expensive.

Mr. B thought that homeowners in arrears on their management fee payments should not be eligible for candidacy on the DVMC.
Mr. H believed that homeowners who are late in their management fee payments should be given some additional time.Aer that, legal methods should be employed.He hoped DVMC members would unite and make their work more transparent.
is record of the meeting captured the deep division between homeowners who paid management fees and those who did not.Arrearages on management fee payments not only affected daily operations of management but also jeopardized the tendering process because the expenditures in the tendering process would have to come from management fees, and the problems with arrearages problems would deter quali�ed management companies from bidding.Nevertheless, many of the DVMC members were themselves fee evaders.In fact, according to the �gure released by the developer, these members collectively owed 659,275 RMB as of May 2000.DVMC tried to resolve the problem by creating a trust account, but the majority of fee evaders still refused to pay the arrearages.e problem was �nally resolved through legal action when the PMB, at the request of DVMC, took several homeowners to court.
e recovery of the majority of arrearages �nally allowed the tender process to occur in December 2002.An evaluation committee, composed of three experts appointed by the government and two homeowners, was established to assess each tender.Six bids were submitted to the evaluation committee.e contract was awarded to a new company.While the developer disbanded the PMB, most of the staff was reemployed by the new company, which marked the beginning of democratic control of the neighborhood.
One important �nding from this analysis is that in China, the process of neighborhood democratization does not necessarily end with the establishment of an HOA.Rather, the establishment is merely a �rst step.It took DV homeowners approximately one year to establish an HOA, but more than two years to assume control over the neighborhood.e resistance of the developer is a key factor, as was found in previous studies.However, con�icts among homeowners also played an important role in this delay between establishment and control.At any given time, multiple actions were undertaken, some of them contradictory.In fact, con�icts over the course of action were always a threat to the unity of homeowners to the extent of severely limiting the ability of homeowners to work towards their common goals.ese con�icts occur because homeowners are far from a coherent group but rather act and think as individuals with their own experiences, skills, knowledge, perspectives, and interests, among other factors.Furthermore, the success of neighborhood democratization is also explained by the fact that homeowners did not acquiesce aer numerous setbacks, delays, disappointments, and tribulations.On the contrary, we can see that actions were constantly evolving due to increased knowledge acquired by individuals who became aware of the de�ciencies in their behaviors.

Impact of Neighborhood Democracy
Because democratic self-governance is considered to be a symbol of a civil society in China, many studies focus on its broader sociopolitical implications.Read [17] argues that because Chinese homeowners are highly dependent on the party-state for their economic success and thus bene�t from reforms, the group is unlikely to push for broader political reform.While this conservative sentiment is shared by domestic scholars such as Xiao [39] and Meng [40], recent assessments have become more optimistic regarding the political impact of neighborhood democracy aer observing homeowners extend their in�uence by developing coalitions of HOAs, submitting legislation proposals to the People's Congress at national and local levels and running as independent candidates for the local People's Congresses [41,42].However, what has been ignored by the literature is the impact of democratic self-governance at the neighborhood level.How does democratic self-governance affect residents' everyday lives?is section considers the impact of such selfgovernance from multiple aspects.
In terms of the quality of governance, the most de�nitive improvement has been in the transparency and accountability of expenditures, which translates into lower management fees.e improvement can be evidenced by comparing the management expenditure statements from 1998 and 2008 (Table 4), indicating that total expenditures decreased in real term from 1998 to 2008 by 14.7%.is decrease in expenditures allowed for a reduction in management fees to the historic low level of 0.55 USD/m 2 , with an aggregated result of a 25.7% drop in income.ese changes are all the more impressive when we consider that the size of the community increased in ten years from 163 households to 228 households-a 40% increase.It is also important to note that the net savings were accomplished without implanting spending cut on services or maintenance.On the contrary, spending on services increased signi�cantly.Rather, large savings were made in administration costs, depreciation taxes, insurance, and utility fees.Some of these savings were so signi�cant that the only plausible explanation was that the 1998 �gures were greatly in�ated.e savings also indicate a conscious effort by homeowners to rein in spending.Overall, the comparison of management expenditure statements provides strong evidence for improved governance.
Democratic self-governance has also made a visible impact on the built form of the neighborhood as many homeowners have conducted extensive home improvement such as the addition of a porch or a facade, the enlargement of the house, or a complete remodeling of the house (see Figures 1 and 2).ese practices were strictly prohibited when the developer was in control of the neighborhood for the sake of aesthetic coherency, as all of the dwellings in DV were designed by a single Canadian architect.Modi�cations to the exterior of the houses thus become very noticeable and introduce a sense of visual diversity to the neighborhood.is  relaxation on home improvements re�ects the homeowner's perception of the neighborhood as a space of conspicuous consumption (c.f.[43]).e celebration of choice and individuation through consumption, a staple of neoliberalism, is particularly evident [44].However, not all homeowners are pleased with this change, as one interviewee complained to me stating the following: my ne�t door neighbors seemed to be a bit sel�sh and all they wanted was to make their private backyard bigger.ey knocked down some trees and fences so they could make the public lawn their own private backyard.
Democratic self-governance also has less material impact on residents' everyday lives.Security guards, stationed on every street corner when the developer was in control, have been removed to outside the compound because their presence was perceived as an invasion of the residents' privacy.Moreover, DV was designed as a lifestyle gated community [32,45].Accordingly, the developer built a multifunctional clubhouse and held a myriad of community activities, and both the clubhouse and the activities have undergone signi�cant changes under the reign of DVMC.For example, community activities were gradually reduced from more than thirty per year to two per year.Additionally, the clubhouse was perceived as a threat to the exclusivity of the neighborhood because it was open to outsiders, and because it was located within the neighborhood, outsiders could gain access to the neighborhood.Consequently, in 2010, when a company expressed interest in reopening the clubhouse, which had been shut down two years prior, DVMC fought �ercely against the idea, going so far as to stage a protest to the district government.Parallel to the �ndings of Pow [46] and Pow and Kong [47], I �nd that the homeowners value and strive to protect the exclusivity and privacy of their gated community.
e classic rational voter model puts forth that, ceteris paribus, individuals are more likely to vote when the stakes and importance of the election are high [48,49].e theory thus predicts lower voter turnout in DV elections aer the exit of the developer as a competing power; however, the voting rate in DV has not proven this theory to be true.On the contrary, a historic high of 80% turnout was recorded in the 2009 election.In the most recent election, held in October 2012, 137 of 228 homeowners sent in ballots.Two reasons explain this sustained level of high voter turnout.Olson [50] posits that group size matters in collective actions.In a small community, such as DV with 228 units, the cost of organizing and promoting an election is lower and the in�uence of a single vote is higher [51].A related factor, proposed by Putnam [52], is that the ability to develop social capital is in�uential in political participation.For DV, even though the level of social activity has declined, the degree of social capital is still high, as demonstrated by the high turnout at annual community activities.
One last point is that, as shown in Table 3, the contemporary DV is completely dominated by Chinese owneroccupiers with very few foreign tenants and absentee landlords.Nevertheless, homeowners remain contested and strat-i�ed, as indicated by the following comment by a member of DVMC: DVMC has no clear mission, no unity of purpose, and meetings were not organized or managed to facilitate good decision-making.Discussions would wander off the agenda.Motions were never proposed and votes seldom taken except by a show of hands on the more banal issues.Meetings became talking shops with no control over discussion with the result that my monthly agendas became longer and longer as new items were added and few were resolved or disposed of.
Even if Chinese homeowners broadly share similar social values, the details of their agendas need not coincide.ese con�icts restrain homeowners in a way similar to how con�icts in interest and strategy did during the early phase of homeowner activism.

Conclusion
e primary objective of this contribution is to demonstrate the value of a longitudinal perspective in producing new insights on homeowner activism in China.ree insights are particularly noteworthy.First, the decrease in rental revenue and occupancy rate and the arrival of Chinese owner-occupiers contributed to the emergence of neighborhood democracy in Dragon Villas.Second, the process of neighborhood democratization does not necessarily end with the establishment of an HOA.Rather, the establishment is merely a �rst step.Furthermore, con�icts between the developer and the homeowners, and among homeowners, played a crucial role in lengthening the process of neighborhood democratization.ird, democratic self-governance has resulted in improved governance, a more diverse built form that articulates individuation through consumption and changes that re�ect the importance of privacy and exclusivity.Taken together, this study has offered concrete evidence on the values and challenges of democratic self-governance in China at the neighborhood level.Given that neighborhood governance is still a new and evolving phenomenon in a frequently changing regulatory environment, future research is needed to determine the full impact of these �rst glimmers of civil society and democratic governance on urban governance in China.
ere are, of course, limitations to longitudinal work.First, there is inherent bias in a retrospective study design, particularly when interviews are involved and subjects are asked to report events some distance into the past (thanks to an anonymous reviewer for raising this issue).Memory becomes distorted by subsequent events and creates errors in the reports.e use of multiple informants can improve data validity and reliability.Moreover, original documents prepared at the time of the actual events can be used as a reliable source of information.Both tactics were employed in this study to reduce the memory bias.Second, it is difficult to �nd a sufficient amount of information over such a long period of time, and this practical problem perhaps explains why there are so few longitudinal studies on neighborhood democracy.ird, this approach generates a large amount of data because of the long timeframe.In this case, the documents I analyzed covered more than 500 pages.erefore, it is difficult to conduct longitudinal studies of this nature beyond few cases.Fourth, because of the problem of attrition, there is always doubt regarding the generalizability of the �ndings.erefore, it is not my intention to argue that the longitudinal approach should replace the crosssectional approach.Instead, both longitudinal and crosssectional studies are essential to the study of neighborhood democracy and democratization.Insights obtained through longitudinal analysis can serve as useful material for crosssectional research and vice versa.

F 1 :
A house in Dragon Villas undergoing extensive renovation (photography by K. Lo).

F 2 :
A house in Dragon Villas undergoing extensive renovation (photography by K. Lo).
T 2: A selection of rental contracts in Dragon Villas, 1995.
Source: Author's compilation using data from the developer.