Abstract
Spatial distribution of urban institutions or property rights is an interesting but little studied topic. This paper analyzes the spatial pattern of homeowners associations (HOAs) in Chongqing, China. By combining data from different sources including a telephone survey, logit models are estimated for the incidence of HOAs. It is found that distance to the nearest subcenter is negatively associated with the probability of forming an HOA in a private community. In contrast, distance to the city center is not significant. A fundamental dilemma is also identified with regard to structural variables such as community size and heterogeneity of interests. Although bigger and more heterogeneous communities have higher potential demand for HOAs to facilitate collective decision-making, forming an HOA is by itself a collective decision that is more difficult to make in those communities. The results of the tradeoff, in the case of Chongqing, are that community size is positively associated with the incidence of HOAs while the number of building types, a measure of heterogeneity of property interests, is negatively associated with the probability of forming an HOA.
Keywords
Introduction
Among a large body of literature on private communities worldwide, little attention is paid to the spatial dimension of their institutions. For example, a fundamental empirical question remains open: is there a spatial pattern of homeowners associations (HOA)? This question may not be very meaningful in countries like USA because an HOA is required by law to be established by the developer and, consequently, the spatial distribution of HOAs is the same as the distribution of private communities in those countries. However, since an HOA is optional in China (as well as many East Asian regions and countries such as Hong Kong and Taiwan), the Chinese city provides an important opportunity for us to study the spatial pattern of HOAs. More generally, research in this vein can help us better understand urban institutions or property rights inside a city from the spatial perspective.
The spatial perspective is the traditional focus of some disciplines such as geography and urban planning. People’s activities and their interactions vary with their residential locations inside the city. Since people’s demand for collective action at neighborhood level arises from those activities and interactions, spatial location affects the demand for collective action, which is the purpose of urban institutions and property rights arrangements. Moreover, people’s activities and interactions at a spatial location are decentralized in nature and often cannot be easily measured. It is then difficult to analyze how they affect urban institutions without introducing the spatial perspective.
The spatial perspective is also important for policy purpose. Since laws have been enacted that require grassroots governments such as street offices in Chinese cities to play some role in the formation and operation of HOAs (Deng, 2014), their attitude towards HOA formation can be described as “responsive” in the sense that, on the one hand, they wouldn’t promote it unless homeowners make the demand; on the other hand, they wouldn’t impede it given that it is their job stipulated in the law. Of course, they want to control the HOA by influencing its composition, especially the selection of the board director. Put another way, grassroots governments’ impact on the incidence of HOAs is very limited, but their influence on the composition of the board of directors might be paramount. 1 The only exception is Shanghai, where the local government was very active in promoting HOAs. It is debatable if government intervention is desirable regarding whether or not to form an HOA. In the case of government intervention, policies that differentiate between different urban areas might be desirable if there is a spatial pattern of HOAs.
The research questions addressed in this paper include: first, is there a spatial pattern of HOAs in Chinese cities? Second, what are the characteristics of the spatial distribution of HOAs? More generally, what factors affect the spatial variation of private community’s institutional structure? Answers to these two questions have implications in the general relationship between urban spatial structure and urban institutions or property rights.
The research methodology in this paper combines quantitative analysis with supplementary qualitative study. Quantitative analysis is based on a telephone questionnaire survey conducted by the author. The author also held several semi-structured interviews with some street office directors. Those interviews provide important insights into the process of forming an HOA.
The study area is Chongqing metropolitan area in southwestern China. Data come from three sources including a telephone questionnaire survey that was conducted by the author. Logit models are then estimated for the incidence of HOAs or the probability of forming HOAs in a private community. The most important finding is that the incidence of HOAs is negatively associated with the distance to the nearest subcenter. Put another way, an HOA is less likely to be formed in a private community that is further away from the nearest subcenter. This finding supports my theoretical argument that there are less externalities and public incomes in private communities far away from subcenters and, hence, less demand for HOAs to facilitate collective decision-making and collective action.
In contrast, distance to the city center is not significant, suggesting that the variable has little impact on the incidence of HOAs. That finding may be due to the fact that Chongqing is a polycentric city and the city center is much less important than in other cities. Another plausible explanation is that the spatial impact on the formation of HOAs may be limited to local regions. City-wide spatial pattern of HOAs might not exist.
A fundamental dilemma is identified with regard to the effects of many structural variables such as community size and heterogeneity of interests. On the one hand, those structural variables that make it more difficult to make collective decisions increase potential demand for an HOA, which is an institution to facilitate collective decision-making. On the other hand, forming an HOA is by itself a collective decision that needs to be made by the homeowners in the first place. It is then more difficult to form an HOA in a community with higher values of those structural variables. The net effect depends on the tradeoff of the above two sides. It is found in this research that community size is positively associated with the incidence of HOAs and the number of building types, a measure of heterogeneity of property interests within a private community, is negatively associated with the probability of forming an HOA.
The contribution of this paper to the literature includes the following: first, this is the first empirical study that analyzes the relationship between urban spatial structure and urban property rights or institutions at large. It contributes to our understanding of the spatial variation of urban institutions. Second, it contributes to Olson’s (1965) theory by focusing on an organization that, on the one hand, facilitates collective decision-making and, on the other hand, requires collective decision-making for its own formation. Third, there is a debate (e.g. Foldvary, 1994; McKenzie, 1994; Zhou, 2014) about whether government intervention is desirable with regard to HOAs. This paper offers a spatial perspective to think about the issue by suggesting that government intervention might only be necessary in some remote regions within the city.
The next section is the literature review, followed by a brief introduction to HOAs in urban China. A theoretical framework is introduced in the fourth section, which ends in a theoretical model for the incidence of HOA. The fifth section describes the study area and data. Empirical analysis is presented in the sixth section, followed by more discussions on relevant issues in the seventh section. The last section is the conclusion.
Literature review
There have been many studies on HOA in the English literature and they can be divided into two camps: for and against HOA, respectively. The camp for HOA often emphasizes its voluntary nature and the formation of a market of private governance that provides local collective goods. Prominent writers in this camp include Foldvary (1994), Beito et al. (2002) and Nelson (2005). In contrast, researchers against private communities, especially gated communities, often criticize their HOAs for a fortress mentality (Atkinson and Flint, 2004; Blakely and Snyder, 1999), discourse of fear (Low, 2003) and restrictions on democracy (McKenzie, 1994). Economic models for private communities include Barzel and Sass (1990) and Helsley and Strange (1998). Empirical studies on this topic include Barzel and Sass (1990), Barton and Silverman (1994), Gordon (2003) and some recent ones such as Cheung and Meltzer (2013).
There is also a strand of growing literature on HOAs in China. First, many researchers (e.g. Read, 2003; Tomba, 2005) have shown enthusiasm for community self-governance through HOA. This enthusiasm is also abundant in the Chinese literature (e.g. Chen, 2009) and Chinese mass media. However, Fu and Lin (2014) pointed out the limitation of HOA as a civic organization. Second, tension between HOAs and PMCs is widely reported in the literature and mass media. For example, Wang and Clarke (2021) documented that many PMCs and local governments hold a mistrustful or even hostile attitude towards HOAs. Chen (2009) collected many stories of fighting between HOAs and PMCs. Third, most researchers agree that the Chinese state plays an important role in neighborhood governance. Deng (2014) analyzed the constraints on HOA imposed by the state. In a series of papers (e.g. Lu et al., 2020; Wu and Zhang, 2022) Wu and his co-authors emphasized the importance of the state in understanding neighborhood governance in China and regarded HOAs and PMCs as “market-like instruments” deployed by the state for economic growth.
Studies on the spatial pattern of urban institutions or property rights are limited. Some empirical studies started to document their spatial patterns. For example, Le Goix (2005) studied spatial locations of gated communities in Southern California and found that large ones are located in suburban or remote areas while small ones are clustered around subcenters (instead of the city center). Song and Liu (2017) observed that many middle-upper class gated communities are located in scenic areas in Nanjing, which are often quite far away from the city center. They found a similar pattern to Le Goix (2005), in which large ones are located in the suburbs while small ones tend to be located around the city center. However, these two studies don’t pay any attention to the institutional structures of private communities, let alone their spatial variation.
Two theoretical papers by Deng (2009, 2017) started to explore the spatial pattern of urban institutions or property rights arrangements. In the first paper, Deng (2009) revisited the Coase Conjecture on durability and monopoly by constructing a model of intertemporal price competition in land monopoly. The interesting feature of the model is the comparison among four institutional arrangements based on the combination of land tenure (rent or own) and local governance (public or private provision). The general idea behind his model is that intertemporal externality may cause different forms of urban institutions to thrive in different scenarios. The same reasons can also explain why most private communities are located in the suburbs. Deng also found that private communities are mostly small and for middle-upper class, and leasehold community is more desirable for the poor.
In the second paper, Deng (2017) extends the Hart and Moore (1998) model to the urban setting by analyzing ex post pricing efficiency of HOA vs. leasehold-type integration. The latter stands for leasehold-based institutions like shopping centers and malls. He found that, first, in a less competitive market, such as a city center, an HOA becomes less efficient and leasehold becomes more attractive. Second, a rich community is likely to prefer an HOA while a poor community prefers leasehold.
It is necessary to point out that the above studies are more or less based on Western institutions and may not be easily applicable to China. For example, because private communities (or developers) in the USA are required by law to establish HOAs, there is no variation of institutional structure within private communities. The above studies simply equate private community with HOA and, consequently, focus on the comparison between private community and leasehold. This is in sharp contrast to China, where an HOA is optional and, hence, there is large variation of institutional structure across private communities. So, a new framework is needed in order to study the spatial pattern of HOAs in the Chinese city.
Introduction to HOA and PMC in China
Residential landscape in Chinese cities is dominated by micro-districts (MD or xiaoqu in Chinese), a type of private community that is similar to a condominium. MD was a planning concept from former Soviet Union that Chinese planners learned in the 1950s. It referred to residential development with integral design and planning; nowadays it is mostly in the form of a private gated community.
A property management company (PMC) is usually hired by the developer immediately after an MD is built. The PMC is responsible for providing some local public goods, such as security, street cleaning, landscaping as well as maintenance of public facilities like elevators, within the MD. Sometimes the PMC is also expected to regulate some homeowners’ behavior although PMC staff often complain that is beyond their duty and expertise. Homeowners need to pay a monthly fee called property management fee to the PMC. The PMC is often the only governing institution in the MD till an HOA is formed.
An HOA needs to be voluntarily formed by the homeowners in an MD. In this sense, it is optional. In urban China local government consists of four levels: city, district, street office and residents committee (sometimes called Community in recent years). 2 During the process of forming an HOA, the street office and the residents committee can provide some guidance in addition to performing the role of regulator. For example, the first condition for forming an HOA in Chongqing is that over 20% of homeowners request the street office to hold a meeting of the homeowners assembly (Ye Zhu Da Hui). Then a preparatory committee should be set up that consists of the representatives of the homeowners, the street office and the developer (if still present). The homeowners’ representatives should account for at least 50% of the preparatory committee members and the head of the committee should be from the street office. 3 The preparatory committee also needs to draft the Homeowners Assembly Decision-Making Rules (Ye Zhu Da Hui Yi Shi Gui Ze) and Management Agreement of the MD (Guan Li Gui Yue). Finally, in its first meeting the homeowners assembly needs to vote for forming the HOA and approve the two legal documents.
Although the street office and residents committee are legally responsible for the “guidance, monitoring and management” of MDs (Deng, 2014), it is fair to say that most Chinese local governments more or less take a merely “responsive” attitude towards the formation of an HOA. On the one hand, local government regards HOAs as a potential threat to its rule. In my interview with a director of street office, he clearly feared that some HOAs might be controlled by people hostile to the government and threaten social stability. On the other hand, government officials understand that neighborhood governance in a private community needs an HOA that represents the interests of homeowners and can facilitate their collective decision-making. Guiding the process of forming an HOA is also their job, which is clearly stipulated in the law. As a result, government officials hold a “responsive” attitude towards forming an HOA in the sense that they don’t actively promote or impede the formation of an HOA; they simply respond to homeowners’ demand once it appears. That might be one reason why about 80% of MDs in big Chinese cities haven’t established an HOA (Zhou, 2014). An interesting exception is Shanghai, where HOAs have been formed in over 80% of MDs thanks to the promotion of local government. It is debatable whether the Shanghai model is desirable or applicable in other Chinese cities.
Once an HOA is formed, it can fire the PMC that is currently operating in the MD. Actually, that is also one reason for many MDs to form HOAs. In this sense, the competition between PMCs is quite strong and becomes the basis for a market of PMCs in neighborhood governance in China. That may be another reason for why the majority of MDs in big Chinese cities haven’t formed an HOA.
It is clear that the demand for HOA does not come from the street office, which simply responds to homeowners’ demand as a daily routine job. Thus, the incidence of HOA is not determined by local government. However, for political purposes local government does want to influence the composition of the board of directors. This is clearly reflected in the procedure of forming an HOA, in which the head of the preparatory committee should be from the street office rather than among the homeowners. In my interview, a director of street office told me that there are usually at least two types of people who want to become the director of an HOA. One type is a person who is truly interested in public affairs and wants to serve the homeowners. This type might be called an activist. Another type is a person whose eyes are on monetary public incomes of the community such as advertisement fees and parking fees. According to the director, the street office wants to make sure the first type of person is elected the director of an HOA. Of course, he also admitted that “we want to elect an HOA director who is willing to cooperate with the street office”.
Theoretical framework
Forming an HOA, just like running an HOA, is a process of collective decision-making in a private community. 4 Hence, both demand and supply of HOA need to be considered when we analyze the phenomenon.
Olson’s (1965) theory on collective action and groups is an important one to the governance of private community. It is based on an important insight: as the group size increases, collective benefit per member decreases. Besides, some collective benefits are also public goods that are non-exclusive and non-rivalry and could easily result in the free-rider problem. Hence, a rational individual will not contribute to the collective good(s) produced by and to the group. Olson’s solution is to provide members with personal benefits so that they have “selective incentives” to participate in the production of collective good. In general, collective action theories predict that structural variables, such as group size and heterogeneity of interests, will be important determinants of the success of collective action.
There is a fundamental dilemma regarding the effects of the above structural variables. On the one hand, an HOA is an institution that is designed to facilitate homeowners’ collective decision-making. Hence, structural variables that make it difficult for the homeowners to make collective decisions also present stronger demand for the HOA to facilitate collective action when the need arises. In other words, a positive relationship is expected between those structural variables and the demand for HOAs. On the other hand, the formation of an HOA is by itself a collective decision that needs to be made by the homeowners in the first place. Those structural variables that generally impede collective decision-making also make it harder for the community to form an HOA. In this sense, it is expected that those structural variables, such as community size and heterogeneity of interests, are negatively associated with the incidence of HOAs. The final net effects of the structural variables depend on the tradeoffs from the above fundamental dilemma facing a private community in China.
One factor that is missing in Olson’s theory is the history of a group. No doubt the longer the history of a group, the more familiar are the members with each other and, consequently, the easier it is for them to reach consensus or make collective decisions. In the context of private community, this means that an important structural variable is the age of the community. The earlier the community was built and the longer the age of the community, the easier it is for homeowners to take collective action including forming an HOA. Besides, many defects in the design of the community gradually become visible and apparent after residents have lived within it for some time. Depreciation of buildings and public facilities also becomes a serious problem over time. It then becomes imperative for the residents to take collective action against those defects and depreciations and the importance of HOA in facilitating collective action becomes evident.
What are possible spatial factors to the distribution of HOAs in urban China? Two factors stand out in terms of their impacts on the demand for HOA. The first factor is based on the fact that suburbanization has been happening in Chinese cities, many of which are developing into polycentric cities. A polycentric city is characterized by the existence of subcenters besides the city center or CBD. It is well established in urban studies that residential density in the suburbs is generally lower, although residential developments appear to cluster around new subcenters. This means that there are less externalities both within and between private communities in the suburbs, especially those far away from subcenters, than in the city center or subcenters. For example, lower density in a Chinese city means there could be more open space, green area as well as more privacy. 5 Change from higher density to lower density is also often associated with change in building type, such as from high-rise building to garden building (Hua Yuan Yang Fang) or villa. Consequently, residents avoid many problems from interacting with neighbors, such as someone planting vegetables or flowers on the roof top, someone smoking inside the elevator as well as garbage thrown from upper floors down the building, a very common problem in a high-density MD. All those problems need to be tackled by a collective organization such as an HOA. Thus, the demand for HOAs to facilitate collective action is generally lower in those lower-density communities far away from city center or subcenters than in those higher-density ones close to the centers.
The second factor is related to the incomes of an MD. An MD in a Chinese city often has some monetary incomes such as fees paid by advertisement companies for advertisements inside the MD and parking fees by drivers from outside. There are usually many advertisements from nearby stores in the form of billboards inside an MD and/or posters on the interior walls of elevators. Apparently, near the center or subcenter there are more stores or commercial activities around an MD that want to do advertisement and the quantity of advertisements is also related to population density, which varies spatially. Hence, advertisement income of an MD is higher if it is located closer to a subcenter or center. Another important source of income is parking fees. Many MDs allow cars from outside to park in its parking lot, thus earning an income. This source of income also varies with spatial location because the demand for parking space is much stronger near the center or subcenter.
In the absence of an HOA, those incomes are controlled by PMC. Therefore, an important reason to form an HOA is to control those incomes, as one director of street office told us in the interview. They even found that some homeowners want to be elected the board director of the HOA for personal gains. As a matter of fact, there are some sporadic reports of corruption related to some HOA board directors. Overall, controlling public incomes to an MD is an important motivation for homeowners to form an HOA. However, since those incomes vary with spatial location, the incidence of HOA should also be associated with an MD’s spatial location.
Although property management service appears to be a factor to the formation of HOA, it is not included in the theoretical model after a second thought. The reason is that property management service is rather standard and many issues between homeowners and PMC arise from either externalities and public incomes, which are captured by the spatial variables, or defects in the physical and institutional design of the community, which are idiosyncratic and not covered by a variable of property management service.
Is it possible that the presence of government is more visible and accessible in locations near the center or subcenter, thus impacting the spatial incidence of HOA? The answer is negative. First, in my interviews with street office directors, they don’t think there is much difference in terms of the presence of grassroots government. Second, street office is responsible for guiding the formation of HOAs while district might care more about subcenter. Those two matters are not related. Third, as pointed out in the preceding section, grassroots government’s attitude towards HOA formation is “responsive”. Hence, the presence of grassroots government may not be an important factor when homeowners initiate the process of forming an HOA.
The above theoretical framework suggests the following theoretical model for the incidence of HOAs in Chinese cities:
I denotes the incidence of HOAs or the probability of forming an HOA in an MD, x is a vector of non-spatial structural variables including community size, heterogeneity of interests within the community and its age, and d and d_sub denote distances to the city center and the nearest subcenter, respectively. The last two variables describe the spatial location of an MD in a city.
Study area and data
The study area is Chongqing metropolitan area, the fourth municipal city in China that is directly under the administration of the central government. The metropolitan area is comprised of nine districts that are connected by bridges and tunnels. Two rivers, the Yangtze River and the Jialing River, run through the city in an east-west direction while three mountain ranges cut across the city in a north-south direction. It is due to this unique topography that Chongqing is often called the “mountain city” in China.
The city of Chongqing was founded very early, but its first period of rapid development was in the WWII when it became the wartime capital of China. Many people from Eastern and Northern China migrated to Chongqing in the face of Japanese invasion. Many factories, especially those in the defense industry, were also moved to Chongqing by the government. After the People’s Republic of China was founded in 1949, it became part of Sichuan Province. Although the city was known to be comprised of several relatively independent urban areas, Jiefangbei has long been recognized as the city center, at least before the onset of economic reform. In 1997 Chongqing was designated by the central government as the fourth municipal city and since then has been developing very fast. Like many other Chinese cities in this period of time, Chongqing’s urban development is characterized by large-scale urban expansion. Although the city grows in all directions, its main direction of expansion is north due to topographical reason as well as the city’s master plan. The former Jiangbei County (in the north) was annexed and renamed Yubei District. It is fair to say that the city has been expanding in a typical fashion of suburbanization and urban sprawl (Liu et al., 2018).
Chongqing is often regarded as a typical case of polycentric city (Liu and Wang, 2016). Nevertheless, a city that includes several “groups” (zu tuan) of urban areas separated by rivers and mountains is not necessarily a polycentric city simply because the commercial centers in those groups of urban areas may not have passed the threshold for a true subcenter. That was the case for Chongqing, at least before the onset of economic reform. After recent four decades of urban expansion, many new subcenters have been emerging. Some such as Guanyingqiao have even surpassed Jiefangbei in terms of its importance to local people. The Planning Bureau of Chonqing City designated in 2005 six subcenters: Shapingba, Yangjiaping, Guanyingqiao, Nanping, Xiyong and Chayuan. In 2011 eight more subcenters were added in the city’s new master plan. It is apparent that Chongqing is now a polycentric city and is also becoming more polycentric in the future.
This research is based on data from three different sources. The first source of data is a list of 724 MDs that was obtained from an advertisement company in 2017. Some news media reported that the number of MDs in Chongqing was about 2000. That list then covered about one third of the population of MDs, although the latter is undoubtedly growing in recent years. This dataset includes some important variables such as the number of households in an MD, the number of entrances or sections (men dong) in the buildings and the age of the MD. Number of households is an apparent proxy for community size while the number of entrances or sections in the buildings is less apparent. It is reported in the literature that many HOAs in China depend on building/flat captain system for efficient organization (Wang et al., 2012). One captain is selected for one entrance in the buildings. Thus, the number of entrances in the buildings can also be a proxy for community size from that perspective.
The second source of data is online web site anjuke.com , a major housing web site in China. Multistage stratified sampling is employed before collecting housing data from the online sources. In the first stage, sample size in each district is set to be half of the number of MDs in that district included in the list of 724 MDs. In the second stage, systematic sampling is adopted for which the sampling frame is MD’s name. Put another way, I sorted the list of MDs in a district by their names (in alphabetic order of their pinyin names) and then every other MD was chosen to be included in the sample. In the end, the sample size is 296. Housing-related data such as FAR and green area ratio were also collected from anjuke.com. Spatial location data such as distance to the CBD (Jiefangbei) and that to the nearest subcenter were obtained in Baidu Map, a popular map APP in China that can provide shortest driving distance to a location.
The third source of data is a telephone questionnaire survey based on the same sample that was conducted by the author in the Fall of 2018 and the Spring of 2019. We first searched on anjuke.com for the names of PMCs in those 296 MDs and then looked up for their phone numbers in Chongqing 114 Life Guide, a WeChat public account of phone directories. After obtaining a PMC’s phone number, we then called it and tried to ask the questions in the questionnaire survey that are brief and mostly about both HOA, if it has been formed, and PMC. Due to labor availability and the difficulty of both finding the phone numbers and doing telephone survey, we ended up doing three rounds of telephone surveys. After each survey the author randomly called about one fourth of the phone numbers that were just surveyed in order to verify the accuracy of survey responses. The final number of valid questionnaires is 135. The response rate is about 45.6%. 6
Apparently, our sample doesn’t include those MDs without PMC. According to our knowledge, PMC is almost ubiquitous in Chongqing. Even public rental housing hires PMC. The only few MDs we know that do not have a PMC are low-income housing, especially resettlement housing for former peasants who lost their land in urban development. Those poor MDs couldn’t afford the property management fee. Besides, there is no fee for a company to have its phone number listed in 114 phone directory. Thus, there is no reason to believe that 114 phone directory only includes those more marketized PMCs. In summary, the sample is representative of the population of private communities in Chongqing and doesn’t produce significant bias. 7
The most important question in the telephone survey is whether or not an HOA has been formed in an MD. It was input into the database as a dummy variable (formHOA). Another important variable generated from this dataset is the number of different types of buildings in an MD. In the telephone survey, the respondent was asked to select from the following building types: multi-story building (duo ceng), high-rise tower building (gao ceng), garden building (hua yuan yang fang) and villa (bie shu). The variable, bldgtypes, records how many types of building there are in an MD. Because some MDs have more than one type of building, people living in different types of buildings, albeit within the same MD, may have very different property interests. Thus, that variable is a good proxy for the heterogeneity of property interests in an MD.
Combining data from the above three sources results in a dataset with sample size equal to 135. It should be noted that some variables have missing values. When they are included in regression analysis, observations with missing values are automatically excluded by Stata.
Empirical analysis
The definitions of variables are provided in Table 1 and their descriptive statistics are provided in Table 2. It is clear from Table 2 that community size as measured by number of households varies greatly across communities, with the smallest being 45 and the largest being 6000. The number of entrances in the buildings varies from 1 to 46. Most MDs have only one type of building while some have three types of building within their territory. The newest MD was built about 3 years ago and the oldest MD’s age is more than 20 years. Its standard deviation is 3.3 years, indicating that most MDs were built less than ten years ago. The MD closest to the city center is about 2 km away while the furthest distance to the city center in our sample is 32.4 km. In contrast, the shortest distance between an MD and the nearest subcenter is only 0.3 km while the longest distance is 13.7 km.
Explanations of variables.
Descriptive statistics.
Because the dependent variable is a dummy variable, logit model was chosen for the empirical analysis. Three logit models were run in Stata and their results are presented in Table 3. Since my interest is in the spatial pattern of HOAs, Model 1 includes only two spatial variables: d and d_sub. More explanatory variables, basically control variables such as number of households (hhnum), number of building types (bldgtypes) and age of MD (age), are included in Model 2. The difference between Model 2 and Model 3 is that the proxy for community size is changed from number of households (hhnum) in Model 2 to number of entrances to buildings (doornum) in Model 3.
Results of logit models.
significant at 10% level. **significant at 5% level. ***significant at 1% level.
Pseudo R2 is very low for Model 1. That is not surprising since it includes only two spatial variables. Pseudo R2 for Model 2 is 0.2014, meaning that about 20% of the dependent variable’s variance can be explained by the model. The explanatory power of Model 3 is close to Model 2.
Logit models yield several results. First, the most important finding that is consistent across the three logit models is that distance to the nearest subcenter is significant at 5% level and the sign of the regression coefficient is negative. It means that the further away from the nearest subcenter, the less likely for an MD to form an HOA. In other words, HOAs tend to be located close to subcenters and their distribution in remote places tends to be limited. In contrast, distance to the city center is insignificant, implying that the city center has little impact on the incidence of HOA except for those MDs located close to it. In the latter case, the city center is the same as the nearest subcenter.
Second, number of building types is significant at 1% level in Model 2 and significant at 5% level in Model 3. The signs of its coefficients are all negative. This means that the number of building types in an MD negatively affects the probability of forming an HOA. Since the number of building types is a measure of heterogeneity of property interests inside an MD, the finding indicates that its effect on the difficulty of forming an HOA outweighs the effect on the demand for HOA as an institution to facilitate collective action.
Third, two measures of community size, i.e. number of households (hhnum) and number of entrances to buildings (doornum), are significant in Model 2 and Model 3, respectively. Their coefficients are all positive, indicating that community size has a positive impact on the probability of forming an HOA. This result implies that the effect of community size on the demand for HOA outweighs the effect on the difficulty to form an HOA.
Fourth, community age (age) is significant at 5% level in both Model 2 and Model 3 and its coefficients are positive. This finding is not surprising because, as my theoretical analysis shows, community age is positively associated with both the ease of forming an HOA and the demand for HOA. On the one hand, the longer the age of an MD, the more familiar are people with each other and the easier their collective decision-making including forming an HOA. On the other hand, many defects in the design of an MD become visible only after long time of use, let alone depreciation of buildings and public facilities. Hence, demand for HOA to facilitate collective action increases with the age of a housing community. The above two effects both predict that community age is positively associated with the probability of forming an HOA.
Logit models in which the explanatory variables are transformed by natural logarithm function are also estimated. 8 The results are presented in Table 4. In general, the results are similar to those in Table 3. In Model 4, log value of distance to the center (ln(d)) becomes significant at 10% level. In Model 6, log value of distance to the nearest subcenter (ln(d_sub)) and that of building types (ln(bldgtypes)) are now both significant at 1% level. Pseudo R2 values are higher in Model 4, 5 and 6 than in Model 1, 2 and 3, respectively. In a word, results found in Table 3 are slightly strengthened in Table 4 after log transformation of explanatory variables.
Results of logit models on log variables.
significant at 10% level. **significant at 5% level. ***significant at 1% level.
Discussion
My empirical analysis indicates that the spatial pattern of HOAs is evident with regard to their relationship to the nearest subcenter. However, no impact from the city center is detected. Two possible explanations are advanced here. First, Chongqing is often regarded as a typical polycentric city in China and, hence, the importance of the city center to MDs across the whole metropolitan area becomes very limited, if not trivial. Second, the spatial impact of urban structure on the incidence of HOA may be limited to local regions around subcenters. Therefore, the city center might be too far away to have any effect on the probability of forming an HOA. It is necessary to note that the above two explanations are not mutually exclusive.
So far my analysis of the spatial factors to the incidence of HOA is restricted to some simple and intuitive reasoning related to the formation of HOA and the demand for collective action within a private community. It is possible that some more complex and sophisticated mechanisms might be working. For example, previous studies (Deng, 2009, 2017) on the relationship between urban spatial structure and urban institutions focus on private community that is governed by HOA. CID (Common Interest Development) is a typical example in USA. They haven’t paid much attention to the phenomenon of a private community being governed by a PMC only. The latter institutional structure is quite common in private communities in East Asia including Hong Kong and Taiwan. It is possible that different institutional structures of private community may have different implications in terms of intertemporal externality or ex post pricing efficiency, which then lead to different spatial patterns of its governing institutions. The scope of this paper determines that this topic has to be left for future research.
A fundamental dilemma regarding structural variables, such as community size and heterogeneity of interests, is identified in this research. On the one hand, those communities in which collective decision-making is more difficult are certainly less likely to form an HOA; on the other hand, they face potentially stronger demand for HOA, which is an institution designed to facilitate collective decision-making and collective action. The tradeoff between the two effects may vary across different communities. It is found in this research that, with regard to community size, the net effect is dominated by the effect on the demand for HOA, but with regard to heterogeneity of interests, the net effect is dominated by the effect on the difficulty of forming an HOA. It is possible that the outcome of the tradeoff may be different in other cities or regions.
A practice of some developers in designing some MDs features several different types of buildings inside one MD. My empirical result indicates that it is difficult to form HOA in this type of MDs because people’s property interests may be very different across different building types. For example, the issues concerned by homeowners in garden buildings may be concentrated on landscaping and green area while the attentions of those in high-rise buildings focus on someone planting vegetables on roof top or garbages thrown from upper floors. The management of this type of MD is more difficult than those with only one type of building. Hopefully in the future the developers will respond to the management problem and accordingly revise their designs.
The fundamental dilemma mentioned above is also a practical one in urban China. A large and heterogeneous community potentially has stronger demand for HOAs, but the obstacle to the formation of an HOA is also formidable. What are possible solutions to this dilemma? How could the process of forming an HOA be triggered? First, the defects in the design of private community or the depreciation of buildings and facilities are often triggers to the formation of an HOA. The growth of HOAs in many private communities in China reflects this line of story. The second solution is local government’s assistance or even promotion like in Shanghai. Of course, it is debatable whether local government should be actively involved in the initiation of forming an HOA. A third solution, which is also an extreme version of the second solution, is a mandatory requirement for the developer to establish an HOA once a private community is built. That might be the reason why US laws require developers to set up HOAs in private communities.
More importantly, the findings in this paper remind us that it is necessary and also important to think about the spatial dimension of the above dilemma. If more government intervention is deemed necessary in the initiation of forming an HOA, it should focus on private communities located far from subcenters, according to my empirical result. Location with regard to the city center is not important, at least in Chongqing.
Finally, an HOA is just a vehicle to facilitate collective decision-making in neighborhood governance. Many HOAs may be inactive after being formed in an MD. This suggests that, on the one hand, some MDs might be in a more harmonious situation than the others and they have less need for an active HOA; on the other hand, forming an HOA is not a promise for good neighborhood governance. It is just the start of a long journey toward good governance.
Conclusion
It is found in this research that there exists a spatial pattern of HOAs: those private communities located further away from the nearest subcenters are less likely to form an HOA. This may be due to the fact that residential density is lower in areas further away from the nearest subcenter, resulting in less externalities among homeowners. Another reason is that MDs in those locations may have less public incomes, which is an important motivation for people to form an HOA.
Distance to the city center has no impact on the incidence of HOA. Two reasons are provided although they are not mutually exclusive. First, Chongqing is a polycentric city due to its mountainous topography and, hence, city center is not as important as in other cities. Second, the spatial factors that affect the incidence of HOAs might be limited to local region.
This paper gives a clear and positive answer to the question about the relationship between urban spatial structure and urban institutions. Put simply, it is confirmed that there is a relationship between urban spatial structure and urban institutions, which were often assumed to be a-spatial. The spatial aspect was often ignored in studies on urban institutions such as the debate on private community. My empirical result shows that the debate should also take into account the impact of urban spatial structure. Institutional structure of the private community varies with its spatial location; so does its social and economic effects. Furthermore, from a policy perspective, government intervention in the initiation of forming an HOA might only be necessary in some local regions within the city that are far from the subcenters. This implication is especially important in China, where local governments’ attitude towards HOA varies.
It is clear from this paper that urban institutions or property rights arrangements are not space blind. Careful analysis could reveal that spatial factors do affect institutional structure or even institutional form of an urban phenomenon such as private community. More importantly, space is not an abstract concept; it is embedded in local culture and nation-wide social, political and economic institutions. That is why local government is an important actor in the governance of private communities in China. That may also explain why externality and public incomes are the two major factors that drive the spatial distribution of HOAs in Chongqing.
Limitations of this research include the following. First, heterogeneity of interests in a private community is proxied by the number of building types. It is possible that differences in homeowners’ other characteristics, such as income, may also be important factors to heterogeneity of homeowners’ interests. Large-scale survey could help us collect information about homeowners in private communities and better measure that variable. Second, all MDs in our sample have hired PMCs. A small number of private communities in Chinese cities do not hire a PMC. That could be an interesting albeit rare institutional structure to study.
Some questions remain open for future research. First, theoretical studies in the literature have examined the relationship between urban spatial structure and urban institutions/property rights more or less in the context of Western cities. It is important to follow up with theoretical studies in the context of China and other similar regions or countries. For example, the dominance of PMC is not restricted to China; it also exists in Hong Kong and Taiwan. HOA is also optional in Hong Kong (Walters, 2002). Second, if locations of private communities can be geocoded in GIS, then more accurate and comprehensive study of the spatial pattern of HOAs is desirable.
Footnotes
Author’s note
An early version of the paper was included in ResearchGate Preprint papers.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.