Individual Social Capital and Community Participation: An Empirical Analysis of Guangzhou, China

Abstract

A social capital framework has been widely adopted to interpret participatory behaviors. While there is substantial literature regarding the effects of community-based social capital on grassroots participation, less attention has been paid to the relationship between different sources of social capital and community participation. This is particularly relevant for understanding community development undergone restructuring of individual social capital, such as China. To address this deficiency in the literature, this paper integrates both individual and social capital that is accessed within and outside a community to analyze their relation to different forms of community participation. Multilevel analysis is based on a large-scale community survey conducted in Guangzhou at the end of 2012. The results reveal a shift in social relations such that personal social resources are now mainly accessed outside the community. They further reveal that social resources outside communities are consistently and significantly related to all forms of participation. This implies that although residents’ personal networks have gradually diffused out of their communities, this has not only not reduced their enthusiasm toward the communities themselves but also facilitated participation in community affairs.

1. Introduction

In his seminal work, Bowling Alone, Putnam found that social capital, defined as generalized norms, reciprocity and mutual trust based on collective social networks, is a main driving force behind civic engagement [1]. He further argued that a loss of social capital directly leads to the failure of democracy, to bowling alone and to other social issues [1,2]. Henceforth, the consequences of social capital have received wide attention, and a veritable mountain of research has concentrated on testing the relationships between social capital, civic engagement and regional and national economic development (e.g., [3,4,5]). Given the unambiguous effects of social capital on economic development, the literature has almost entirely accepted the generalized positive effects of communitarianism and trust and portrayed the social capital as the key to a myriad of community problems—from life satisfaction, mental health and population aging to social security and even the rejuvenation of small towns (e.g., [4,5,6]). It has revealed that communities with more social capital (stronger in mutual trust, common reciprocity and unified moral regulations) are less problem-ridden, have better basic infrastructure, better government, higher levels of employment, and higher levels of health and happiness of residents (e.g., [4,7,8]). As the significant contribution of social capital to community participation (e.g., [1,3,8]), the social capital framework has been adopted as an important construct for interpreting participatory behaviors.

In light of the importance of social capital in Chinese society, the social capital framework is particularly relevant to understanding behaviors of grassroots participation in urban communities. As in market societies, the low degree of Chinese participation in community affairs has been proven to be due to weaker social capital within communities [9]; these communities were once places where intimate social interaction occurred, but they have changed into locales comprising strangers who do not know each other’s names, do not trust each other and even safeguard against each other [10,11,12,13]. Due to China’s special culture and regime background, it presents a somewhat different picture from the prevailing situation in Western countries. China’s rural–urban difference in social capital volume leads to different levels of community participation [12]. Rural communities tend to have higher proportions of grassroots organization voter turnout than urban communities due to stronger mutual trust and attachment to the community as well as closer social ties [12]. Residents of urban communities are less participatory due to lower volumes of social capital [10]. Moreover, new forms of communities have proliferated in the post-reform era. Newly built commodity housing estates and urban villages are associated with lower levels of social capital and participatory behaviors than traditional work-unit yards [13,14,15].

Despite the surge of interest in social capital and its effects on community participation, less attention has been paid to the relationship between different sources of social capital and community participation [16,17]. Extant studies mainly focus on the social capital within the community, i.e., neighbor acquaintances, community cohesion and solidarity that is cultivated from in-group interaction and networks [18,19,20]. However, the literature fails to recognize the unequal social resources accessed outside a community amongst residents within the same community. This is particularly relevant for understanding participation behavior undergone a restructuring of individual social capital, such as China [21]. Individual social capital in China has experienced restructuring in a way that has gradually dispersed beyond community boundaries. Employment relationships may have replaced neighborhood relationships as the main sources of individual social capital [13]. This poses the question of whether a large amount of social resources and support available from an extra-community decreases residents’ passion for their current communities and reduces their motivation to participate in community affairs. Therefore, this paper aims to explore the relationship between different sources of social capital and community participation. Particular attention is paid to social capital that is accessed outside a community and its potential role in assisting residents to be involved in community affairs.

This study relies on a sample survey conducted in Guangzhou at the end of 2012. Note that the data are a little bit stale; the development of communication technology, policy change, and urban restructuring may have influenced the dynamics of relationships between sources of social capital and community participation. It is, however, an initial attempt to investigate the effects of different sources of social capital in Chinese cities. This study contributes to the literature twofold. First, this study is the first to investigate whether and to what extent the social capital accessed outside a community impacts community participation. Second, current operationalizations of participation fail to measure different types of involvement. This study classifies community participation into three forms based on factor analysis and presents how individual social capital affects all forms of participation.

In the following sections, we first discuss how participatory behaviors are jointly shaped by individual and community social capital, and we then review the literature pertaining to Chinese cities. This is followed by our working hypothesis. Next, we discuss the data and the measurement of variables. Following that, empirical findings based on the survey data are presented. In the final section, key findings are summarized, and implications are discussed.

2. Literature Review

2.1. Relationship between Social Capital and Civic Engagement

As it is a multidimensional construct, prior studies have examined social capital from different dimensions, including the individual vs. the collective, bonding vs. bridging and cognitive vs. structural [22,23,24]. In this study, we focus on individual and community social capital. While acknowledging that these two kinds of social capital are sometimes mutually conducive, we distinguish them herein in order to determine their respective effects on participatory behaviors. As for community social capital, in Putnam’s and others’ view, it refers to the community-based mutual reciprocity, trust and norms based upon in-group social processes and social relations [1,2,25]. The main mechanisms of impact of community social capital on community participatory behaviors are that residents are subject to in-group social norms and expect reciprocal benefits from others, which then mobilizes them to participate in civic affairs. Current studies employing manifold approaches and instruments confirm Putnam’s definition of community social capital as promoting participatory behaviors on the whole. For example, mutual trust and sense of duty have been found to be positively associated with participation in local organizational activities and in collective efforts to form a block association [26,27,28]. Likewise, neighborliness and solidarity among local residents have been shown to facilitate local efforts to maintain community improvement programs and resolve dilemmas involving collective action [28]. Moreover, the in-group trust can be converted into trust in the government, which significantly improves government output and economic growth [5]. However, some research has revealed that community social capital does not facilitate all forms of participation. For instance, Hays and Kogl [22] observed that frequent interactions among neighbors were not associated with participation in organized activities or programs.

Although there have been many studies on the impacts of community social capital on community participation, less attention has been paid to the impacts of individual social capital. As in the work of Lin [19] and Olson [29], individual social capital is defined as a social resource obtained from individuals’ social ties. The influence mechanism for individual social capital on participatory behavior is that people with higher degrees of individual social capital are related to encouraging individual senses of cooperation and cultivating the public spirit, thereby increasing participation propensity [30]. Empirical studies concerning collective affairs have revealed that higher degrees of individual social capital encourage cooperation and, hence, local participation [31]. In contrast, the pathway that lower levels of individual social capital undermine participation initiatives is first making people feel isolated and alienated from social networks and the wider society. Then people in this context have less confidence in their actions and intrapsychic motivations and do not think they can make a difference. Thus, they tend to withdraw from social and political participation. On the other hand, if residents feel isolated and alienated from social networks, this probably further decreases confidence in their actions and intrapsychic motivations, and they tend to withdraw from social and political participation.

In general, there has been little empirical support for the relationship between individual social capital and community participation, except in the literature on contentious participation. Those with extensive, strong social ties are more inclined to protest, as they may be more informed and efficacious [32]. Furthermore, despite the restructuring of social relations that has been induced by changes in urban space, little research elucidates the extent to which individual social resources, including those accessed both inside and outside communities, are related to community participation. To this end, it is essential to explore the ways in which individual social capital is relevant to all forms of community participation and distinguish the relative importance of individual social capital inside vis-à-vis outside community on participatory behaviors [18,19,33].

2.2. Community Participation in China’s Housing Marketization

Participatory behaviors in urban China have gradually changed due to the transformation of urban communities as well as the development of grassroots agencies in recent decades, which is different than in the West. Prior to the market-oriented reform, urban communities were characterized by the state work-unit (danwei in Chinese), which provided services in all spheres. Autonomous, voluntary and grassroots-driven participation was not allowed [12]. However, much of this has changed. Housing com-modification and economic liberalization led to an unprecedented spatial and social restructuring of urban communities in early reform phases [13,21]. Due to the dissolution of the work-unit system, the work-unit became less important in community governance, and residents’ committees (RCs) and nascent grassroots organizations such as homeowners’ associations (HOAs) have played increasingly important roles in community governance, giving residents of various kinds of communities platforms for achieving the common good. The RC is a territory-based social institution created by the central government that has been established in all Chinese cities. It is responsible for the basic social management of urban communities, and its duties include maintaining household registry rolls, translating government initiatives to the grassroots level and executing local government policy [34]. Housing reform has led to high rates of homeownership in urban China, as more than 80 percent of households now own their homes [35]. This homeownership identity leads households to consider their residential areas to be their home territories, generating a sense of responsibility for their living environments [36]. HOAs have emerged more or less spontaneously and have become instrumental in helping manage community affairs in order to champion common interests and assert collective control over the property. Although they have less legal and financial autonomy than their American counterparts, the growth of HOAs in China illustrates, to some extent, an improvement in community governance autonomy. Convenient access to online forums has also helped foster community spirit, hence facilitating young professionals to participate and coordinate their actions to tackle shared problems [37].

While the development of nascent grassroots organizations, homeownership identity and convenient networks encourage residents to become involved in community governance, the downward trend in community social capital and the restructuring of individual social network impact residents’ attitudes toward their communities. Current residential communities have been transformed from locations with dense social networks to places characterized by anonymity among neighbors, experiencing a stark decline in community-based social capital [36,38]. It has been shown that this lower level of community social capital constrains participatory behavior. For example, Gui and Huang [10] and Xu et al. [12] use surveys to corroborate the role of community social capital (e.g., social interactions and perceived mutual reciprocity among neighbors) in facilitating participation in community affairs. Residents of urban villages and commodity housing expressed a lesser sense of belonging to the community and were less socially active compared with those in the work-unit compounds [9].

The effects of individual social capital (Guanxi in the Chinese context) have been shown to be related to individual health, helping one to achieve social goals (e.g., finding a job) and encouraging environmental activism and social movements [16,39,40]. Wang et al. [41], using survey in Shanghai, examined different types of neighboring and revealed that intergroup neighboring helps facilitate participation in community activities. Lin [42] argued that in China, the importance of social resources rivals those of political power and professional skills and is crucial in one’s attainment. However, the consequences of the Chinese restructuring of individual social capital for participatory behaviors have not been widely studied in China. Recent case studies on collective action offer a glimpse into the effects of individual social capital in contemporary China [16,40]. In the cases of “not in my backyard” and “rebuilding Enning Road”, links to authorities, experts and the media allow individuals to access crucial information, seek legitimizing arguments and widely disseminate related information, hence acquiring sufficient public support [16,40]. Individual social capital impacts the development of collective action in that such action is more likely to be successful if the participant’s private social capital is strong and extensive.

2.3. Hypothesis Development

Much more consideration should be given to the association between social capital and community participation. In this paper, we aim to explore whether and the extent to which the extra-community social capital has impinged on participation behaviors across the city of Guangzhou in the 2010s. The proposed model is developed as presented in Figure 1, which consists of three main hypotheses. Based on the review of the literature, we propose the first hypothesis of this paper:

H1. 

After controlling for the demographic variables, community social capital is positively associated with grassroots participation.

Figure 1. The proposed model.

Figure 1. The proposed model.

Due to the restructuring of social networks in contemporary China, there is a great deal of uncertainty regarding the impact of individual social capital on participatory behaviors. We herein distinguish sources of social ties, i.e., social networks inside and outside communities, in order to examine how the individual social capital is related to participatory behavior. The above literature review provides a rationale for the individual social capital inside the community as a positive predictor of participation. Thus, we propose the second hypothesis of this paper:

H2. 

After controlling for the demographic variables, individual social capital inside communities is positively associated with grassroots participation.

As for extra-community social capital, social resources from the extra-community offer close social bonds and support in addition to providing varied platforms for residents’ involvement, which has likely decreased passion for the community as well as time for community activities. In other words, individually based connections probably pull individuals away from forming intentions of community participation. Viewed in this way, social ties from the extra-community exert restraining rather than mobilizing effects. Thus, we propose the third hypothesis of this paper:

H3. 

After controlling for the demographic variables, individual social capital outside communities is negatively associated with grassroots participation.

3. Data and Method

3.1. Data Source

The data for this study are drawn from a large survey completed in Guangzhou by the end of 2012. Guangzhou is located in the central region of the Pearl River Delta and has a population of nearly 13 million in 2012. It is the third-largest city with regard to economy size in China, following Beijing and Shanghai. The city of Guangzhou consists of eight districts (excluding county-level cities such as Nansha and Huadu), as presented in Figure 2. The research team comprised academics from Hong Kong Baptist University, Sun Yat-sen University (Guangzhou), and Duke University and conducted in-door, face-to-face survey interviews with the study participants. The survey provided rich information on neighborly relations, social networks and community activities. A multi-stage stratified random sampling technique was adopted in order to maximize the representation of the sample. During the first stage, within the outer ring road of the city of Guangzhou, the boundaries of three sampling unit strata, i.e., the inner core (52 streets), the inner suburbs (45 streets) and the outer suburbs (42 streets), were drawn based on historical demarcation, land use, population size and population density. In the second stage, communities were randomly chosen (see Figure 2): 17 from 12 streets within the inner core, 14 from 11 streets within the inner suburbs, and 8 from 7 streets within the outer suburbs. Following that, a number of households were chosen from within selected communities according to home address and population size of the total community using an interval sampling strategy. The number of households selected for the survey ranged from 21 to 144. In the last stage, one participant of at least 18 years of age from each household was targeted for the survey. If a household refused to take part in the survey, their next-door neighbor was recruited. During the interviews, more than 30 trained interviewers were recruited. The response rates were high; 1801 out of 1809 participants successfully completed the questionnaires and agreed to engage in the interview at the beginning.

After excluding participants with missing data on control variables and social capital variables, 1774 participants were included in this study. In order to reveal to what extent this sample is representative of the population, we compared the age–gender distribution of the sample with that of the 2010 Population Census of Guangzhou City. As for age, the share of people aged 21–64 is 88.03% of the sample, whereas the share presented by the population, census is 81.91%. As for gender, females made up 55.68% of the sample and 58.03% of the population census. Thus, the demographic composition of the sample was not too much different from that of Guangzhou City.

3.2. Dependent Variable: Measures of Grassroots Participation

Extant studies in the Chinese context have tended to treat community participation as an overall concept. Palmer et al. [31] and Zhu [43] are among the few exceptions; however, the former mainly analyzes the participatory behaviors of migrant workers, and the latter’s dichotomous outcome regarding participation activities is insufficient for determining outlying community participation in contemporary Chinese society. This paper contributes to this scholarship by unpacking community participation into different forms of engagement based on factor analysis. Variables used in the factor analysis were from respondents’ self-reported counts of participation in 11 community activities (as shown in

Table 1. Factor analysis of participation items.

	Factor 1	Factor 2	Factor 3
1. Voting for RC members	0.775
2. Giving advice to RC	0.737
3. Voting for HOA members	0.739
4. Giving advice to HOA	0.728
5. Discussing community affairs at online forum		0.583
6. Complaining about incivility in the community		0.489
7. Refusing to turn in management fee		0.790
8. Protests or petitions			0.587
9. Joint letter			0.652
10. Appealing to the higher authorities for help			0.675
11. Exposing community issues to media			0.678

Notes: Only factor loadings larger than 0.4 are shown.

) during the past 12 months. The final results of the factor analysis using varimax-rotation method by SPSS are shown in Table 1. It reveals that 11 community activities elucidate three distinct categories of participation. The typology captures both conventional and non-conventional facets of participation in urban China. Factor 1 encompasses items regarding whether residents took part in elections for or advised an HOA or RC, which are types of grassroots organizations in Chinese communities, and participation is thus defined as associational involvement. This kind of participation takes a formal, organized and cooperative but less active form. This reflects how residents participate in and interact with China’s grassroots systems of governance.

Comparatively, factors 2 and 3 are more informal, spontaneous and active forms. These reflect two distinct forms of individual activism. These factors compare and contrast activities in which the actor involved are institutional (factor 2) and activities that are not permitted or approved (factor 3). Specifically, factor 2 refers to mild actions that influence community issues, such as discussing community affairs in online forums, reporting concerns about community issues to the government and complaining about incivility in the community. Factor 3 encompasses activities that take a more contentious form, including letter petitions, protests and appealing to the higher authorities for help. In this type of participation, deliberate actions or activities are undertaken in order to influence political outcomes by targeting relevant political or societal elites or organizations. Thus, this requires more initiative and commitment than the first two types. Yip’s [44] studies on housing activism point to important differences between mild action and contentious action in the context of Chinese communities, which supports our interpretation of factors 2 and 3. Following Yip’s research [44], factors 2 and 3 are defined as institutionalized and non-institutionalized action, respectively.

Three dependent variables were calculated using dichotomous outcomes (participants taking part in at least one group activity received a value of 1 or otherwise received a value of 0).

3.3. Independent Variables

Table 2. Descriptive statistics for independent variables (N = 1774).

Dependent Variables	Type	Min.	Max.	Mean	SD
Age	Continuous	18	79	44.58	14.56
Male	Dummy	0	1	0.44	0.49
Married	Dummy	0	1	0.82	0.38
Child in house	Dummy	0	1	0.61	0.49
Years of education	Continuous	6	21	12.93	3.45
Local Hukou	Dummy	0	1	0.72	0.45
Homeownership	Dummy	0	1	0.79	0.40
Years of residence	Continuous	0	50	7.30	5.42
Class	Dummy	0	1	0.42	0.49
CCP	Dummy	0	1	0.21	0.41
Individual social capital
Intra-community	Continuous	−0.33	10.36	0.004	1.00
Extra-community	Continuous	−1.39	4.07	0.12	1.00
Community social capital	Continuous	3	4	3.45	0.15
Neighbor acquaintances	Continuous	0	200	11.81	21.59
Community size	Continuous	107	10,000	1961	2830
Community homeownership rate	Continuous	51%	100%	0.79	12.92
Community migrant rate	Continuous	3%	68%	0.28	0.15
Community location	Categorical
Inner core	Dummy	0	1	0.33	0.47
Inner suburb	Dummy	0	1	0.51	0.50
Outer suburb	Dummy	0	1	0.16	0.37

reports the descriptive statistics of all independent variables used in the analysis. To synthesize individual-level and community-level variables into a single framework, a two-level analysis method was adopted. The first level included personal socio-demographic attributes, individual social capital inside and outside communities, community social capital, and neighbor acquaintances. The second level included the community’s demographic composition and the location of the community.

The measurement of individual social capital inside and outside communities is the main objective of this study. Following Lin [19] and Fu [45], the position-generator method is adopted in this study to calculate resources embedded in social networks. The method has been evidenced to offer a representative structure of the positions in a given society [24,45]. Social science researchers usually assign an occupational prestige score based on social prestige and social-economic role [45,46]. Participants were asked to report their acquaintances with individuals with jobs are from a list of 23 occupations (e.g., janitors, university lecturers, journalists, lawyers). Occupational prestige scores were assigned to occupations [45,46]. Based on respondents’ position generator answers, three indices for social networks could be calculated: (1) network extensity, or the volume of network resources, measured using the number of different occupations accessed by participants; (2) upper reachability, measured using the highest occupational prestige score of the occupations accessed by participants; and (3) the range of the occupational prestige scores for all accessed occupations. As shown in

Table 3. Means and correlation analyses of the three dimensions of ISC.

Dimensions of ISC	Network Upper Reachability	Network Extensity	Network Range
Network upper reachability	1
Network extensity	0.706 **	1
Network range	0.682 **	0.905 **	1
Intra-community	6.46	0.28	23.01
Extra-community	50.58	4.51	1006.21

Notes: ** significant at 0.01.

, the three indices are highly correlated (all over 0.682). The component score of the network resources extracted from a principal components analysis (volume + upper reachability + range) was employed to reflect the multidimensional nature of social networks. To elucidate the impacts of the social resources inside and outside communities, social capital based upon sources of social networks was calculated. This revealed that personal social resources were derived mainly from extra-community networks (see Table 2 and Table 3), which confirms the results of prior studies finding that social networks have transcended community barriers [13]. This means that in contemporary society, the traditional local social networks have been replaced by non-local networks.

Community social capital was calculated using four items on a 5-point Likert scale (from 1 = strongly disagree to 5 = strongly agree) to capture the degrees of trust, reciprocity and social cohesion within each community. Respondents were asked the extent they agreed with the four statements [47]: (1) ‘Residents in this community can be trusted’; (2) ‘Residents are helpful in the community’; (3) ‘This is a cohesive community’; and (4) ‘Residents can solve community problems together’. The variable of neighbor acquaintances is employed to reflect the quantity of community social ties. It was measured by asking respondents the number of neighbors in a community known by name. This ranged from 0 to 200, with an average of 11.81.

Socio-demographic variables were included to control for individual variations. These variables included age, marital status, the presence of children, household registration status, education, annual per capita household income, homeownership and length of residence. These variables are consistent with the social investment and social position variables used in prior civic engagement studies [31,43]. Among the 1774 participants, 55% were female, their average age was 45 years, 82% were married, 61% had children in their household, 72% were local residents, 79% were homeowners, their average years of schooling were 7, 42% considered themselves middle class or above and 21% were CCP members.

As for the community-level variables, community homeownership rates and migrant concentrations were defined as the share of homeowners (mean = 79%) and migrants (mean = 28%) in a community. The average size of communities among the 39 communities was 1961.

4. Results

Figure 3 gives the percentages of three forms of community participation. A total of 56.6% of the respondents participated in associational activities, and 27.6% participated in institutionalized actions, but only 11.1% of the respondents participated in contentious actions. Three observations can be derived from the figure: First, only associational participation took place in more than half of the population, which confirms prior studies in that our data were again skewed toward low levels of grassroots participation [9,10]. Second, the highest percentage of the population engaged in associational participation, indicating that people were more inclined to participate passively. Third, despite its low percentage, individual activism, including both institutionalized and non-institutionalized action, serves as a critical starting point for residents beginning to voice their attitudes on community affairs, using mild or even contentious ways to achieve their goals.

The results of the multilevel regression analysis are given in

Table 4. Regression analyses of community participation on individual and community social capital.

		Associational Involvement		Individual Activism
				Institutionalized Action		Non-Institutionalized Action
		Model 1		Model 2		Model 3
		B	S.E.	B	S.E.	B	S.E.
Fixed effects
Constant		−5.938	2.396	−0.555	2.333	1.419	2.658
Age		0.002	0.005	−0.019 ***	0.005	−0.019 **	0.008
Gender (1 = male)		0.143	0.114	0.045	0.119	0.224	0.170
Marital status (1 = married)		0.204	0.154	0.164	0.167	−0.327	0.224
Child in house (1 = yes)		−0.071	0.122	−0.085	0.128	0.122	0.183
Years of education		−0.001 *	0.021	0.018 *	0.023	−0.013	0.033
Hukou (1 = local)		0.513 ***	0.138	0.398 **	0.155	0.462 *	0.236
Homeownership		0.836 ***	0.161	0.672 ***	0.191	0.548 *	0.287
Years of residence		0.072 ***	0.013	0.042 ***	0.013	0.048 **	0.017
Class		−0.070	0.118	−0.081	0.123	0.091	0.177
CCP		0.261 *	0.149	0.007	0.150	0.073	0.209
Individual social capital	Intra-community	0.099 *	0.064	0.081 *	0.053	0.197 **	0.062
	Extra-community	0.283 ***	0.062	0.326 ***	0.061	0.513 ***	0.082
Community social capital		0.979 **	0.676	0.902 *	0.660	1.412	0.737
Neighbor acquaintances		0.011 ***	0.003	0.004 *	0.002	0.006*	0.003
Community size		−0.055	0.051	0.013	0.051	0.030	0.056
Community homeownership rate		0.008	0.011	0.021 *	0.011	0.012	0.012
Community migrant rate		0.006	0.009	0.006	0.009	−0.008	0.011
Inner core		0.395	0.233	0.187	0.233	0.102	0.254
Outer suburb		0.592 *	0.305	0.112	0.299	0.136	0.341
−2log likelihood		2032.92		1894.11		1065.78
BIC		2189.50		2050.72		1222.35

Notes: *** significant at 0.001; ** significant at 0.01; * significant at 0.05.

. Below, we present the fixed-effect results for both Level-1 variables, which include personal attributes and individual social capital inside and outside communities, as well as the results for the Level-2 or community variables.

Level 1 Variables: A. Control variables. The results show that age was insignificant in Model 1 (associational involvement), but this variable was highly significant and negative in both Model 2 (institutionalized action) and Model 3 (non-institutionalized action). Young people were more likely to express and protect their rights, while older people were more accepting of the current situation and continued to organize passively. Although the gender variable was insignificant in all three models, in each instance, the coefficient estimate was positive. This indicates that males were more likely to join activities across all three dimensions. Considering measures of socioeconomic status, educational attainment had a significant negative effect on associational participation. However, it showed a significantly positive effect on individual activism. This means that education can improve residents’ right consciousness and expressions of themselves. Homeownership, local hukou and length of residence had strong associations with all forms of participation, reflecting the importance of both economic commitment and rootedness. CCP status was only significant in Model 1, indicating that CCP membership raised the likelihood of associational engagement.

B. Individual social capital inside and outside the community, community social capital and neighbor acquaintances. Individual social capital outside communities was significant in all models. This suggests that even though a community’s status as the major source of social ties is not assured, its residents still care for their home grounds and community public services. Furthermore, the source of social networks outside communities is mainly employment, meaning that employment is not only a source of income for modern citizens but also a significant means for accumulating individual social ties. Community social capital was only significantly related to associational involvement and institutionalized action. This means that the mobilizing effect of community social capital facilitates residents to cooperate with the grassroots government on community issues and address community issues [48], but it is insufficient for engaging residents in contentious movements. Neighbor acquaintances were significant in both Model 1 (associational involvement) and Model 2 (institutionalized action), but this factor was insignificant in Model 3 (non-institutionalized action). To this end, non-institutionalized action is more likely to be driven by the social capital outside the community rather than other types of social capital.

Level 2 Variables: A. Community attributes. The homeowner share of the population was positively associated with institutionalized actions, which reaffirmed the importance of homeowners’ implication in individual activism. The migrant share of the population and community size were not significant. The coefficient estimates of the locations of the communities were in line with previous studies [49], wherein residents living in the outer suburbs were more likely to engage in associational activities. This is probably because suburban housing estates have congregated young professionals who are in search of environments conducive to bringing up a family. Therefore, these community members may be prepared to contribute to community affairs, such as getting involved in RCs and HOAs to improve estate management.

5. Discussion

At the beginning of the paper, we questioned the association of the individual social capital inside and outside communities with participatory behavior. Our findings revealed that individual social networks have dispersed beyond community boundaries, and extra-community social networks are the main sources of such networks for individuals. It further revealed that different dimensions of social capital had differential effects on participatory behaviors. Community social capital and social networks inside communities were significantly related only to associational involvement and institutionalized action, and social networks outside communities were associated with all forms of participation. Community-based connections are gradually becoming diffuse in the context of unprecedented urban spatial and social transformation. Today, the individual social capital of residents is mainly employment- and interest-based, whereas residents living in the same community are strangers to each other with relationships only as nodding acquaintances. Therefore, community social capital, which is dependent upon community-based connections, only exerts a limited mobilizing effect. The strength of the community’s social capital enables residents to engage in grassroots activities, but it is insufficient for pulling residents together to take the initiative in further developing their homes.

The impact of social capital (or guanxi in Chinese) is omnipresent in Chinese cities [16,39]. Empirical studies have verified that those residents involved in civic participation and collective action were more likely to achieve success if they had social networks with officials or media workers [16,40]. The findings of this study again verify the influence of social capital by revealing the role of resources from social networks accessed outside communities in community behaviors. It not only implies the influence of social capital but also suggests that the source, measurement and effects of the individual social capital can lead to a new understanding of community participation and development. The seminal work by Elias and Scotson [50] suggests that the marginal groups are less involved in community affairs due to stigmatization and exclusion. Following this line of theory, Wang et al. [41], using the case of migrants in China, reveal that intergroup neighboring help migrants to break down the barrier formed by stereotypes, thereby facilitating collaboration and involvement in neighborhood activities. According to our findings, it is probably because the intergroup neighboring increases social resources, thereby overcoming the barrier to joining in community activities. Thus, the unequal power configuration amongst different population groups largely contributes to residents’ participation behavior. Herein, we suggest future studies of community behaviors should look beyond traditional predictors of community participation, such as hukou [9,41], housing tenure [37], socioeconomic (income and occupation) [22] and intra-community social networks [7,12,15], and consider more about the heterogeneity in residents’ extra-community social capital.

Some limitations of this study deserve attention. First, we explored the significance of the association of extra-community social capital with community participation while considering other effects underlying the restructuring of Chinese social relationships. However, we were not able to draw a causal inference. Second, informal housing (e.g., urban villages and self-built housing) was not included in the present survey, which excluded a large number of rural migrants. Future studies are needed to further verify the generalization of our findings in other community contexts. Third, important information about social media use, which has significantly changed forms and channels of participation at the community level and hence relationships between social capital and grassroots participation, is missing due to the nature of the survey. To overcome those limitations, future studies should provide a broader picture of the relationship between social capital and civic engagement.

6. Conclusions

This paper presents the first empirical evidence about the impacts of social capital accessed outside a community on participation in activities within a neighborhood. Based on the survey data from the city of Guangzhou, China, we assessed how individual social capital and community social capital jointly and differently shaped participatory behaviors. Results from multilevel regression analysis reveal that social capital outside communities significantly contribute to all forms of community participation. Results also show residents’ social capital is mainly from outside the community. Findings confirm the importance of social resources in community engagement. It implies that the current social capital framework explains only part of the variations in community participation and suggests scholars should look beyond traditional predictors by considering the impacts of the unequal distribution of social resources from outside communities. To enhance participation in community activities and rebuild the community, policymakers and organization leaders in China are suggested to give special attention to marginal groups who have scant social resources.