COETHNICITY AND THE DILEMMAS OF COLLECTIVE ACTION

Eliciting participation in public goods provision embodies a standard collective action problem, which describes situations in which groups of interdependent actors—each acting rationally and in their own self-interest—fail to undertake a joint action that would benefit everyone. Because the costs are great and the benefits diffuse, scholars like Olson (Reference Olson1965) have argued that individuals will be motivated to participate in a joint action only when inducements or coercion are used. However, the frequency with which people do, in fact, cooperate, in the social world cries out for an explanation. Many political scientists have used rationalist and institutional frameworks to explore how collective action dilemmas can be resolved. For example, new institutionalist scholars argue that institutions provide a setting for repeated interaction and, therefore, serve as mechanisms for resolving social dilemmas (Ostrom Reference Ostrom1991; Shepsle and Weingast Reference Shepsle and Weingast1981). Other scholars, however, have cast doubt on these efficient explanations of social order, and have instead explored the social basis of collective action. These scholars highlight the “vehicles of culture”—norms, values, and identities—that provide the “cement of society” and facilitate collective action (Elster Reference Elster1989, 248).

Ethnic identity is one such “vehicle of culture” that may play a role in resolving collective action dilemmas. Individuals tend to use identity points to organize their social environment because they provide information that helps one define his or her social role in relation to others (Hale Reference Hale2004). Ethnicity, like class or gender, is a type of identity, in which, “membership is determined by attributes associated with, or believed to be associated with, descent” (Chandra Reference Chandra2006, 398). An individual is invested with an ethnic identity if she is characterized by certain ascriptive markers acquired genetically or through “cultural and historical inheritance.”

A number of studies have illustrated how shared ethnic identity can facilitate mutually beneficial collective action, advancing different mechanisms by which coethnicity promotes cooperative behavior. Some scholarship highlights the prevalence of in-group favoritism in social environments, suggesting that individuals tend to care more about the welfare of their ethnic group members (Horowitz Reference Horowitz1985; Tajfel et al. Reference Tajfel, Billig, Bundy and Flament1971). Homogenous communities, then, will have an advantage over heterogeneous ones in local goods provision because community members will be more apt to bear the costs of providing goods that benefit their “kind” (Butovskaya et al. Reference Butovskaya, Salter, Diakonov and Smirnov2000). Other scholars have advanced a “technology” hypothesis, arguing that co-ethnics have an advantage in accomplishing collective tasks because they are better able to communicate with one another and, thus, are able to function more efficiently together (Habyarimana et al. Reference Habyarimana, Humphreys, Posner and Weinstein2009). Essentially, members of the same ethnic group can take advantage of common cultural materials (e.g. language) that facilitate coordination. Moreover, an individual may be better able to “read” a co-ethnics behavioral cues, improving the likelihood that he/she will engage in joint action with them (Bacharach and Gambetta Reference Bacharach, Gambetta and Cook2001; Habyarimana et al. Reference Habyarimana, Humphreys, Posner and Weinstein2009). Regardless of the mechanisms involved, ethnic homogeneity is thought to facilitate local goods provision by lowering transaction costs that serve as barriers to collective action.

SECURITY PROVISION IN CAPE TOWN

Yet when I compared experiences with local security initiatives in communities across Cape Town, I found that the relationship between ethnic homogeneity and successful public goods provision is far from reflexive. I chose to investigate security initiatives as an indicator of public goods provision because public safety is a salient issue in Cape Town. South Africa, as a whole, is plagued by exceedingly high rates of violent crime. Crime in Cape Town, however, is particularly egregious, as the city has continuously experienced the highest prevalence of murder and drug-related offenses in the country.Footnote ¹ Service delivery failures have been particularly apparent in this domain, as the municipality has struggled to provide adequate security for their residents in the face of endemic violence.

Since my empirical aim is to investigate the local provision of public goods, I chose to examine community policing efforts, in particular. In the post-Apartheid era the South African government has made a concerted effort to reform the institutions of law enforcement and develop a community-based approach to security provision. After the fall of Apartheid, the new regime saw radical police reform as an essential part of their commitment to a non-racial democracy. In particular, the government desired to rebuild trust between local communities and the South African Police Services (SAPS), since SAPS functioned as the repressive arm of the Apartheid government, carrying out much of the violence against non-white persons during this era. As such, a new community-based model of law enforcement was explicated in the Interim Constitution, which included the establishment of neighborhood watches as well as municipal and provincial-level consultative bodies known as Community Policing Forums (CPFs). In addition to the goal of improving community-police relations, CPFs were also designed to enhance SAPS’ crime-fighting capacity by improving communication between law enforcement, government, and local neighborhoods. In practice, there is considerable variation in how well these bodies perform across districts. CPFs tend to be more effective in Cape Town's wealthier communities—where resources are more readily available for community policing—and in those areas where local businesses are actively involved in the process (Pelser Reference Pelser1999). Scholars have documented some of the challenges that CPFs have faced across South Africa. For one, crime levels have risen in South African townships, which impact the perceived effectiveness of CPFs and, therefore, support for the model (Brogden Reference Brogden2002). Moreover, CPFs have floundered in some communities because of the weak institutional capacity of SAPS, a partner (and manager) in community policing. Many officers have not received adequate training in this model of policing, thus compromising its effectiveness.

In determining field cites in which to investigate local CPFs, I chose two neighborhoods with similar demographic characteristics. Delft South is a residential district in the suburb of Delft, a poor community that occupies land adjacent to Cape Town International Airport on the marshy terrain known as the Cape Flats. Hangberg is a residential district in the suburb of Hout Bay, a seaside community on the Atlantic Seaboard, 20 miles south of Cape Town's Central Business District. Both communities are mostly homogenous in their racial makeup. Hangberg is mostly populated by Coloured families,Footnote ² while Delft South is populated largely by Black African families with some Coloured residents living in the neighborhood. Both communities form part of the larger, more ethnically heterogeneous suburbs of Delft and Hout Bay, respectively.

While both of these communities are similarly homogenous, they have had different experiences with respect to local public goods provision. While Delft South has an active CPF, Hangberg's CPF has struggled to garner participation from the community in safety initiatives. After conducting field research in these communities, the contrasting environments of social capital stood out as the most proximate explanation as to why Delft South and Hangberg experienced divergent outcomes of local security provision. The community characterized by an environment of bonding social capital, based on coethnic ties, was less able to work together to improve public safety, while the neighborhood that lacked these strong bonds saw greater levels of participation in such efforts. To many, the suggestion that coethnic social capital can undermine public goods provision may sound counterintuitive, since it is assumed to facilitate mutually beneficial collective action. But in certain contexts, coethnic ties can work against local goods provision; in the next section I lay out a theoretical framework that demonstrates under what conditions this may happen.

THE ROLE OF SOCIAL CAPITAL

Social capital has been described as facets of social environments—such as norms, networks, and relationships of trust—that facilitate mutually beneficial cooperation (Coleman Reference Coleman1990; Putnam Reference Putnam2000). If we break this concept into its constituent parts, we see that shared ethnic identity fosters the development of each. For instance, social networks are often established on the basis of shared ethnic membership. Scholars have found that, in many different settings, coethnics tend to be tied together in dense systems of interpersonal relationships, interacting more frequently with in-group than out-group members. For instance, in an experimental game designed to test how “findable” coethnics were, Habyarimana et al. (Reference Habyarimana, Humphreys, Posner and Weinstein2009) found that, game participants were more successful in tracking down coethnic “targets” than non-coethnics persons because of the prevalence of identity-based networks. Moreover, shared ethnic identities often give rise to shared norms. Norms can be described as informal and commonly held cultural understandings that govern individuals’ behavior (Chwe Reference Chwe2001). Because they are “community-based”, norms can develop in smaller social units, like ethnic groups. Many scholars have documented the existence of ethnic norms, or what Kuran (Reference Kuran1998) defines as “…the behavioral codes that its members are expected to follow to retain social acceptance.” Ethnic norms not only shape members’ behavior, they also serve as a “badge marking membership” (Whitt and Wilson Reference Whitt and Wilson2007, 656) and, therefore, can reinforce individual identification with the group and increase cohesiveness among coethnics (Bhavnani and Miodownik Reference Bhavnani and Miodownik2009).

Finally, shared identities foster relationships of trust between coethnics. Ethnicity serves as a type of “social radar” (Hale Reference Hale2004), the empirical markers of which provide individuals with low-cost information about others that can be used to generate expectations about their behavior (Chandra Reference Chandra2004). These ethnic cues act as a kind of shortcut, helping individuals bypass the need to collect personal information about others’ intentions or competence. Shared ethnic membership, therefore, has been shown to reduce uncertainty in social interactions (Foddy and Yamagishi Reference Foddy, Yamagishi and Cook2009). As such, it serves as a potent source of trust; individuals are often more willing to place their trust in a coethnic, having extrapolated about their intentions on the basis of group membership. A vast array of studies—both experimental and survey-based—have corroborated the link between coethnicity and trust (Barr Reference Barr2003; Kasara Reference Kasara2011). As we can see, norms, networks, and trust comprise the reservoir of social capital that coethnics enjoy.

This reservoir, in turn, helps to facilitate mutually beneficial collective action among coethnics. Because it encompasses a belief that others will refrain from acting exploitatively and engage in reciprocity, trust between coethnics reduces the transaction costs that impede cooperative behavior. Meanwhile, coethnic norms facilitate mutually beneficial collective action because they encourage individuals to engage in cooperative behavior with in-group members, as several experimental studies have documented (Barr Reference Barr2003; Fershtman and Gneezy Reference Fershtman and Gneezy2001). For instance, Habyarimana et al. (Reference Habyarimana, Humphreys, Posner and Weinstein2009) conducted a series of experimental games using randomly chosen subjects from heterogeneous neighborhoods in Kampala, Uganda. They found that a majority of subjects believed that cooperation with coethnics would be reciprocated, and this was reflected in their behavior. In other words, they were able to identify a specific coethnic norm at work—one that encouraged reciprocity for coethnics. Finally, coethnic networks facilitate collective action because they provide a mechanism for sanctioning group members that fail to contribute to a joint effort. Part of the effectiveness of norms comes from the threat of sanctioning for those who violate them, and sanctions are applied more effectively within ethnic groups than between them because coethnics are often tied together in dense social networks, making it easier to ostracize a violating member (Habyarimana et al. Reference Habyarimana, Humphreys, Posner and Weinstein2009).

The empirical link between coethnicity, social capital, and collective action helps us understand why many homogenous communities are able to successfully provide local public goods, like community-based security. But if coethnicity facilitates the development of social capital and coethnic social capital enables mutually beneficial collective action, under what conditions do these mechanisms fail?

STRONG BONDS AND LOCAL SECURITY PROVISION

I argue that, in some contexts, these same components of coethnic social capital can lead to collective action failures. This happens when coethnicity creates an environment where norms, networks, and relationships of trust are insular, exclusive and constricting. In other words, coethnicity will fail to resolve social dilemmas when it generates bonding social capital. Robert Putnam describes bonding social capital as that which, “…brings together people who are like one another in important ways (ethnicity, age, gender, class, and so on)…” (Putnam Reference Putnam2002, 11). It is the kind of social capital that creates strong bonds between individuals and tends to “reinforce exclusive identities and homogenous groups” (Putnam Reference Putnam2000, 22). In his seminal work on network theory, Granovetter describes these bonds as “strong ties”, which tend to form dense networks in communities that comprise them (Granovetter Reference Granovetter1973). Coethnicity facilitates the development of strong ties.

We can conceptualize the effects of bonding social capital in ethnically homogenous communities by returning again to the issue of local security provision. Providing collective security requires individuals to participate in policing their own neighborhoods, which often involves conducting volunteer street patrols or assisting the police in thwarting illegal activities in the area by calling their attention to suspicious activities or individuals. Sociologists have referred to such action as “social control”—a concept that describes the “…capacity of a social unit to regulate itself according to desired principles—to realize collective, as opposed to forced, goals” (Sampson Reference Sampson, Saegart, Thompson and Warren2001, 94). Many scholars argue that a community's stock of social capital facilitates social control, which, in turn, improves public safety. Where there is a density of social ties—that includes intergenerational links—communities are better able to regulate the behavior of individuals (Sampson, Morenoff and Earls Reference Sampson, Morenoff and Earls1999); social control becomes difficult, however, when neighborhood ties are few or shallow or when a community suffers from what Janowitz (Reference Janowitz1975) has termed “social disorganization.” In this sense, scholars argue, bonding social capital helps communities act collectively to confront local problems (Larsen et al. Reference Larsen, Harlan, Bolin, Hackett, Hope, Kirby, Nelson, Rex and Wolf2004).

But strong ties may also be harmful to local security provision in certain contexts. In neighborhoods plagued by violence and drug use, participating in community policing may involve being called upon to identify coethnics who are potentially involved in the drug trade and/or violent activities. In this way, local goods provision improves the welfare of the community but comes at the expense of a small segment of the community involved in illegal activities. I argue that, in communities characterized by high levels of bonding social capital among coethnics, individuals will be hesitant to contribute to local policing efforts because such activities may target certain segments of the community and, ultimately, result in harm to their own social environment. In high bonding communities, the presence of dense networks means that an individual will likely maintain ties with a significant portion of community members. Granovetter has described such a community—one that is “…completely partitioned into cliques, such that each person is tied to every other person in his clique but no others outside of it” (Granovetter Reference Granovetter1973, 1373). This type of environment, he argues, both shapes, and ultimately, constrains one's behavior. With respect to law enforcement, individuals will recognize that participating in policing efforts that may implicate a fellow community member could result in one's social ostracization.

Norms play a central role in this dilemma. In regard to policing, “norms of silence” often develop in communities with strong ethnic bonds, restricting opportunities for individual action. That is, community members tend to behave according to the expectation that they refrain from “ratting out” their coethnic neighbor. An individual may believe that if she refuses to behave according to this “norm of silence”, those in her social network will begin to distance themselves from her because of this violation. Therefore, she will forego participating in neighborhood watches, joining in street patrols or attending public safety meetings, so as not to appear a “rat” and inflict irreversible damage to her social environment. In this way, norms function as social control, prescribing and proscribing certain types of behavior (Cook Reference Cook2005) and placing “restrictions on individuals’ freedoms” (Portes Reference Portes2000, 532).

As Cook (Reference Cook2005) has explained, this kind of social control can only develop in a relatively closed community with high levels of bonding social capital, where social networks make ostracization a tangible threat. Where networks are tightly-knit, sanctions are more easily carried out as information about norm violation is easily obtained; individuals will, therefore, behave accordingly. Moreover, in a community with high levels of bonding social capital, networks also tend to be insular. That is, community members maintain strong ties with coethnics in their community, but maintain fewer “bridging” ties with non-coethnics or others outside the community. Because of this network structure, the social consequences of ostracization from the ethnic group are grave. In the context of local security provision, the threat of becoming a social outcast dampens the willingness of community members to contribute to community policing efforts.

On the other hand, we can expect individuals to be less hesitant about contributing to local policing in communities without high levels of bonding social capital. Without strong ethnic bonds, “norms of silence” may never develop or may not become embedded in the cultural milieu, so the prospect of “ratting out” a coethnic involved in illegal activities would not be as problematic. Moreover, in communities that lack bonding social capital, networks are less dense, less tightly-knit and less insular; therefore, networks will be less effective mechanisms for sanctioning. And when an individual maintains “bridging” or “weak” ties with persons outside of the ethnic group, the consequences of social ostracization are less dire (Granovetter Reference Granovetter1973; Uslaner Reference Uslaner2012). In short, in communities without high levels of bonding social capital and where bridging ties to outside social groups are present, individuals may not expect to be ostracized for norm violation, and if they are sanctioned by their immediate network for speaking out, they may still be able to maintain a supportive social system with outside groups. I expect that, in such environments, one will be more willing to contribute to community policing. In the next section, I present a comparative case study of the communities of Hangberg and Delft South, which tests this theoretical framework.

THE EVIDENCE

To establish causality, I used a comparative analytical method akin to a “most similar systems analysis” design (Sartori Reference Sartori1970), in which two cases with comparable characteristics and contrasting outcomes are analyzed and explained. To collect data on community policing, I visited municipal police stations and community improvement district headquarters in the sample communities I chose, conducting semi-structured interviews (that were recorded and later transcribed) with local police officers, chairpersons of CPFs, and members of neighborhoods watches. Initial contacts with CPF officials were made through my position as a visiting researcher at the University of Cape Town, after which a “snowballing” method was used to establish further contacts in each neighborhood. The interviews that were conducted focused on rates of participation among residents in community policing efforts, the functions and objectives of these organizations, and the challenges and obstacles that these bodies face. I also collected primary documents, including CPF newsletters and meeting minutes in my sample communities.

To gather data on social capital I conducted semi-structured interviews with members of community groups, directors of community improvement districts, religious leaders, and ordinary citizens in both Delft South and Hangberg.Footnote ³ I was able to gauge community characteristics from these interviews, including perceptions of community cohesion and levels of inter-personal trust between neighbors. In addition, I used a number of primary sources, including the written minutes of neighborhood watch meetings, electronic correspondence among neighborhood members (which I was granted access to) as well as secondary scholarly accounts of the areas. Ultimately, I used a process-tracing technique, which allowed me to work backward from the two outcomes in question to analyze which factors led to their divergence.

Table 1 presents a simplified picture of the similarities and differences between Hangberg and Delft South. As we can see, in addition to their ethnic homogeneity, both Hangberg and Delft South share other similarities. Both communities are economically marginalized and face the dual challenges of persistent housing shortages and unemployment. A significant number of families in both areas reside in informal housing, living in poorly constructed tin shacks or in the backyards of relatives. Poverty rates in both communities are high, even as per capita income is higher in Hangberg than in Delft South. Because Coloureds were placed higher on the racial hierarchy during the Apartheid regime and enjoyed the benefit of a Cape Town Coloured preference labor area denied to black Africans, the people of Hangberg tend to be more economically secure than Delft South residents, having worked for decades in a thriving commercial fishing industry. However, more recently both communities have been plagued by high rates of unemployment. Many workers in Hangberg have lost their jobs with the decline of commercial fishing in Hout Bay and have begun participating in the risky and illegal business of poaching crayfish and abalone. Similarly, Delft South residents have struggled to find work, since the suburb is located on the economically depressed Cape Flats and transportation links to more economically vibrant parts of the city are poor. Many service-industry and manufacturing jobs are located 34 kilometers to the west in the central business district or in the Northern and Southern suburbs.

Table 1. Comparing Delft South and Hangberg

Both communities also struggle with the persistent problems that often accompany high rates of unemployment, namely alcohol abuse, drug abuse and crime. According to a municipal report from 2009, Delft is among the top five police precincts in Cape Town with the highest rates of murder and sexual offenses.Footnote ⁴ Since 2009, crime rates have remained steady in the area, with intermittent surges in violent activity. In Hangberg, rates of robbery, theft, and child abuse are high. In both communities, criminal activity is connected to drug trafficking. In Hangberg, for instance, organized crime syndicates run the illicit fishing trade, paying their poachers in drugs instead of money.Footnote ⁵ This has exacerbated drug addiction among poachers, and consequently, has fueled a rise in property crime as users turn to theft to fund their drug consumption.

Yet, while Delft South and Hangberg share a number of structural similarities, these communities have experienced different outcomes with respect to the local provision of security. Delft South residents are active in community policing efforts, participating in many of the official activities arranged by Delft's CPF and organizing several of their own neighborhood watches throughout the sector.Footnote ⁶ Members of the neighborhood watch frequently engage in day and night patrolling, especially on the weekends and, in addition to autonomous patrolling, watch members often accompany the South African Police Service on search and seizure operations. Delft South residents also participate in other community initiatives organized by the CPF, including youth days, anti-truancy programs, or “Walk-abouts” (where community members confront suspected gang members or drug dealers and persuade them to cease their activities). They participate in monthly “Imbizos”, in which the community meets to discuss public safety issues and in more ad-hoc public meetings that often follow a particular incidence of violent crime.Footnote ⁷ Finally, Delft South residents have participated in more recent organized marches against taxi violence, violence against children and the killing of police officers.Footnote ⁸

According to SAPS officials, Delft South is one of the most participatory sectors in the suburb.Footnote ⁹ Police officials note that the Imbizos—open to all residents of greater Delft—are well attended by the black African community living in Delft South. And in sector-specific meetings of the Community Policing Forum, Delft South residents attend in large numbers, regardless of whether the issues being discussed are relevant to their own households. A SAPS official recalled that over 300 residents attended the Imbizo the month before; in less participatory sectors, he noted, CPF organizers “struggle to get 50 people together.”Footnote ¹⁰ He noted that even residents without school-aged children regularly attend Imbizos focusing on school truancy. Interestingly, some community members who are not officially registered as members of the neighborhood watch still contribute to providing security. For instance, men in Delft South sometimes join together to confront shebeen owners that continue to operate after hours.Footnote ¹¹ Local sheebeens—that is, illicit bars that serve alcoholic beverages without a proper license—are often the focal point of crime in early morning hours. By regulating the behavior of shebeen owners, Delft South residents are contributing to security provision in the neighborhood.

The community of Hangberg has not enjoyed the same level of success in community policing as Delft South. While there are local institutions in Hangberg that support collective security, such institutions receive little support from the community at large. Hangberg is included as a sector in the Hout Bay Neighborhood Watch (HBNW), a large community-based program that organizes security initiatives throughout the suburb. But while Hangberg maintains a working relationship with the HBNW, the organization has had difficulty in garnering participation among Hangberg residents in HBNW activities.Footnote ¹² Only a small number of residents have registered as members, and only a fraction of those members participate in day and night patrolling. In contrast to the community of Delft South, few Hangberg residents attend public safety meetings hosted by the Community Policing Forum. Moreover, few residents have expressed an interest in strengthening collective security efforts in the community, despite the persistent problems with crime.Footnote ¹³ The Hout Bay branch of SAPS has also struggled to generate a police partnership with Hangberg, as many residents view SAPS with suspicion. The suspicion stems from allegations of corruption within SAPS and a general feeling that law enforcement has neglected Hangberg, failing to respond to reports of crime with the same vigor as they have in other Hout Bay neighborhoods.Footnote ¹⁴ While other areas of Hout Bay have strengthened their partnerships with SAPS in order to confront rising crime, Hangberg has turned away from the institution at a time when their problems with drug abuse and organized crime connected to illicit poaching have increased.

CONTRASTING ENVIRONMENTS OF SOCIAL CAPITAL

What explains why these communities differ so greatly in their efforts to police their own neighborhoods? And why has Hangberg—a homogenous community of long-standing residents—struggled to achieve effective community policing when it has all the ingredients that are presumed to enable successful collective action? One of the most salient explanations of the variation is the contrasting environments of social capital between these two communities.

The seaside village of Hangberg is a tightly-knit coethnic community that exemplifies a “high bonding” environment. Residents here maintain close relationships with one another and many locals are part of larger families whose members also reside in Hangberg. The cohesiveness of the community is largely an artifact of apartheid-era segregationist legislation, which mandated that Coloured families be forcibly removed from Hout Bay's valley area to this seaside neighborhood. In the years that have followed the end of Apartheid, Hout Bay's racially demarcated settlement patterns have remained largely intact. And many of Hangberg's “original” families—those subject to forced removals from the valley—still reside in this community. While Hangberg's residents are closely tied to one another other, residents maintain few ties with communities in greater Hout Bay, such as the majority black township of Imizamo Yethu or the majority-white neighborhoods in the valley.Footnote ¹⁵

Hangberg also maintains a strong common identity, one that is intimately intertwined with their place of residence and their ethnicity as Coloured South Africans. It is also an identity that differentiates Hangberg from other communities in Hout Bay. As Tefre (Reference Tefre2010, 182) notes, “There is a clear understanding in Hangberg of who ‘we’ as a community, are, that demarcates them from their neighboring communities, and this identity as ‘coloureds’ and the original inhabitants of Hout Bay seems to be shared by all in Hangberg.” Hangberg's Coloured identity has grown stronger in recent years, as residents perceive that the area has been neglected by city officials that expend more resources in confronting problems in the black community of Imizamo Yethu.Footnote ¹⁶ It seems that the racial dynamics of inter-community relations in greater Hout Bay has exacerbated in-group/out-group perceptions in Hangberg, another indicator of the community's high bonding social capital.

But how has bonding social capital contributed to the failure of collective security provision in Hangberg? In short, it appears to have dampened the willingness of individuals to contribute to community policing initiatives. In Hangberg, a “norm of silence” has developed within the community.Footnote ¹⁷ The norm implies that, when a crime is committed, community members should refrain from “ratting out” those involved in law enforcement. Since the community is tightly-knit, informing the police about a crime and its alleged perpetrator often means one would be betraying the brother of one's neighbor or the local shopkeeper's niece. Residents understand that the consequences of such an action would be social ostracization, which would be costly. Maintaining good relationships with their coethnic neighbors is essential to Hangberg residents since these networks function as a crucial social support in this economically depressed area.

Moreover, many in Hangberg lack social ties with individuals outside of these coethnic networks.Footnote ¹⁸ In this sense, the community can be characterized as lacking bridging social capital, which helps to explain why community policing has failed here. Without ties to individuals in other communities, the consequences of ostracization for a Hangberg resident would be harmful to their material and social well-being. Individuals here are acutely aware of being perceived as “pimpers”, or police informants, and fear the repercussions of being labeled as such. For these reasons, most residents avoid being seen at community policing forum meetings and refuse to participate in day or night patrolling.Footnote ¹⁹

The dearth of bridging social capital in this community may also help explain Hangberg's failure to engage in an effective partnership with SAPS. Hawdon (Reference Hawdon2008) argues that a community's level of social capital impacts residents’ perception of police trustworthiness, which in turn, influences the likelihood of successful cooperation with the police. Police legitimacy is often lacking in communities with dense bonding ties and low levels of bridging social capital, he explains, since “…officers, as outsiders, are not likely to be trusted, further deteriorating the lack of legitimacy with which they enter a situation” (Hawdon Reference Hawdon2008, 194). Hangberg exemplifies this type of “high-bonding, low-bridging” community. While allegations of police misconduct have certainly depleted levels of trust between residents and the police, Hangberg's insular social environment also contributes to the perception of police illegitimacy among residents, as the police are viewed as “outsiders” who have historically neglected the Coloured community here.Footnote ²⁰

Finally, Hangberg residents shy away from community policing out of concern for the greater Hangberg community. There is a perception that participating in CPF activities would generate feuds between families and that such hostilities could threaten to tear the very fabric of the village.Footnote ²¹ Because Hangberg maintains a strong identity as a Coloured community, its residents express a desire to maintain the community's cohesiveness. To do so, they refrain from participating in any activities, like joining a neighborhood watch that could undermine this goal. The desire to maintain community cohesion has also generated another social norm in Hangberg, namely the private settlement of disputes (Tefre Reference Tefre, Peter and Kyed2014). Involving police authorities in conflicts—such as theft, assault, or domestic violence—may result in criminal proceedings for the alleged perpetrator. Such a situation has negative consequences for a victim, her family, and the family of the alleged perpetrator. But it may also strain the relationship between all of the persons and families involved, and by extension, the greater community. As such, many families choose to arbitrate matters among themselves through mediation or, depending on the severity of the wrong, with the use of violence as a form of sanction. Such mechanisms of dispute settlement conform to what Donald Black (Reference Black2014) terms “self-help”, a form of social control in which members of communities use violence to punish and deter deviant behavior. Self-help is particularly prevalent, Black argues, in communities, like Hangberg, that are “neglected by law” (Black Reference Black2014, 40), or where residents perceive there to be police neglect. In the context of community policing, Hangberg's self-help norms have undermined cooperation with SAPS and participation in the local CPF.

In sum, in the context of this community, it appears that coethnicity has lead to collective action failures. Shared ethnicity has facilitated the development of bonding social capital in Hangberg, which manifests in dense and insular social networks, relationships of “thick” trust (Putnam Reference Putnam2000, 136), and robust social norms. Evidence suggests that these norms and networks have shaped the behavior of Hangberg's residents in regard to the public good and have decreased their willingness to contribute to it.

Delft South's experience with local goods provision stands in contrast to Hangberg's. Because Delft South lacks the kind of bonding social capital that characterizes Hangberg, it has been able to overcome collective action problems related to community policing. In interviewing residents and reading scholarly accounts of the area, I found that social networks are more heterogeneous in Delft South than they are in Hangberg, as residents maintain weak ties with other races and ethnicities. While Delft South comprised a majority of black Africans, some coloured Capetonians live in the community; for the most part, relations between black Africans and Coloureds are cordial. Some families even rely on each other for assistance, sharing food or needed household items.Footnote ²² In researching social behavior in Delft South, Oldfield (Reference Oldfield2004, 196) found that “informal connections between neighbours of different races link families together, despite their different languages and places of origin.” She notes that Coloured and African families often look after each other's homes when they travel away from Cape Town, share cleaning responsibilities on their street and help with one another's gardens. In addition, Muyeba and Seekings (Reference Muyeba and Seekings2011) found that many residents in Delft often rely on members of different races for childcare (a notably high trust behavior).

Social bonds among coethnics in Delft South also appear less taut than in Hangberg, since many residents maintain ties with individuals outside of their immediate social and spatial circle. The type of social environment in Delft South is largely a function of its history. Delft South is a creation of the post-Apartheid era of government-led reconstruction, formed during the mid-1990s as a part of the African National Congress’ Reconstruction and Development Programme (RDP). The RDP was a policy framework meant to redress Apartheid-era socio-economic injustices like the acute shortage of adequate housing for Black and Coloured citizens. Situated just east of the airport, Delft South provided an adequate space on which to build a government-funded housing project. Many of Delft South's residents reside in small, state-subsidized housing units, having relocated to the area from other suburbs in greater Cape Town when they were allocated housing in the municipal lottery. Because this community is relatively young, the social environment in Delft South is less cohesive than in many surrounding communities. Oldfield (Reference Oldfield2004, 192) explains that many residents “… continue to build their lives around social and political networks that link to their previous homes in former African and coloured Group Areas.” As a result, the social networks that comprise Delft South are far from insular, as they connect residents to nodes outside of the local community.

It is important to note that Delft South residents indeed share ethnic bonds. A majority of residents are Xhosa-speaking; their families originally migrated to Cape Town from the Eastern Cape in search of work opportunities before and after the fall of the Apartheid regime. Shared ethnicity has helped Xhosa-speaking individuals in Delft South bear the hardships of life on the Cape Flats; for many residents, coethnics have provided financial or material support in times of need.Footnote ²³ But even though the shared Xhosa culture has brought people together in Delft South, the community still lacks a history of co-residence. As such, the social bonds between neighbors are less strong.

In sum, coethnicity has assuredly generated social capital in Delft South. But it has not generated bonding social capital—that which supports exclusive ties that reinforce social homogeneity (Cheong et al. Reference Cheong, Edwards, Goulbourne and Solomos2007). Since Delft South lacks an environment of bonding social capital, strong coethnic norms have not developed. The “norms of silence” that operate in Hangberg—which prevent residents from participating in community policing lest they are perceived as “rats”—do not exist in Delft South. Moreover, even if such norms existed, it would be difficult to sanction a violating member of the community in Delft South. Because social networks are heterogeneous and diffuse, there is not a substantial threat of social ostracization. Without restrictive norms or credible threats of sanction, there are simply fewer hurdles that prevent community members from contributing to community policing efforts in Delft South.

Several theoretical implications emerge from these cases. For one, Hangberg and Delft South's experiences with community policing provide a compelling illustration of the dynamics of informal social control in ethnically homogenous communities, a concept that has been theorized in sociology literature. As scholars have elucidated, communities self-regulate in different ways: informal social control can take place in the private sphere, which includes networks of kin, family and friendship groups, the parochial sphere, which include networks of community groups and the public sphere, or networks that include links to formal institutions and organizations external to the community (Hunter Reference Hunter, Suttles and Zald1985). Carr (Reference Carr2003) argues that social control is most effective when there is a strong “interplay” between the parochial and public levels of control. In the insular community of Hangberg, linkages to both formal institutions of law enforcement and other outside groups are weak. In this respect, parochial and public networks fail to act in concert with one another to produce effective social control. Moreover, Hangberg's prevailing social norms—which privilege silence over cooperation with the police because of fears of social ostracization—suggest that private social networks, built on coethnic ties, disrupt the efficacy of both the parochial and public sphere of social control. In Delft, however, community organizations and law enforcement are able to work together because they are not constrained by the dense and insular private networks that coethnicity sometimes generates. This corroborates a similar finding in Carr's research on one Beltway neighborhood in Chicago where the “dearth of dense social ties” contributed to “the diminished role of private and traditionally parochial forms of control” (Carr Reference Carr2003, 1285) and elevated the role of outside formal institutions, such as law enforcement and city bureaucracies. Linkages to these institutions helped to improve public safety in the neighborhood.

Furthermore, the case of Hangberg, in particular, provides an example of the “dark side” (Ostrom Reference Ostrom, Dasgupta and Serageldin2000, 176) of social capital (Levi Reference Levi1996; Portes Reference Portes2000; Portes and Sensenbrenner Reference Portes and Sensenbrenner1993). Social capital is often discussed in normative terms and assumed to be an ingredient that helps communities thrive. But while communal norms and social trust can generate positive externalities for neighborhoods, they can also, “…trap people within harmful social arrangements” (Pretty and Ward Reference Pretty and Ward2001, 213). Pretty and Ward (2001, 213) explain, “…[Norms] may encourage conformity, perpetuate adversity and inequity, and allow certain individuals to get others to act in ways that only suit themselves.” And as Sampson, Morenoff and Earls (Reference Sampson, Morenoff and Earls1999) found in their research on informal social control in Chicago, dense networks, and strong social ties can constrain and inhibit mutually beneficial collective action in some contexts as much as it can enable it in others. In many respects, the bonding social capital that encapsulates Hangberg has done exactly this. It has prevented a community from confronting a problem that residents wish did not exist.

Finally, while social capital appears to be a central factor in explaining the contrasting experiences with local security provision in Hangberg and Delft South, it is important to note that other variables may help explain these divergent outcomes. For one, mistrust between Hangberg residents and SAPS has indeed contributed to lackluster participation in the neighborhood CPF. Similar findings have been noted in sociological studies across the United States, such as Tyler and Fagan (Reference Tyler and Fagan2008) who find that the perception of police legitimacy significantly influences both whether people report crimes to local police and whether they participate in neighborhood watch groups. Police legitimacy, they explain, is largely a function of evaluations of procedural fairness as well as performance, and it is in these two areas in which SAPS falls short in the eyes of many Hangberg residents. The residents of Delft South, however, do not view SAPS officials with the same level of suspicion and this may contribute to the higher levels of participation in community policing there. In addition, the presence of organized crime in Hangberg may be an important variable in explaining failures in local security provision there. Residents not only fear the social ostracization that may come from implicating a neighbor, they also fear violent reprisals from gang members. These fears tend to suppress participation in community policing in the area.Footnote ²⁴ While such alternative explanations are necessary to paint a complete picture of community policing in Hangberg and Delft, they are incomplete. If for example, mistrust between the police and the community in Hangberg explains collective action failures there, it is essential to consider the role that social capital plays in this dynamic, since Hawdon (Reference Hawdon2008) found an empirical link between social capital and perceptions of police legitimacy. In general, one cannot fully understand the dynamics of collective action in a community without analyzing the social environment that encompasses it. And while many scholars recognize the importance of social capital to community policing (Larsen et al. Reference Larsen, Harlan, Bolin, Hackett, Hope, Kirby, Nelson, Rex and Wolf2004; Pino Reference Pino2001; Sampson Reference Sampson, Saegart, Thompson and Warren2001), it is valuable to analyze this relationship in contexts not yet explored and to parse the role that ethnicity plays in the process.