HISTORICAL PRECEDENTS FOR COMMUNITY POLICING

The introduction of community policing in Tanzania diverges from accounts of the concept as ‘a Western solution to an African dilemma’ (Brogden Reference Brogden2004: 635). In police and popular imagination, contemporary ulinzi shirikishi is not entirely a new innovation, and certainly not an example of the imposition of a western-designed model. The ongoing reform of the TPF of which community policing is a component has had very little external financial support, and was informed by extensive internal consultation. Furthermore, the way in which participatory policing is organised today and the assumption that citizens should play a role in securing their own neighbourhoods have a historical precursor in the form of ruling party sponsorship of ‘vigilantism’ during the 1980s and early 1990s.

Sungusungu organisations emerged in the early 1980s in the midst of a countrywide surge in violent crime. Drought, an international oil crisis and an economic downturn contributed to severe shortages, which coincided with the return of troops from Tanzania's war with Uganda and an associated influx of weapons. In towns commercial and residential properties were subject to armed raids, and in rural areas cattle theft was rampant (Heald Reference Heald2002: 1). Police attempts to respond to cattle raiding were ineffective, and officers were often suspected of collaborating with thieves (Heald Reference Heald2002: 2). In this context, Sukuma villagers in central Tanzania developed a local solution and formed village defence groups known as sungusungu.Footnote ² By 1987, Abrahams (Reference Abrahams1987: 182) observed that such institutions were common across Sukuma and NyamweziFootnote ³ inhabited areas in the regions of Mwanza, Shinyanga and Tabora.

Sungusungu comprised an elected village-level leadership of typically male elders with expertise in medicine and divination,Footnote ⁴ and all able-bodied male residents could be called upon to participate in the recovery of stolen cattle, armed with bows and arrows that were often poisoned (Masanja Reference Masanja, Forster and Maghimbi1992; Bukurura Reference Bukurura1994a: 4).Footnote ⁵ Village residents contributed a small amount of money to a sungusungu fund, which was used to provide food and supplies for those sent to apprehend thieves, and was subsequently replenished through fines paid by offenders (Bukurura Reference Bukurura1994a: 12).

Sungusungu appear to have been very popular amongst local populations, and to have been credited with reducing rates of violent crime, apprehending suspected thieves and returning large numbers of stolen cattle (Abrahams Reference Abrahams1987: 187). Those suspected of committing a crime were subject to brutal methods of investigation and punishment. Charges were not laid before the accused; rather the entire village population was expected to cooperate in ostracising the individual and their family until they were ready to confess their misdeeds to the sungusungu committee. Those who refused to confess to an offence were subject to torture and punishment, which sometimes resulted in death (Abrahams & Bukurura Reference Abrahams, Bukurura and Hann1992: 94). In the words of one former secretary of sungusungu interviewed in Mwanza: ‘they did their investigation with a stick … you would be beaten first, then you had to surrender. That's it’ (Ward Executive Officer-C 2011 int.).Footnote ⁶

At the time, some early academic reflections on sungusungu presented a somewhat idealised picture of a ‘people's organisation’ (Masanja Reference Masanja, Forster and Maghimbi1992) or a peasant-based movement of ‘popular resistance’ against an alien and oppressive state (Campbell Reference Campbell1989). However, this neglects the extent to which sungusungu practices played a role in perpetuating hierarchical intra-community relations. Whilst the initial objective of sungusungu appears to have been the recovery of stolen cattle, they were also known as bashosha bakima, or those who recover women, referring to the retrieval of women who had left their husbands or had eloped, thereby depriving their parents of bridewealth payments (Bukurura Reference Bukurura1994a: 21–3). The sungusungu also tried and punished suspected witches, usually older single women (Bukurura Reference Bukurura and Abrahams1994b; Mesaki Reference Mesaki, Haram and Yamba2009).

The TPF initially attempted to curtail the activities of sungusungu, and many early participants were prosecuted and imprisoned (Abrahams & Bukurura Reference Abrahams, Bukurura and Hann1992: 95). Judicial opinion of sungusungu was also largely negative, and in 1987 the High Court ruled sungusungu to be unconstitutional and illegal (Widner Reference Widner2001: 377). In contrast, the ruling party, CCM, was quick to recognise the possible value of sungusungu, both ideologically and pragmatically. As early as 1982 President Nyerere praised sungusungu, describing them as a revolutionary movement deserving of support and stated that they should be equipped with weapons whilst urging the release of all those who had been arrested (Masanja Reference Masanja, Forster and Maghimbi1992: 204). He subsequently pardoned convicted sungusungu and issued a general amnesty in 1989 (Mwaikusa Reference Mwaikusa, Semboja and Therkildsen1995: 172–3). Support for the movement was both ideologically and practically motivated. Sungusungu was consistent with Nyerere's ideology of Tanzanian socialism and its emphasis on local self-reliance and grassroots participation in development; however, the utility of an additional local security provider was also highly apparent at a time when state law enforcement bodies appeared incapable of responding to security threats. Thus, embracing sungusungu ‘happily chimed with both necessity and ideology’ (Heald Reference Heald2002: 9).

In 1989 the People's Militia Act was amended to incorporate sungusungu, formalising their role in crime prevention to some extent.Footnote ⁷ The formation of local security organisations modelled on sungusungu was subsequently promoted around the country as part of a government anti-crime drive spearheaded by then Minister of Home Affairs, Augustine Mrema. Participation in neighbourhood patrols was made mandatory for all able-bodied males between the ages of 18 and 60 in rural and urban areas, and was monitored by CCM's locally elected representatives, the party's ‘ten-cell leaders’,Footnote ⁸ who organised timetables for patrolling and were responsible for reporting shirkers to the police. Those evading patrol duties or refusing to pay a fine imposed by sungusungu were prosecuted and sometimes jailed (Kakoti Reference Kakoti1998: 61). At the same time, residents were encouraged to contribute cash, materials or labour to build local police posts, and rewards were offered to incentivise citizens to share information with the police (Mallya Reference Mallya1994: 75; Kakoti Reference Kakoti1998: 57–8).

During this period pre-existing sungusungu organisations were increasingly brought under the auspices of the local administration. Hangaya reports that secretaries of sungusungu were asked to take the place of absent village secretaries and rank and file sungusungu participants were deployed to collect the development levy (Hangaya Reference Hangaya1989: 37–8). There are also reports of sungusungu being utilised to enforce participation in communal work initiatives and nation-building projects, police public meetings and pursue tax defaulters, and sungusungu representatives were included in District Security Committees (Abrahams Reference Abrahams1987: 191; Shivji Reference Shivji1990: 17; Heald Reference Heald2002: 9–10; Grawert Reference Grawert2009: 397).

Sungusungu in urban Mwanza had a slightly different trajectory than in other urban areas, as the city is located in the Lake Zone,Footnote ⁹ where sungusungu originated, and had a majority Sukuma population. As early as 1983 the government ordered the deployment of sungusungu from rural areas in the towns of Mwanza, Shinyanga and Tabora (Campbell Reference Campbell1989: 46), in order to eradicate ‘racketeers, loiterers and prostitutes’ (quoted in Shaidi Reference Shaidi1989: 258). Local sungusungu groups were subsequently established between 1984 and 1987 to deal with high rates of armed robbery and burglary. Whilst sungusungu in Mwanza retained the leadership structure, practices and paraphernalia of rural sungusungu, it is commonly understood to have been ‘agizo ya chama’ (an order from the party), and mirrored the administrative structure of CCM, so that each branch office of CCM had a corresponding sungusungu office.

Party sponsorship of sungusungu had a significant impact on rates of reported crime. During 1991–2, at the peak of sungusungu operations around the country, a 60% reduction in crime rates was recorded, and notably reported armed robbery declined by 72% (Safer Cities Dar es Salaam 2000: 47–8). However, the compulsory nature of participation made sungusungu highly unpopular in urban areas and night patrols proved unsustainable (Mwaikusa Reference Mwaikusa, Semboja and Therkildsen1995: 71). Complaints were reported in the largely government-controlled media of harassment of civilians by sungusungu and false arrests, and the risks involved in patrolling were also contentious (Kakoti Reference Kakoti1998: 63–4). Ultimately, most of the groups established as a result of government orders were short-lived, due to difficulty in sustaining compliance or internal conflict between members (Mwaikusa Reference Mwaikusa, Semboja and Therkildsen1995: 171).

The trajectory of sungusungu from a popular movement to ‘another instrument of state coercion’ (Heald Reference Heald2006: 273) is by no means unique. Rather, the extent to which participation in sungusungu came to be determined by the interests of the party and the administration and to be enforced from above can be described as ‘typically Tanzanian’ (Mwaikusa Reference Mwaikusa, Semboja and Therkildsen1995: 170). Whilst popular participation in development was a core tenet of Tanzanian socialism, this was to occur in accordance with ruling party directives and in pursuit of objectives determined from above.

Following independence in 1961, bringing about ‘development’ (maendeleo) was central to the claims to legitimacy of the new nationalist regime, as elsewhere in Africa, and development was placed at the heart of discourses of nationhood (Jennings Reference Jennings2007). Participation in development was thus an act of ‘nation-building’ and was framed as an obligation of citizenship: ‘to be a Tanzanian was to assist in the development of the nation’ (Jennings Reference Jennings2007: 72). ‘Self-help’, referring to small development projects that were to be initiated and undertaken by communities themselves, was incorporated into official development strategy from the early 1960s (Jennings Reference Jennings2003). The importance of popular participation in local development was reinforced by the 1967 Arusha Declaration, in which then President Nyerere articulated the ideology of Tanzanian socialism, or ujamaa, underpinned by the principle of ‘self-reliance’ (kujitegemea).

Although early self-help typically did reflect independent community-level cooperation in pursuit of local priorities, the party and administration increased control over the planning and funding of local development initiatives throughout the decade, and participation was increasingly enforced through legal means. In 1969 the Ward Development Committees Act granted the committees power to compel participation in self-help, and to impose fines for failure to do so, as well as removing decision-making further from the village, making self-help more akin to ‘compulsory labour’ (Jennings Reference Jennings2003: 182). Government officials presented themselves as agents of ‘modernity’, in juxtaposition to the ‘underdeveloped’, ‘backward’ and ‘traditional’ masses (Schneider Reference Schneider2007: 28), whose role was not to determine development priorities, but to provide material resources and labour in accordance with directives from above. Shivji (Reference Shivji1990: 50) thus characterises the experience of self-help as ‘a long nightmare of harassment, beating, extortion of bribes etc. from ordinary citizens by the organs and officials of the state’.

Autonomous forms of organisation outside of the ruling party were increasingly no longer tolerated as Tanzania became a de jure one party state in 1965 and local government organisations were abolished in 1967. Although government ideology continued to emphasise the importance of popular participation, this was to occur only through party institutions rather than locally autonomous organisations (Tripp Reference Tripp, Hyden and Bratton1992: 228–9).

As was the case for state-sponsored forms of sungusungu, it proved difficult to sustain participation in and popular enthusiasm for self-help dictated from above (Pratt Reference Pratt, Mwansasu and Pratt1979: 205; Caplan Reference Caplan, Forster and Maghimbi1992; Schneider Reference Schneider2007: 25). Development models predicated upon ruling party dominance were undermined further when a multiparty system was reintroduced in 1992. Maintaining participation in sungusungu became increasingly difficult in some areas due to the strong association between sungusungu and CCM. As one ten-cell leader in Mwanza explained: ‘Initially it [sungusungu] was compulsory … because there were no opposition parties. The multiparty system has brought greater freedom’ (ten-cell leader B-3 2011 int.). The brutality exercised by sungusungu also made them unpopular, and many leaders were arrested once they could no longer rely upon protection from the ruling party. Although sungusungu has continued to operate in many areas, particularly in the Lake Zone where it originated, and other areas with high levels of Sukuma and Nyamwezi migration, in the majority of urban areas, including Mwanza, sungusungu had largely ceased to operate by the mid-1990s.

CONTEMPORARY COMMUNITY POLICING: SUNGUSUNGU TO ULINZI SHIRIKISHI

Community policing as understood in Tanzania entails both improving popular perceptions of the police force, in part to improve the TPF's ability to solicit information from citizens, and giving citizens a more active role in crime prevention by encouraging the formation of local security organisations under the auspices of local government. As Heald (Reference Heald2002: 9) observes regarding CCM sponsorship of sungusungu during the 1980s and 1990s, community policing is ideologically palatable, in this case connoting more ‘democratic’ and public service-oriented policing, but also reflects a pressing need to increase the human and material resources available to the police.

First, community policing is designed to improve trust in the police and to increase understanding of their work. The Tanganyika Police and Prisons ServiceFootnote ¹⁰ was established in 1919, and was, like other African police forces, a colonial creation, orientated towards upholding the colonial order rather than providing a public service (Killingray Reference Killingray1986; Anderson & Killingray Reference Anderson, Killingray, Anderson and Killingray1991). During the colonial period, police recruitment privileged ethnic groups deemed particularly suitable for martial roles, including those from other British colonies, and official policy sought to prevent officers from serving in their home areas (Burton Reference Burton2003: 69–71). Thus, Burton (Reference Burton2003: 83) describes popular perception of the police in Dar es Salaam as having been similar to that of an ‘occupying force’. The orientation of colonial police forces towards upholding the ruling regime was typically little altered following independence, as African governments realised the utility of control of the police in consolidating their power over newly independent states (Marenin Reference Marenin2009b: 349). During Tanzania's de jure one-party era (1965–1992) the TPF, like other branches of the administration, was closely tied to CCM. The return to multiparty politics was not accompanied by related alterations in the legal and regulatory structure of the police force. CCM retained the presidency and a parliamentary majority when multiparty elections were held in 1995, as it has done in subsequent elections in 2000, 2005 and 2010, and police partisanship has been reported during every election campaign, at local and national levels (Hoffman & Robinson Reference Hoffman and Robinson2009: 132; EU 2010: 23–8; Ewald Reference Ewald2011: 237–42; Makulilo Reference Makulilo2011: 251). Since the most recent general election in 2010 the violent policing of opposition party meetings and demonstrations has resulted in civilian fatalities, leading the independent Legal and Human Rights Centre (LHRC) to conclude that the TPF ‘is not free from political control’ (LHRC 2013: 26). Furthermore, the police are consistently ranked as Tanzania's most corrupt public institution in national surveys (e.g. LHRC 2012; Transparency International 2012).

The LHRC reports that 246 citizens were killed by the security forces between 2003 and 2012 (2013: 20–1). Police are often quick to resort to use of weapons and it is rare for alleged perpetrators of armed robberies to be apprehended or to appear in court as police are likely to shoot suspects on sight. Human rights organisations and the media have also documented many cases of horrific police brutality towards those arrested and the use of torture to extract confessions (CHRI 2006: 15–16; Human Rights Watch 2013). More recently, excesses perpetrated by security forces during ‘Operation Tokomeza’, intended to eradicate poaching, have been widely reported in the national media and led to the removal from office of four ministers.

Community policing is in part designed to mitigate these impediments to popular trust in the police in order to facilitate intelligence gathering and encourage popular cooperation with the TPF, and some innovative strategies have been pursued to this end. The mobile telephone numbers of senior police have been released to the public to enable citizens to circumvent low-level corruption and maintain anonymity when reporting crimes, addressing popular assumptions that local police release data to those accused of committing crimes in return for bribes. Police media campaigns have been designed to educate the population about the work of the police, and a ward police officer (polisi kata) has now been allocated to many of the country's wards. However, a discussion of the implications of these reforms for police-popular relations is beyond the scope of this paper, and the remainder of this section addresses the second objective of community policing, whereby responsibility for neighbourhood security is to be shared with communities themselves.

Tanzania's police-to-population ratio is extremely low, being one officer per 1156 citizens at the beginning of 2013 (The Guardian 2.2.2013). Although the current figure of 38,847 officers represents a considerable increase from 29,057 officers in 2008, it remains difficult to attract potential recruits as rates of pay are low and the hours worked make it difficult for police to supplement their income through other activities, as is the norm for other public servants (TPF 2008a: 36). The 2009 police annual report, for example, acknowledges the challenge posed by an: ‘Unmotivated workforce resulting in poor morale since TPF personnel are unable to meet professional and family commitments due to poor remuneration and poor conditions of service’ (TPF 2009: 26). Limited mobility due to a lack of funds to buy vehicles or fuel makes it difficult for police to respond to incidents in a timely fashion and restricts the geographical coverage of the police force, both in rural areas, where uniformed police are rarely seen (Stavrou & O'Riordan Reference Stavrou and O'Riordan2005: vii), and in urban areas such as Mwanza, with sprawling and inaccessible informal settlements.

Whilst the relatively wealthy can hire guards from one of the rapidly proliferating commercial security providers operating in the country, this is not an option for the vast majority of the population, who are instead expected to secure themselves through the ostensibly low-cost initiative of community policing. In the context of severe resource constraints, the police have recognised the utility of ‘tapping on available resources i.e. the citizens’ (Tibasana Reference Tibasana2002: 167), and residents of low-middle income areas are encouraged to participate in local policing in a form very similar to the CCM-sponsored sungusungu of the 1980s and early 1990s described above.

Since 2006, police have held seminars for elected and appointed leaders within the lowest tiers of local government in order to explain the concept of ulinzi shirikishi and how to go about establishing a local security group. Leaders are instructed to organise public meetings at which a security committee (kamati ya ulinzi) is to be selected and a strategy for organising night patrols should be decided. As was the case for sungusungu, it is anticipated that all able-bodied men within each mtaa (plural mitaa) or sub-ward, the lowest level of local administration in urban areas,Footnote ¹¹ should take responsibility for patrolling.Footnote ¹² Other residents undertake to provide goods or money to ensure the guards have equipment necessary for patrolling (such as batons, whistles and mobile telephones) and may receive some posho (subsistence allowance). The security groups are incorporated into the existing hierarchical organisation of local governance, whereby the young security guards are overseen by the security committee, which like other committees comprising the governance of the mtaa, such as those for health and sanitation, is subordinate to the mtaa chairperson.

According to guidance on establishing ulinzi shirikishi provided by the TPF, the duties of local security groups include: preventing crime and offering timely assistance to the community when required; educating the community about strategies of preventing crime and bringing about ‘development’ (maendeleo) in the area; providing a first response in times of trouble or natural disaster such as fire or floods; the prompt delivery of statements to the police regarding any suspicious behaviour; and ridding the area of any person(s) carrying out criminal activity (TPF 2008b: 15). Guidelines provide little information about how local guards are to go about performing these functions and variation is apparent between different administrative areas. Typically, however, night patrols form the core activity. In contrast to the enmity that characterised relations between sungusungu and the police, the community police are expected to deliver any suspected offenders to the nearest police post.

In addition to the similarities between the organisational models of sungusungu, as promoted by CCM, and community policing, the concept of ulinzi shirikishi is often closely linked with sungusungu in popular understandings of the policy. Furthermore, the history of popular participation in local policing is credited with facilitating the introduction of community policing today. Community policing groups are sometimes called sungusungu and in some parts of Mwanza, Kisukuma terms for the sungusungu leadership structure have been recycled for ulinzi shirikishi. In all three sub-wards in which research was carried out, people who had previously held leadership roles within sungusungu were closely involved in community policing. Media reports often use the terms ‘community policing’ and ‘sungusungu’ interchangeably. Although a common narrative describes community policing as an improved version of sungusungu, due to greater collaboration with the police and ostensible respect for human rights, some residents did not understand ulinzi shirikishi to be something distinct from sungusungu. For example, one ten-cell leader explained: ‘It's not that sungusungu has died. Its name has changed. Now we call it ulinzi shirikishi, but its origin is sungusungu … there's no difference’ (ten-cell leader-B 4 2011 int.).

Thus sungusungu has to some extent provided a useful referent for police, who are able to build on existing structures, or memories of them, to construct the apparatus of community policing today, albeit with additional concern for the protection of human rights and adherence to state law. The following section outlines additional similarities between the operation of ulinzi shirikishi and sungusungu, in terms of the understandings of citizen participation forged during the one-party era that continue to influence the practice of local development in Tanzania.

CONTINUITIES IN LOCAL DEVELOPMENT: UJAMAA TO PARTICIPATION

The importance of self-help and self-reliance to the development strategies of the one-party state was outlined above. Popular participation in local development, of which ulinzi shirikishi is an example, remains a key tenet of local governance in Tanzania today, and is a central way in which many collective goods and services are provided (or not provided), reflecting the extent to which participatory development has become ‘development orthodoxy’ (Cornwall Reference Cornwall2006: 62). Contemporary international support for participation is predicated upon the assumption that involving ‘communities’ in the planning and implementation of local development strategies, will produce outcomes that are closer to popular demands, more effective and more sustainable, and will empower citizens to better engage with democratic political processes. In Tanzania, however, it appears that these perspectives on the role of participation coexist with older understandings of participation as an obligation of citizens entailing the contribution of goods and resources towards kujenga taifa (building the nation) (Marsland Reference Marsland2006).

Analyses of the local governance of development in urban and rural areas have demonstrated considerable continuities in practice that appear somewhat contradictory to the stated aims of many participatory or ‘community-based’ projects sponsored by international donors. Rather than providing input into development policy and shaping priorities, citizens are often involved largely as providers of resources, such as labour, information or money (Marsland Reference Marsland2006; Chaliga Reference Chaliga2008; Rabé and Kamanzi Reference Rabé and Kamanzi2012). Local participatory institutions, such as community-based organisations (CBOs) may be largely envisaged as a means of assisting government, particularly in mobilising local resources (Dill Reference Dill2009), much as was expected of the lowest tiers of local government under the ‘minimalist orientation’ of ujamaa (Green Reference Green2010a: 24). Government officials continue to see themselves as responsible for telling ignorant citizens how to achieve development (Green Reference Green2010a: 26), and participation ‘has often come to mean teaching the community to act in the way that … experts require them to’ (Marsland Reference Marsland2006: 68). Ongoing local government reforms in which principles of participation are enshrined have done little to alter the top-down nature of development planning, and communities are therefore largely expected to recognise their responsibilities within government-determined projects, rather than contribute to the development of new ideas and ways of doing things (Chaliga Reference Chaliga2008; Green Reference Green2010a, Reference Green2010b: 1246; Norman & Massoi Reference Norman and Massoi2010; Rabé & Kamanzi Reference Rabé and Kamanzi2012). As was the case during the socialist period, popular participation is frequently underpinned by the threat of fines or even imprisonment (Green Reference Green2010a: 25, Reference Green2010b: 1241).

Ulinzi shirikishi has proved little different in this regard, which can to some extent explain the ‘ambivalence of success’ (Cleaver & Toner Reference Cleaver and Toner2006: 216) noted in the introductory section. Although ‘community-based’ projects may improve residents’ quality of life, in this case in terms of urban safety, it should not be assumed that this reflects ‘community’ ownership and inputs, and that ‘community’ institutions do not represent distinct groups with private interests (Cleaver & Toner Reference Cleaver and Toner2006: 216–17).

Evidence from Mwanza suggests that community policing has had a very positive impact on perceptions of local safety. Local government leaders and police officers unanimously praised ulinzi shirikishi as a successful innovation that had substantially reduced crime. The large majority of residents surveyed also credited community policing with having improved neighbourhood safety where it was operating. For many, ulinzi shirikishi provides a policing service that was deemed largely absent before, as limited police capacity means that officers are rarely seen in low-middle income residential areas and are widely perceived to do very little to address reported crime. However, perceived positive outcomes in terms of neighbourhood safety did not correspond with a sense that community policing was a ‘community-based’ initiative, and sustainability tended to be predicated upon the coercive power of the mtaa government to compel residents to provide labour and material resources.

As noted above, the process of introducing community policing in Mwanza's sub-wards began with police instructing leaders as to the responsibilities of local government and other community members with regard to security. This knowledge was then taken back to the mitaa and presented to public assemblies where proposed systems for neighbourhood security were endorsed. As one resident explained (survey respondent – 10 2011 int.):

Those responsible for community policing were supposed to be selected by the wider community, and thus subject to some form of electoral accountability. However, the elections, conducted at public meetings, entailed a person's name being proposed, usually a high-status individual, and confirmation of the nominee by those assembled (see also Lange Reference Lange2008: 1125). Thus those who claimed to have participated in the selection of the leaders for community policing did not find it easy to identify criteria they had used to select candidates. Survey respondents commented that they had not used any criteria as participation meant only assenting to the appointment of the person nominated through a show of hands. Others described the leaders as having been chosen by the mtaa government or chairman to join them in local leadership, meaning that ‘they chose themselves in the office’ (survey respondent-14 2011 int.). Only 13% of the 108 people surveyed across the three mitaa said they had played any role at all in selecting the leaders of community policing institutions, despite the fact that leaders often described themselves as having been chosen by the ‘community’. Survey respondents were thus more likely to refer to the community police as a distinct group of people with their own interests and objectives, rather than a ‘community owned’ institution over which they had influence.

The principal opportunity urban residents have to solicit information from leaders and request justification for actions taken is during public meetings, which the law requires be held at least every two months. However, these gatherings typically offer little opportunity to subject leaders to answerability or to have an input into local policymaking (Chaliga Reference Chaliga2008; Rabé & Kamanzi Reference Rabé and Kamanzi2012). Public meetings are forums ‘which appear inclusive but within which very few people have authority to speak’ (Green Reference Green2010b: 1256). Residents may thus often view meetings as an occasion when they are informed of decisions that have already been made by authorities at the mtaa level or above (Rabé & Kamanzi Reference Rabé and Kamanzi2012).

Regarding community policing, after the consent of the community had been obtained at a public meeting the participatory element of ulinzi shirikishi largely entailed popular contributions of money or labour, and volunteering information about local crime. This typically proved very difficult to sustain, in part due to considerable collective action problems resulting from the lack of accountability and transparency that characterises local governance in Tanzania, which allows perceptions of corruption to flourish. Rather than reflecting an outpouring of popular energies in pursuit of mutual aims, where community policing was sustained it was largely due to its codification in by-laws and the ability of leaders to impose fines or report shirkers to the police for non-compliance, and at the cost of considerable time spent visiting the houses of those scheduled to patrol and loudly threatening them until they emerged from their homes.

In this context, participation was often described as a result of coercion, whereby, ‘they force you to do it. It's compulsory, otherwise you get fined’ (survey respondent-13 2011 int.). Rather than increasing social cohesion necessary to address local development problems, as anticipated by TPF literature regarding ulinzi shirikishi (TPF 2008b: 20), enforcing participation in community policing tended to damage relationships between leaders and other residents who resented being dragged from their beds or woken in the night as their neighbours were harangued by the community police. Community policing thus entailed considerable intrusion into the homes and lives of other residents, and leaders were widely perceived as needlessly aggressive when carrying out their duties, to the extent that some claimed that leaders had used or threatened violence when collecting them for duty.

Complaints from residents and difficulties in sustaining popular participation in patrols were often dismissed by leaders, and local police, on the basis that those who complained were either criminals themselves, or needed more education about community policing in order to understand its importance and their responsibilities within the system. This perspective was also apparent in the way local police officers spoke about community policing, and they credited the reforms with having made people realise that they shared responsibility for local security with the police.

When explaining the need for ulinzi shirikishi, police place considerable emphasis on the lack of police manpower, which makes participation a responsibility of citizens. This is illustrated by an excerpt from a training session provided by a police sergeant in Mwanza for members of a CBO, observed on 19 November 2011:

The expectations police have of ulinzi shirikishi are consistent with state perspectives of the role of self-help in local governance more broadly. As Dill (Reference Dill2009: 730) argues with regard to CBOs in Dar es Salaam: ‘a CBO's raison d'être was to assist the government’.

Despite demonstrating continuities with one-party era strategies of popular mobilisation, however, the experience of ulinzi shirikishi in urban Mwanza also suggests that compulsory participation in local policing and other ‘typically Tanzanian’ (Mwaikusa Reference Mwaikusa, Semboja and Therkildsen1995: 170) forms of participatory development may increasingly be contested as party political competition on the Tanzanian mainland intensifies. The association of local development initiatives with ruling party electoral strategies and a more general perception that citizens are able to refuse directives to participate are discussed in the following section. This is also not historically unprecedented, but rather represents an additional parallel between sungusungu and ulinzi shirikishi. Although sungusungu in its state-sponsored form thrived with the support of CCM, the institution's coercive power was undermined by the emergence of opposition to the ruling party, precipitating sungusungu's decline.

THE POLITICS OF LOCAL DEVELOPMENT IN TANZANIA

Like Tanzania's other large urban centres, Mwanza has become an opposition stronghold over recent years. Although CCM has retained the presidency and a parliamentary majority since the reintroduction of multiparty competition in 1992, elections held in 2010, a month before the commencement of this research, saw a surge in opposition fortunes, and both of Mwanza's two urban constituencies, Nyamagana and Ilemela, were taken from CCM by the Chama Cha Demokrasia na Maendeleo (CHADEMA). CHADEMA was also much more successful than in the past at the local level, winning 11 of 21 wards in the city, and a CHADEMA affiliated councillor was selected to be mayor.

Individual candidates in mitaa and village-level elections must also be affiliated to a political party, and governments at this level may contain representatives of multiple parties. This can lead to conflict, as bringing ‘development’ to one's constituency brings potential electoral gains, and local politicians may thus seek to block development initiatives that are likely to be credited to other actors, particularly if they represent an opposing political party (Marsland Reference Marsland2006: 73–6; Lange Reference Lange2008; Rabé & Kamanzi Reference Rabé and Kamanzi2012: 61). In addition, ulinzi shirikishi is associated with CCM due to the party's previous sponsorship of sungusungu and promotion of other forms of local self-help.

In the three sub-wards in which research was carried out, local governments were dominated by CCM candidates, in contrast to some neighbouring mitaa where party political competition within committees had prevented establishment of community policing. However, leaders in all three mitaa acknowledged that some residents continued to associate the policy with CCM. One security committee secretary, for example, claimed that ‘vyama vinasumbua’ (parties cause trouble), as people sometimes protested that they could not guard because ‘it's CCM's work, it's CCM's idea’ (Secretary of community policing-C 2011 int.). Mwanza is not unique in this regard and senior police at the national level report a perceived link between CCM electioneering and community policing (Tanzania Daima, 15.12.2009; The Guardian 27.3.2011). The former Inspector General of Police, for example, in 2011 urged citizens not to use their political affiliation as an ‘excuse’ not to participate (Daily News 12.7.2011).

In addition to these partisan objections to participating in ulinzi shirikishi, mtaa government members, ten-cell leaders and police complained of more general challenges in organising collective development efforts due to a shift in political culture associated with the transition to multiparty politics. Residents were described as having the ‘uhuru’ (freedom) and knowledge necessary to question leaders and refuse to participate. In the absence of opposition parties citizens could be forced to comply with instructions to some extent, whereas today in the multiparty context people are more likely to question directives, to demand information about how money that is collected is being used, and to consider providing ‘development’ or policing to be something for which the government should take responsibility. As one ten-cell leader, who had previously been a sungusungu kamanda (commander) in the early 1990s, explained (ten-cell leader-C 5 2011 int.):

This forms a stark contrast to when sungusungu was first introduced in Mwanza, when people had no choice but to follow CCM directives because ‘the party was the government’ (Ten-cell leader-C 6 2011 int.). An additional challenge to community policing in the multiparty context arises from the role ten-cell leaders are expected to play in organising local collective action. Ten-cell leaders are useful in organising ulinzi shirikishi due to their knowledge of other residents and ability to monitor participation, as well as performing tasks such as collecting money. However, the legitimacy of their role in local governance is sometimes questioned because their authority is not recognised by opposition party supporters, who perceive them to be CCM representatives.

The capacity of CCM and its network of ten-cell leaders to enforce participation in top-down schemes during the one-party era should not be overstated, as noted above. However, it is possible that the emergence of more viable challenges to CCM dominance offers citizens a new means of resisting coercive practices of local development that impose costs on citizens.Footnote ¹³