STREET POLITICS

Thieving, hustling and cheating, touting minibuses, fencing stolen goods and selling marijuana encompass a broad range of enterprise in terms of the ‘wages’ they provide, the networks and levels of organization in which they are embedded, and their position as licit or illicit, legal or illegal activities. What all these have in common is that they all occur on the street and are intrinsically linked to practices, identities and meanings of streetlife. Yet this street economy is not a free market of relations and economic and social opportunities; power dynamics and hierarchies are also at work.

Petty crime and hustling constitute the bulk of street activities. Notably, petty theft and assault, especially at night, are historically the most common offences on the streets of Addis Ababa (Brown Reference Brown1973; Tesfaye Reference Tesfaye and Beyene1988; Wondimu Reference Wondimu2004). Those who commit these offences are not always professionals; they are individuals who alternate self-employment with occasional engagements in property crime to get by (Wondimu Reference Wondimu2004). A level of professionalization and organization, however, does exist and relates to certain street businesses, including touting minibuses, fencing stolen goods and to some extent the retail sale of marijuana. As I learned in my time on the streets, members of these organized groups usually belong to particular networks, for instance of old friends coming from the same neighbourhood. Hence, for a hustler or occasional thief, becoming a professional fence or being part of the group of minibus touts is rarely easy or straightforward (Di Nunzio Reference Di Nunzio2012b).

From the perspective of government institutions, tackling petty and ordinary crime has only rarely figured on the political agenda, either of the current ruling party or of its predecessor, the Derg – the military junta that led a socialist regime in the mid-1970s and 1980s. The existence of street crime was not per se a reason for political intervention or political confrontation.Footnote ⁵ Moreover, ordinary crime persisted throughout and co-existed with political conflict, as Tesfaye (Reference Tesfaye and Beyene1988) showed in his analysis of crime incidence at the time of the Derg. At the same time, as a veteran of streetlife whom I interviewed emphasized while talking about the Derg period, politics did not concern street thugs (duruye), suggesting to me that disengaging from politics was a fundamental tactic for being able to pursue one's activities on the street. This did not necessarily imply that offenders were not prosecuted during the Derg. GB, for instance, was a well-known gangster in inner city Addis Ababa in the 1970s and 1980s and was arrested many times, as a former soldier of the Derg and a resident of the inner city told me:

In this context, disengaging from politics helped to moderate the level of punishment faced by street actors. It enabled individuals to be involved in illegal activities on the street without the risk of being charged and given sentences that would have been far more severe had politics been involved.

Regime change in 1991 did not lead to an appreciable change in the state's approach. Especially in the first ten years of its rule, the EPRDF dealt with streetlife as a terrain of ‘crime’ rather than ‘politics’. A statement I often heard on the streets that was attributed to the late Meles Zenawi, the long time prime minister and chairman of the EPRDF, shows how young men in the inner city understood the government's approaches to crime: ‘stealing is a job but if you are caught it is an offence’. Whether this statement is apocryphal is hard to tell, but the currency it has gained suggests that actors in the street economy recognized that being caught was not always a direct consequence of illegal activities: going to prison was an event that might occur, but also might not. That margin of probability made engaging with crime a ‘job’, a viable way of getting by.

However, politics has entered the street economy and streetlife more pervasively during particular historical junctures when the street constituted a terrain of political conflict. Even in these moments of unrest, government narratives were not necessarily about ‘crime’ per se. The regime's discourses and political actions criminalized the allegiances between political dissent and streetlife. In other words, those involved in the street economy became potential targets of state repression not because of specific crimes, but because at such moments they were considered to be part of a large constituency of politically ‘dangerous’ subjects. During these periods, then, state institutions focused on those groups, often organized ones, that were believed to play a political role on the street, either by mobilizing or violently repressing them. During the Red Terror at the end of the 1970s, for instance, crime and violence in the city became almost completely embroiled in the political conflict between the military junta and the civilian left, as Yeraswork Admassie, a sociologist at Addis Ababa University, explained to me (see also Zewde Reference Zewde, Tronvoll, Schaefer and Aneme2009; Donham Reference Donham1999; Lefort Reference Lefort1983; Markakis and Ayele Reference Markakis and Ayele1986). Niches of the street economy characterized by a certain level of organization witnessed a politicization of conflicts that, until then, had revolved around the control of particular illegal businesses. For instance, in Merkato, Addis Ababa's main marketplace, the clashes between two gangs, the China Group and the Genghis Group, over the extortion of local merchants became gradually embroiled in the violent struggle between the squads of the EPRP (Ethiopian People's Revolutionary Party), which opposed the Derg, and those of the Meison (All Ethiopian Socialist Movement), which initially supported the military regime (Yeraswork Admassie, personal communication).

The recent history of relations between the street and politics was similarly characterized by the regime's periodic attempts to capture the streets in moments of political conflict and by the moves ‘thugs’ made to navigate state politics. At the end of the 1990s, the government's concern with enrolling young people in the war against Eritrea resulted in a wider focus on mobilizing streetlife. The options offered to many in the inner city were twofold: go to the front and have their criminal records cleared, or stay at home and risk ending up in jail. Many young offenders, including some of my key informants, joined the army and fought in the Ethiopian–Eritrean war.

The riots of 2001 and 2005 brought the politics of the state into streetlife and the street economy more pervasively. On both occasions, the participation of young people in the riots was a manifestation of the mutual distrust that had long characterized relations between the population of the capital and the ruling party. The rural history of the EPRDF and its privileged and ideological links with the peasantry did not align with the grievances and expectations of the urban population (Tronvoll Reference Tronvoll, Tronvoll and Hagmann2012: 277). This, together with the EPRDF's persistent repression of public protest, led to the mushrooming of dissent and unrest in the city. In April 2001, the brutal repression of student demonstrations further contributed to the triggering of riots and protests on the streets of Addis Ababa. Four years later, the 2005 riots bore witness to an emerging alliance between the street and the opposition parties. That year, my informants in the inner city recalled, opposition supporters were seen campaigning on the street and reaching out to individuals. As one man with past involvement in the street economy told me, the opposition capitalized on the fact that many people on the street hoped to gain a position in government institutions if the opposition took power:Footnote ⁶

To EPRDF leaders, this alliance suggested a convergence between the criminal attitudes of Addis Ababa's street youth and the opposition's hunger for power.

Since the 1970s, and especially in moments of political conflict, the trope of the unruly and criminal marginalized youth has been part of the political discourse in urban Ethiopia. The Derg used the category of the lumpenproletariat in its periodic attempts to either mobilize or repress streetlife. Following the 2001 riots, a senior government official told the BBC that ‘gangsterism’ lay behind the unrest.Footnote ⁷ Similarly, in 2005, the ruling party labelled the protesters ‘dangerous vagrants’ (adegegna bozene) and jobless youth to emphasize the unlawful nature of street protests and demonstrations. Within this long history of criminalization, however, the way in which the EPRDF defined and deployed the notion of the dangerous vagrant was an unprecedented attempt to define the characteristics of those believed to be potential actors of unrest and crime. The issuing of the proclamation against the dangerous vagrants a year before the 2005 election (Federal Negaritgazeta of the Federal Democratic Republic of Ethiopia 2004: 2533) provided the ruling party not only with a legal tool to tackle unrest, but also with a very detailed description of the economies of hustling and exchanges that take place on the street.Footnote ⁸ Whereas previous notions of the lumpenproletariat and the gangster youth had offered a generic picture of the activities and attitudes of ‘street youth’, the proclamation on the dangerous vagrants established the ‘street’ and ‘unemployed’ youth as a political and juridical subject and a target of specific policies and interventions (cf. Mains Reference Mains, Tronvoll and Hagmann2012).

REHABILITATING THE ‘DANGEROUS VAGRANTS’

The proclamation against ‘dangerous vagrants’ was not designed to punish a particular crime, such as robbery, theft or burglary. A dangerous vagrant, as police officers pointed out to me, is a particular kind of offender. Lieutenant S., a crime investigator, told me that there is a difference between a thief (leba) and a dangerous vagrant. Thieves are, to some extent, professionals. Adegegna bozene are vagrants who disturb the ‘tranquillity’ and ‘order’. Hence, a dangerous vagrant is an offender because he continues to commit crimes, as Officer A., a community policing officer, told me, or because, as the proclamation states, he is ‘unemployed’ and has committed an offence. ‘If you have a job, you are not an adegegna bozene,’ a young attorney pointed out to me.

As a result, the response to the behaviours of dangerous vagrants, Lieutenant S. and his colleagues argued, cannot be just punishment, but must have a broader focus on ‘peace’ and ‘rehabilitation’. In this regard, Officer T., a community policing officer, whose professional story I will recount later, argued that government action has had a significant impact on the ways in which young people behave on the streets:

He then concluded by telling me that the change that he had seen in the sefer was also deeply rooted in the consciousness of young people themselves. In his view, many young people had been rehabilitated:

The change that Officer T. had seen in the behaviour of young people, I argue, was a result of the ways in which politics entered and reshaped streetlife. Rehabilitation was primarily a political process, which revolved around detaching ‘street youth’ from the opposition while integrating them into a formalized, regulated and politicized street economy. In the years after 2005, the notion of dangerous vagrants provided the ruling party with a way to make a clear distinction between the predicaments of the unemployed youth and the political interests of the opposition. Unlike the opposition parties, which the ruling party saw as mainly interested in toppling the legitimate government (Tronvoll Reference Tronvoll, Tronvoll and Hagmann2012), dangerous vagrants were portrayed as individuals troubled by social and economic problems, particularly unemployment. Unemployment was considered the reason for crime and dissent, and therefore rehabilitation revolved around providing employment. Employment tackled crime, while dissent was addressed by characterizing the rehabilitation of ‘unemployed youth’ as a process of political mobilization.

After calm was restored and opposition politicians imprisoned, government officials, party members and youth organizations worked to reshape the politics of the street. This included strengthening the links between the local government office and those niches of the street economy that had already gravitated towards the influence of the EPRDF. In addition, it simultaneously entailed the establishment of an institutional framework enabling local government officials to mobilize individuals who were engaged in the street economy but with low levels of professionalization or organization.

In the inner city, the strategies of the ruling party focused on those areas of the street economy involved in the transport sector: touting minibuses and collecting parking fees from car drivers. These were licit businesses that provided a service, such as managing the flow of minibuses in a particular motorpark or attending to cars, that, until recently, had been organized in an unregulated and illegal manner. The minibus touts began to appear in the last years of the Derg when minibuses started to circulate in Addis Ababa. The control of the motorparks and minibus stops was a profitable business and, as the above quote from Officer T. mentions, it triggered violence between competing groups. The regulation of territories controlled by each group of touts was among the first government interventions on the street in the late 1990s. While this reduced street violence linked to touting minibuses, it did not lead to a political mobilization of the street economy. It was only during the 2005 riots that the alliance between the ruling party and the minibus touts was forged. On 8 June 2005, taxi and minibus drivers staged a three-day strike that virtually paralyzed the city in support of students protesting the April election (Fortune 2005). In response, the Addis Ababa Transport Authority told the taxi drivers to resume their services within two days, on penalty of having their licences withdrawn (The Ethiopian Herald 2005). The minibus touts played an important role, first as promoters of the strike action, and later as enforcers of the government order. My informants told me that the leaders of the minibus touts were urged by the local government officials to stop the strike going further. In the years following the 2005 riots, minibus touts discovered the benefits of their collaboration with the ruling party. Following government interventions, minibus touts are now formally recognized as associations of ‘private investors’ in the urban transport economy. A man in his twenties who worked in one such group made sense of this transformation thus: ‘We were tara [literally, queue-keepers] with the stick, now we are with the mahaber [association or cooperative].’

The formalization of minibus touting revolved around strengthening existing relations and shifting political alliances. The establishment of ‘parking guys’ in 2008 enabled the ruling party to further expand the reservoir of individuals directly and indirectly linked to the party. Parking guys issue parking tickets every half hour to cars parked on those streets assigned to them by the local government and collect payments from the drivers. Before 2008, ‘parking’ was largely the street business of individuals, except in some areas where small private companies were involved. On the eve of the 2010 national elections, there were three to four parking groups in each kebelle (neighbourhood council) in Addis Ababa's inner city. These cooperatives provided employment for many who had previously found other ways of getting by, while increasing the number of people who were dependent on local government bureaucracy for their survival and thus were expected to support the ruling party when needed. Politics clearly framed the activities of the parking guys; as one of them reckoned: ‘You know, our job depends on the outcomes of the next elections.’

The politicization of streetlife since 2005 has suggested to parking guys and minibus touts in particular that engaging with the politics of the ruling party is now a fundamental condition for operating on the street. In other words, the political mobilization of street life reversed the understanding that ‘disengaging from politics’ – as, for example, thugs such as GB did during the Derg – was an essential tactic to avoid repression, or at least severe sentences. Due to the impact of government policies, many young people with past and present engagements in the illegal street economy have recently joined government programmes to get by, but also because it is fundamentally a way of staying away from the machinery of state repression. ‘Why steal when you can make the same money by working?’ a young man in his late twenties with past involvement in mobile phone snatching and theft asked, referring to the fact that he would have ended up in prison had he not joined the parking guys.

This process of formalization and politicization of streetlife, though successful, has its limits. Although parking guys receive a regular monthly income, their salaries are still low and their work has not brought the expected opportunities for social improvement. Also, the ruling party in the inner city has focused on a specific niche of the street economy, while the economies of hustling, thieving and fencing stolen goods remain active. In this regard, the presence of a well-organized gang of fences on one of the main streets bears witness to the endurance of street crime in the inner city. The work of the fences serves both thieves and the shop owners interested in buying stolen goods. Special arrangements with some shop owners in the area enable the fences to keep their business going. In practice, if a fence is caught, he will never reveal the shop owner with whom he is doing business. In exchange, the owner will pay to bail out the fence. As a result, the business of fences continues, as does their capacity to respond to the offer of stolen goods, meaning that engaging with theft remains a viable way of making do.

A NETWORK OF INFORMERS

The implementation of community policing schemes served the ruling party's concern with establishing itself at the grass roots in Addis Ababa, especially in the inner city. It enabled the ruling party to promote the participation of citizens in policing their own community, but within a framework that the ruling party positioned within the existing political structure. One of the core measures of the implementation of community policing schemes in the mid-2000s was the use of Peace and Security Committees. Each kebelle had a certain number of Peace and Security Committees; as a kebelle official told me, each sub-kebelle division (or tabia) had its own, as he called it, ‘peacekeeping committee’, usually with five or six members. These might include chairmen of burial associations (iddir) as well as affiliates of the ruling party, such as the leaders of local mass associations. The duty of these committees, as the text of the government's Urban Good Governance Package (Ministry of Works and Urban Development 2006) reads, was to mobilize:

While policy documents described the committees as a new institution by skilfully using the donor-friendly language of citizen participation, a previous incarnation of Peace and Security Committees existed in the early 1990s in both urban and rural areas (Vaughan and Tronvoll Reference Vaughan and Tronvoll2003: 39; Andargatchew Tesfaye Reference Tesfaye2004: 28; Hagmann Reference Hagmann and Kyed2007).

The institutional history behind the establishment of these committees not only suggests that the community policing schemes reflect local trajectories of the state, but, more importantly, it reveals that the primary motivation for the formation of these committees was not to open avenues for self-determination, but to make ordinary citizens active agents of the ruling party's apparatus of control. From the perspective of the EPRDF and its commitment to realize a collectivist and developed Ethiopia (Vaughan Reference Vaughan2011; Hagmann and Abbink Reference Hagmann and Abbink2011), this had a broader ideological significance: a vision of communities policing themselves is the highest realization of a dream of a collectivist and self-organized society made up of responsible and dedicated citizens.

Interestingly, the socialist regime that preceded the EPRDF – the Derg – engaged with a comparable vision of people's direct engagement with policing. During the Derg, armed citizen militias were major instruments of state violence against political opposition during the Red Terror (Zewde Reference Zewde, Tronvoll, Schaefer and Aneme2009). In the mid-2000s, however, the involvement of citizens in the ‘policing’ of their communities did not imply participation in the state machinery of repression. These schemes have not outsourced the use of ‘state’ violence against, for instance, leaders and activists of political opposition. The ruling party has continued to rely on the judicial system to publicly and openly criminalize political opposition and critical voices in the media using a variety of charges, including, most recently, terrorismFootnote ⁹ (Pausewang et al. Reference Pausewang, Tronvoll and Aalen2002; Tronvoll Reference Tronvoll, Tronvoll and Hagmann2012). Instead, the duties of community policing have mainly revolved around vigilance with regard to potential factors and actors of political unrest.

Gathering information, in fact, was one of the most important duties of the Peace and Security Committees. As a local official told me, members of peacekeeping committees reported weekly to members of the cabinet of the local government office (kebelle), who, in addition to their regular duties, supervised peacekeeping activities in their electoral constituencies. The justice kebelle official in charge of coordinating peacekeeping activities explained that the collaboration between the kebelle, the committees and other institutions was arranged in a pyramidal structure. The city administration was at the top, followed by the kebelle officials, who worked with the police and peacekeeping committees. Information about serious crimes and ‘sensitive issues’, such as politics or unrest, usually went upwards and was passed on to the city administration or even to the topmost levels of the federal government.

However, the fact that the political hierarchies and concerns of the state framed the activities of the peacekeeping committees does not alone explain the basis on which the ruling party managed to expand its capacity to control the local population. We need to understand how peacekeepers gathered information and how their activities served the political activities of the ruling party.

Foucauldian analyses of surveillance have often revolved around identifying variations of Bentham's Panopticon, through which local institutions develop their ability to control and expand their capacity to see (Foucault Reference Foucault1979). However, the streets and neighbourhoods of inner city Addis Ababa are far from constituting an Ethiopian Panopticon. Narrow side roads and the lack of street lighting can hide more than they reveal of what thieves, hustlers or political opponents are up to. In this context, state surveillance was thus not about ‘seeing’; the success of the structure of control depended on its capacity to ‘hear’ information travelling through communities. In other words, ‘peacekeepers’ (selam askabari) or, as they are pejoratively known, akataris and bandas (literally ‘spies’) or joros (‘ears’), were effective ‘informers’ for the government if they were able to participate in and literally ‘hear’ the gossip of their community. As I myself experienced, people often talked about their friends, neighbours and acquaintances, and news about people in the community quickly spread. In the work of the peacekeepers, the boundary between spying and gossiping was blurred: it might be said that informers gossiped with the kebelle's officials about the community just as neighbours gossiped about other neighbours. Gossiping was a way of mapping the community. Spying worked in the same way, but served forms of political control. As the ruling party built its network of ‘ears’, it produced a political community of loyalty. When individuals shared information with the government, they potentially betrayed the trust of those neighbours, acquaintances and friends who had passed them the gossip.

Community policing was successful for a range of reasons. It expanded the capacity of the government to ‘hear’ and to ‘know’. It strengthened relations between the ruling party and its most loyal supporters. It indirectly produced a climate of suspicion and mistrust that amplified the effects of peacekeeping on the streets and in neighbourhoods. Many in my field site could navigate the ruling party's machinery of surveillance and control by, for instance, interacting cautiously with individuals who were commonly believed to be informers because they were members of the ruling party. At the same time, some people worried whether their friends, neighbours and acquaintances were actually informers, and tended to refrain from doing anything that could draw the government's attention to them. As a young man living in my field site put it: ‘You could never start a [street] business because you cannot trust anyone, even some of your friends.’

At the same time, the ways in which ‘peacekeeping’ was carried out reveal the limits of the ruling party's apparatus of surveillance. I had the opportunity to meet a few individuals who were believed to be informers of the government. They were not necessarily experts in intelligence or disguise: they simply relied on their own networks to keep an ear on what was going on. As they did so, the range of relations and interactions they participated in affected the kinds of information they could and could not access. The different profiles of informers enabled the ruling party to tap into a variety of networks, including the minibus touts and the parking guys. This expanded the capacity of the ruling party to keep an eye on the networks that gravitated under the influence of the EPRDF, but did not necessarily open new channels for gathering information on what happened outside the reach of the community of political surveillance.

In fact, strengthening Peace and Security Committees and increasing the synergy between the police and the local government office actually reduced the channels through which the ruling party was able to collect information. Officer T., the policeman mentioned above, had been in the police force for eleven years but had started as a community policing officer just a year before I interviewed him in 2010. His job was partly to spread awareness about crime prevention and community policing in the community and generally prevent crime. His duties also included collecting information about offenders through the networks he built through his community policing activities. Among his networks and contacts, Officer T. told me that the local government officials were an important resource. They had useful information, much of it obtained from the government-controlled Peace and Security Committees. As Peace and Security Committees became primary sources for gathering information about crime, the street and the communities, the interaction changed between policemen and the ‘thugs’ on the street.

According to a veteran of streetlife, policemen had stopped talking to thugs (duruye). ‘In the past,’ he reckoned, ‘policemen would talk with the thugs,’ either to collect information or to find some form of ‘mediation’. After a few months in the field, I realized that policemen also felt that their relations with street thugs had changed. This became clear when I happened to follow two of my informants in one of their mellas (street dealings). They were helping a neighbour get her mobile phone back. Her brother had stolen it from her and had been caught by the police and held in detention at the police station. One of my informants confidently walked into the police station to talk to the investigators, and then headed to the common cell where the brother was held to get information on the fence to whom the latter had turned to sell his sister's phone. Seeing this, one policeman at the door commented to another: ‘These are old-style duruye. They are not afraid to talk to policemen…’ The narration of the veteran of streetlife I mentioned above and the policeman's reference to an old-style duruye – implicitly different from a new-style duruye – were not mere recollections about the past. Rather, by comparing the present and the past, both the former thug and the policeman were mapping their current positions on the street. They were making sense of the fact that the strategies of the ruling party had transformed the dynamic of their interactions and, potentially, had shaped the kind of information that one could get from the other. As policemen stopped talking to thugs, their ability to know about, and hence to deal with, ordinary crime did not improve – with consequences, as the last section of this paper shows, both for residents of the inner city and for the efficiency of the ruling party's strategies of crime prevention and community policing.

VIGILANTES IN ADDIS ABABA'S INNER CITY

One night in early October 2010, a young man from a neighbourhood in Addis Ababa's inner city died on his way home, strangled by thieves wanting to rob him. His attackers had tried to use the technique of čebu, in which robbers seek to make a victim lose consciousness by hitting his neck, but in this case it had ended badly. A young man living in the same neighbourhood as the victim told me that people had decided to organize themselves to prevent this happening again:

H., the manager of one of the groups of minibus touts (tara askabari) in my field site, became the headman of the night watchers. He did not see himself as a leader, but as a manager who organized the shifts every night: ‘We are ninety people now. Between eleven and twenty people walk at night,’ he said. ‘They start at ten and go until two. Then if they catch someone they will go on until three or even four in the morning.’

Over the days following the death of the young man, many people joined up. When I joined one of their night rounds, I saw many people I knew. Some were occasional ‘hustlers’ or individuals with past and present involvement in the street economy; others were workers. There were also young educated people who wanted to take care of their neighbourhood. After a few days, a meeting was held between the group, a police officer and kebelle officials. At the meeting, the night watchers said that they would organize themselves since the police were unable to keep order on the streets. A few days later, a night watcher told me: ‘Police are few, they sometimes come and they sometimes don't. People wanted to clean the area of thieves and we organized ourselves.’ Another observed more cautiously, ‘I don't blame them [the policemen]; they earn 600 birr (US$36) a month and it is nothing.’

This commitment to order and security was, however, different from the increasing political concern for ‘peace’ that framed government-supported community policing activities. As the group grew, the people of the neighbourhood and the police themselves described the night watchers as ‘peacekeepers’ (selam askabari). Officer T., the community policing officer, told me that the police saw the emergence of these ‘new peacekeepers’ in addition to the existing kebelle-supported peacekeeping committees as positive. The former went out at night and caught thieves while the latter shared information with the police and the kebelle. In the eyes of the officer, the two were complementary. Some of the ‘new peacekeepers’ did not necessarily see it in the same way. ‘We are in touch with the police but we don't want any kebelle or things like that. We are on our own.’ Thus, one of the night watchers emphasized that their activities did not necessarily engage with a politically framed idea of peace: they were there to catch thieves, nothing else.

However, the distinction between ‘peace’ and ‘catching thieves’, while valid for some night watchers, was not made by all. H., the leader of the night watchers, had an interesting story: he joined the minibus touts between the late 1980s and early 1990s, becoming the manager of his group in 2006. During that time, his career as a taxi tara was linked to different individuals and events. He had risen in the business because of his friendship with a man who had controlled several minibus motorparks in the old city centre since the 1980s. Many in my field site believed that H. became influential because of his role in the 2005 demonstrations when the ruling party had turned to the minibus touts to stop the strike of the minibus drivers. Since then, H. had become the manager of one of the biggest minibus motorparks in Addis Ababa's inner city.

The story of H. and his nomination as headman suggests that the distinction between involvement in the government's politics of control and ‘catching thieves’ was more blurred than some of the new peacekeepers claimed. He was a well-known, well-respected and established ‘street businessman’, which made him ‘suitable’ for the position. The same young man who told me that the night watchers ‘are on their own’ described H. in these terms: ‘We choose as our leader a taxi tara. He has a job, he has money and he knows a lot of people.’

The irony of this is that H. gained his position and strength on the street precisely because of his relation with the ruling party's politics of streetlife from which the young night watcher wanted to distance himself. H. was aware of his political position on the street and told me himself that he had become headman because of his involvement in the ruling party since 2005: ‘I have been in contact with the police for the job I do; for this reason I came to do this.’

The story of a parking guy and his involvement in night watching adds other insights, helping to better understand how some former ‘thugs’ experienced and navigated the blurred line between the politics of control and ‘catching thieves’. As with many parking guys, he had previously engaged in the illegal street economy. In doing this, he had not only found ways of getting by, but also of elaborating ideas of respect and self-worth. In 2009, he joined a government-supported cooperative of parking guys, looking for a stable income. When the young man was killed on his way home, he joined the night watchers. ‘You know it is a good thing. They [the thieves] sometimes do setete [house robbery] on their way back. These are people who come from another sefer, perhaps Shiromeda [north-eastern part of the city] or even from the rural areas.’ He continued: ‘They are daily labourers or work in construction sites. I can't blame them, they don't have money and they want to do a city life. But they are fucking dangerous, they have čube [knife] and they could stab you.’ A few weeks later, he quit. ‘I did not feel comfortable with catching thieves … I was a thief in the past too!’ This short engagement is revelatory. He joined because he thought it was a good thing for the neighbourhood, but then left because of his past as a thug (duruye). He came to see himself as being on the other side, and felt he could not continue as a night watcher; his life history and the respect that he had gained were at stake. ‘Catching thieves’ clearly showed the contradiction between his present and his past. He still saw himself as a tough man, but the political transformation of the street had come to define the terrains of his action, transforming him into a kebelle-supported parking guy. When a young man of the neighbourhood died, strangled by thieves, he felt he had to do something. ‘Catching thieves’, however, would have taken him too far because it revealed evidence of his subjugation, of his becoming docile.

The experience of this parking guy, the headman's career and the attempts of some night watchers to distinguish between ‘catching thieves’ and ‘peacekeeping’ reveal important aspects of the relationships between the ruling party and the street. The emergence of vigilantism in Addis Ababa's inner city aimed to address the simple fact that, despite the fuss about government peacekeeping activities, neighbourhoods remained insecure. The concern of some night watchers to distinguish themselves from the ruling party was an attempt to keep from indirectly contributing to the expansion of the party's machinery of surveillance. The increasing pervasiveness of the ruling party was considered to be an unwanted intrusion that limited the lives of ordinary citizens; many people in the field site cautiously interacted with individuals who were believed to be informers. In these circumstances, when residents in the inner city decided to organize themselves to patrol the streets, they were not only voicing doubts about the capacity of government institutions to deal with ordinary crime, they were expressing their dissent from the politicizing narratives encompassing the ruling party's approaches to crime prevention and community policing.

To many night watchers, patrolling the street at night seemed an efficient and politically uncompromised manner of fighting crime. It did not depend on information gathered from a politicized network of spying and surveillance. It directly targeted crime at night, when residents had personally encountered thieves. Having said that, night watching is not new in the inner city. Night curfews, for instance, were important strategies to contain crime in the early history of Addis Ababa (Assen Reference Assen, Zekaria, Zewde and Beyene1987) and throughout the Derg (Clapham Reference Clapham1988: 129). Although the EPRDF had long disengaged from imposing curfews, residents of the inner city continued to rely on a self-imposed curfew: being home by ten was considered the most efficient way of reducing the likelihood of falling victim to robbery. Night watchers grounded their understanding of crime and their perception of the efficacy of their action on what they saw as a logical consequence of the ways in which communities in the inner city dealt with crime: if someone was walking late at night on the streets and was not a person from the neighbourhood, he was up to no good and hence a potential thief to be handed over to the police.

Still, night watchers inhabited an ambivalent political space and terrain of action. While some night watchers were trying to define an autonomous space of action on the street, the ruling party was able to sneak in through the webs of connection and affiliations that local government officials had built. The fact that the headman was himself involved in the politics of the street and that a parking guy felt that being a night watcher was in contradiction with his past as a thug suggests that night watching was far from being a politically disengaged terrain of action. In their stories, blurring and drawing distinctions between ‘peacekeeping’ and ‘catching thieves’ were ways of negotiating the position that night watchers, both individually and as an institution, held in relation to the ruling party and its ability to penetrate and reorganize terrains of street practice. The night watcher who used to be a thug and ended up as a parking guy, for instance, both blurred and redrew the boundaries between ‘catching thieves’ and ‘political control’ in order to reconcile his former identity with his current status as a government-supported street worker. In doing this, he negotiated his relationship with the ruling party's politics of the street. Meanwhile, the headman's connections blurred these boundaries, bringing the state in, and thus making vigilantism a political terrain of action, indirectly linked to the strategies of the ruling party and its attempts to control the street.