Civil society-State partnerships for the promotion of citizen security in Brazil

Fiona Macaulay



This article examines the different modes of engagement between civil society and the state in the area of citizen security in Brazil. It begins by considering both the progress made in opening up new spaces for civil society interventions (in the role of advisor, watchdog and even service deliverer) across a number of policy areas, and the specific difficulties posed by the criminal justice system. It continues by analyzing the activities of non-state organizations in two fields: policing and the prison system. It concludes that the danger of producer capture is much greater in the former, because police are suspicious of civil society monitoring of their role, and the culture of community policing has not yet taken hold. However, the prison system has been more open to change, with some very creative partnerships between the state and local NGOs transforming the management and ethos of some small prisons.

• • •


The involvement of civil society in social policy delivery has become, in recent years, one of the dominant tropes of the New Policy Agenda.1  Of course, there are many points at which ordinary citizens may become included in social policy – in designing policies, allocating resources, giving advice to government bodies, delivering services on the ground, monitoring implementation and giving feedback to state agencies. Citizens, both as individuals and in groups, may be called upon as lay experts, clients, and end-users of services, and as more generic stakeholders in the social fabric. In some cases civil society participation is mere window-dressing, with policy-makers proceeded unhindered down familiar paths. There have, however, been some very significant attempts in some policy areas, or in particular geographical locations in Latin America, to make participation meaningful, a tool as much for citizen empowerment as for improvement of public services. Insofar as citizens can exercise a tangible influence on policy outcomes, and have sufficient resources and institutional stability to resist co-option and maintain autonomy, this engagement will be termed civil society-statepartnership for the purposes of this paper.

In recent years criminal justice reformers have been attempting to extend to the criminal justice institutions the principles of civil society participation already well established in other, “softer”, policy areas. This reflects in large part the evolution of the human rights community from reactive, ad hoc protests against institutional violence, to a proactive stance aimed at analyzing and restructuring the system. This paper examines the fruits of such attempts in the areas of policing and citizen security, and of penal policy (prisons and sentencing). It finds remarkable progress and innovation in some areas, yet few advances and entrenched institutional resistance in others. I distinguish between two principal modes of civil society engagement: (1) monitoring and oversight; and (2) of constructive engagement and partnership. The first mode is inevitably antagonistic to some degree, with the community performing a watchdog role, and the authorities reacting, generally, with secrecy and hostility. The second mode is more creative, but requires not only that civil society is moved to engage in the areas of rule of law and justice, but also that the state agencies relinquish some of their power and prerogatives, and provide the institutional infrastructure necessary for this interface.

Civil society and the state in Brazil

This paper focuses on the case of Brazil, and specifically on the policy area of crime and justice. On the one hand, Brazilian civil society is acknowledged to be relatively dense (if patchily so), which in itself is a function of the many institutional tools for enabling participation that have been made available since the transition to democracy. The 1988 Constitution was key in this respect: the drafting process was one of the most participatory in Latin America with 122 grassroots amendments presented by social movements, totaling over 12 million signatures, many of which succeeded in altering the final text.2  In particular, the new Constitution institutionalized various forms of popular input into governance and policy-making: plebiscites and referenda, public hearings, people’s tribunals and, most relevantly for our purposes, the creation of a plethora of state-civil society councils, at all three levels of government, to advise on a range of social policy areas (Draibe, 1998; Tatagiba, 2002).3

These mechanisms may be broadly categorized into three main groups: (1) the public policy management councils (known as conselhos gestores), which are statutory in character and are charged with overseeing particular ongoing social policies (health, education, social services, children’s welfare). They have legal powers to define priorities, set out budgets and monitor policy implementation; (2) more ad hoc councils set up to deliver special government policies (for example school meals, employment, housing, food distribution and rural development); and (3) thematic councils, which tackle issues such as race, disability or women’s rights. The latter have no statutory character and may be set up on local initiative.

All three types occupy an institutional space that is enshrined in legislation of some description, federal or subnational, making these spaces of “invited participation” (Cornwall, 2002). This guarantees them some level of resourcing and continuity, although political clientelism and co-option are a constant threat. All three tend to have a mixed composition, generally half representatives of civil society and half representatives of relevant government agencies. The “council” model of civil society-state relations has undoubtedly deepened the level of civic association in Brazil: it is estimated that by 1999 there were approximately 45,000 members of health councils alone up and down the country (Tatagiba, p. 48).

In particular, the Brazilian Workers’ Party (Partido dos Trabalhadores – PT)has been a key player in promoting and consolidating these institutional spaces and a pioneer in its municipal and state administrations in opening up the political and policy process to forms of social participation such as the famed Participatory Budget.4  These participation spaces and processes have the potential not only to increase the capacity of both civil society and the state to operate in their separate spheres, but also to bring them together as joint stakeholders in the effective solution of social problems. The party has both used the advisory council model as it currently exists, and also sought to modify it in several policy areas in order to make it less prone to co-option and more responsive to the opinions and needs of organized, as well as unorganized, civil society.5

Civil society and the criminal justice system

All bureaucracies tend to be insular and self-serving but their degree of resistance to external influence varies, and not all policy areas are equally open to civil society engagement. The criminal justice system has traditionally been the most closed as it is composed of institutions that constitute part of the state’s monopoly (in theory at least) of coercive power. Operators of the criminal justice system tend to possess a marked esprit de corps, based on their professional training and on the social control responsibilities that they exercise. In consequence they tend to be extremely resistant to outside scrutiny or interference in the operation of their institutions.6

In Brazil the professional associations of judges, public prosecutors, and police officers have been able to flex their collective muscle in a number of ways, with the police blocking long-awaited constitutional reforms,7  and the judges fending off measures that they regard as an attack on their autonomy.8  Surveys conducted with judges and public prosecutors in the mid 1990s revealed that 86.5 percent of judges were completely opposed to any form of external control over the judiciary, whilst prosecutors showed a slightly more democratic face, with only 35 percent totally opposed to outside monitoring of their own institutions. Nonetheless, they still felt that any such body should be composed primarily from their own ranks (Sadek, 1995; 1997). However, a sequence of subsequent court-related scandals eroded that position and judges now grudgingly accept the necessity of a mixed judiciary-civil society oversight board as a means of recuperating lost legitimacy. This measure was finally approved in December 2004 in a long-awaited judicial reform bill. Similar surveys of the civil (investigative) police station chiefs revealed that any form of inspection of their activities was consistently ranked lowest in terms of contribution to better policing, although the creation of community policing councils was cautiously welcomed (Sadek, 2003).

This problem of producer capture and corporate mentalities is by no means specific to Brazil. Indeed it is the product of the very way in which modern states handle social conflict, crime and deviancy. As many critical penologists have pointed out, in the retributive model of justice, crime is viewed as a violation of the state. The justice system therefore determines blame and administers pain in a contest between the offender and the state, in which the victim or the wider community are largely absent or silent (Zehr, 1990). Conflicts have become the “property” of the state (Christie, 1977), on which logic state agents build their edifice of professional expertise. This expertise is used both against fellow legal system operators and lay people, as a means of defending their monopoly over different aspects of the law and order apparatus.

The Brazilian justice system institutions are marked by atomization and hyper-autonomy at both an institutional level and individual operator level, with rivalry and competition between the different institutions of the criminal justice system – the two branches of the police (civil and military), prosecutor’s office, courts and prisons – as well as between the different branches of state government which control them. For example, the civil police in Brazil are not merely an investigative force, as in other countries, but also have a quasi-judicial function. The police investigation mirrors that conducted by the courts, thus making the police chief – who must have a law degree – a de facto investigating magistrate, and the police station a “registry” office staffed by a legal “clerk”. Such “lawyerization” of the police (Cerqueira, 1998) puts the police in competition with the judiciary and Prosecution Service for control of criminal investigation. It is this context that shapes the degree and mode of civil society input into the justice system.

All the above has made it very difficult for groups in civil society to redefine the terms of the debate about law and order. Neild (1999) points out that the terminology employed is crucial to how ideas of “security”, and the relationship between the state and citizen are framed. The concept of “national security” establishes the idea of force majeure and effectively allows the security forces a free hand in pursuing, by all means necessary, some notion of a national interest. The militarized character of the major police force in Brazil, created in its current form under the authoritarian regime of 1964-1985, continues to reflect the national security logic of that period.

The term currently most used in Latin America and in Brazil is that of “public security”. Here the good being protected is still the interest of the state and of the public authorities, albeit often at a very local level. Those who are powerful enough to capture the public sphere and its resources are also able to access its law and order agencies. However, those who are excluded by virtue of their social class are left, by definition, unprotected. According to Article 144 of the Brazilian Constitution, the police’s mission is the “preservation of public order”, as defined under the section dedicated to “The defense of the state and of its democratic institutions”, in which “public order” and “social peace” are paramount concerns. The figure of the citizen is absent, even in a document that articulates the fullest statement of civil liberties. On a rhetorical level, at least, the needs of the state continue to override those of the individual.

The most recent coining, that of “citizen security”, removes the power to define fear, crime and security from the state and the socio-political elite and delegates it to members of the public. In this formulation, the state authorities are at the service of the population, not vice versa. Citizen security and safety are, in ideal terms, based on policing by consent, not by repression, on punishment for rehabilitation, not retribution, and founded on the principles (and constraints) of universal human rights and civil liberties. All three of these conceptualizations of security have been in currency in Brazil, and are employed at different moments by the state authorities, by the mass media, and by civil society. For example, even whilst the current PT government is clearly an proponent of citizen security, as laid out in its own policy guidelines,9  it is still under pressure in some quarters to acknowledge the drugs trade and narco-violence as an issue of national security (the so-called “colombianization” of Brazilian inner cities). Periodic calls for “hard line” policing methods and a visible oscillation at state government level between tough and more “community-oriented” policing strategies illustrate the dynamism of this continuing debate over the very terms of reference, and the importance of civil society engagement with the state in this arena.



In the area of policing, civil society organizations have been created to achieve two ends: (1) to monitor the activities of the police, particularly in relation to accusations of human rights abuses; (2) to work together with the local police in community-police liaison councils in order to allocate policing resources according to local needs and priorities.


Following the transition to democracy in Brazil, there has been a steady increase in crime and violence, matched by a similar rise in police abuses – excessive use of force, summary executions and use of torture on criminal suspects. This is not the place to rehearse the various analysis of police dysfunction in Brazil (Chevigny, 1995; Human Rights Watch, 1998; Pereira, 2000); suffice to say that police inefficiency and systemic abuse of human rights are overdetermined by underresourcing, corruption, lack of training, procedures and discipline, impunity inherent in the bias of the military courts (which try offenses by military police) and the internal affairs units, long run institutional practices, and a public security mindset that reflects and reinforces social stratification and inequalities.

It was clear by the mid-1990s that the police needed to be brought under civilian oversight of some kind. The state government of São Paulo under PSDB founder Mário Covas pioneered a new institution, that of the Ouvidoria da polícia, set up in 1995. More followed, initially in states governed by the left or centre-left.10

The ouvidorias are generally housed in the offices of the state secretariat for law and order, or equivalent, and are therefore part of the executive.11  Their brief is, literally, “to hear” (from the verb “ouvir”) complaints about police misconduct, corruption or omission from the public,12  prepare an initial case summary, pass on the complaints to the police internal affairs units, and track the progress of the investigation. The ouvidoria may also refer cases to the Prosecution Service. Although generally translated as “ombudsman’s offices”, they do not possess the independence and wide powers that such entities have elsewhere. Police internal affairs units continue to monopolize the resources and remit to carry out investigations into alleged police misconduct, and often obstruct or refuse to open an inquiry. Therefore theouvidorias constitute in institutional terms a form of semi-independent internal control.

Nonetheless, they have achieved the highest degree of transparency of all the police oversight mechanisms.13  They broke new ground in publishing the first reliable figures on police shootings of civilians, as well as on police fatalities on and off duty. The ouvidorias have also contributed significantly to breaking the culture of police impunity in Brazil. Members of the public are guaranteed anonymity, crucial in overcoming the population’s real and justified fear of police reprisals. Complainants are now increasingly emboldened to report abuses openly, a shift that must reflect greater confidence in the state authorities. In 2000 most complaints to the Rio de Janeiro ouvidoria were made anonymously: from January to July 2001, some 150 complaints were made in person. Rio, in common with around half the states in Brazil, now has a witness protection program for use in such cases.14  When the ouvidorias encounter bureaucratic inertia, obstruction or hostility, they can have recourse to the media, using a “name and shame” strategy,15  and the number of complaints against police tends to rise noticeably when incidents receive widespread media coverage

Strong links to civil society have been crucial for the ouvidorias to maintain their legitimacy and stance of independence from the administration. The São Paulo ombudsman is appointed from a list of three candidates put forward by the state Human Rights Council, and is backed by a board of leading lawyers and human rights activists. The Pará office is governed directly by the state police advisory committee (CONSEP) and the most successful ombudspersons to date have come from a background of human rights activism and hence have high credibility.

As the police force has traditionally been a closed institution and public consultation on policing virtually unknown, the ouvidoria is the first government institution to solicit the views of members of the public and performs an invaluable feedback function. The notion that the public should have a right to oversee, control and determine the actions and priorities of the police represents a significant cultural shift in Brazil of which the ouvidorias are both a reflection and a constitutive element.

Due to the inherently conflictual nature of oversight mechanisms, which are obliged to criticize the institutions they are supervising, “partnership” may seem an odd term to use for theouvidorias. Certainly the police regard them more as sparring partners than as collaborators. However, it would be wrong to assume that the police are simply the instrument of the state authorities, or under their heel. Often the elected authorities are challenged by autonomous enclaves within the security apparatus, which they can only undermine with the active support of civil society.

Community-police liaison committees

One of the primary means of moving towards a model of policing based on consent and cooperation is to institute spaces in which the police and local community are able to come together to debate local needs and priorities. Police-community liaison committees (Conselhos de Segurança – CONSEG) were pioneered in Maringá, Paraná state, in 1974.16  São Paulo state followed suit under the progressive, democratic government of Franco Montoro, regulating these new bodies in 1985 and 1986. By 2002 the state claimed to have over 800 CONSEG operating in more than 520 municipalities.

Ideally, these bodies exist to encourage cooperation with the local police force and a “community policing” operational style, to overcome traditional mistrust and suspicion, and to effectively “municipalize” policing, that is, make it responsive to neighborhood needs rather than to priorities set at the level of the state government. The CONSEG could serve, in principle, as part of an attempt to modernize the police, rendering it an accountable and responsive public service rather than a repressive state bureaucracy driven by its own agenda. It is also claimed that the reorientation of the police combined with the local community’s involvement in monitoring and reporting crime and undertaking preventive action can reduce crime levels significantly. For instance, the city of Lajes in Santa Catarina reported a 47.7 percent drop in theft and robbery following the installation of ten CONSEG.17

However, as with so many aspects of the criminal justice system in Brazil, no empirical studies have been carried out on these councils, despite their impressive numbers. What seems clear from a reading of the extremely bureaucratic regulation of these bodies is that they still fall firmly under the control of the police and public security apparatus of the state. The legislation for Paraná states that “The Community Security Council must give all necessary support to the Public Security bodies, given that its job is to cooperate, represent, check up and make demands on the Public Security authorities and other organized elements of society, but without interfering in the running of the former”. In São Paulo state the local civil police chief and military police commander are automatically co-opted members of any council and take the initiative in seeking out the organized elements of the local community (“forças vivas da comunidade”) which are earlier defined as “representatives of the municipality, of local associations and other bodies providing relevant services to the community”.18  Much of the regulation is taking up with procedures for elections and the proper use of logos, coats of arms and even an official CONSEG song. Membership of the CONSEG is numerically small and closed, due to the conduct of internal elections.

Anecdotal evidence suggests that they are also not all “representative” of the local community, being mainly composed of local businesspeople. Much of their activity seems to centre on raising money to buy police equipment (as basic as new tires for police vehicles) for which generosity they may expect preferential attention in return. Indeed, the CONSEG seem to present a good case of “mutual capture”, whereby the police play a guiding role in setting up, running and finding members for the Council, whilst the members enjoy privileged access to a public good. This is a problem of which the state authorities are not unaware, identifying as a common obstacle the “coming forward of leaders not equipped for community work: people who want to extract some personal, financial or electoral advantage from the CONSEG”, and thus undermine their legitimacy. Indeed, there is a very thin line between this type of capture and the kind of alliance that elements of “uncivil” society have made with local police, assisting death squad activity aimed at eliminating those classed as social undesirables.19

An alternative model was instituted in São Paulo City under the administration of PT mayor Marta Suplicy (2001-2004). The municipal police force was revamped as an “ideal model” of preventive policing, and new civil society-police structures set up. The federal constitution allows municipalities to set up municipal police forces for the purposes of protecting city property. Whilst this is a limited remit, there have been recent moves in Brazil to “municipalize” policing, in part to circumvent the enormous structural obstacles to a complete reform of the state-level policing apparatus.

Benedito Mariano (who had been the first police ombudsman in Brazil, during Mário Covas state administration), was brought in to head the city’s new secretariat for public security, and doubled the numbers of municipal guard from 4,000 to 8,000 (including a 30 percent quota for women officers). The Guard’s contact with the local community is intended to be a cornerstone of their preventive strategy – indeed, their work is probably closer to what might be termed “community” policing than most other experiments that go under the name in Brazil. The community is consulted on a regular basis through Community Committees set up in six regions of the city. In this case, every meeting of the Committee is open to all comers, although a standing committee is elected. The co-opted members include the Regional Inspector of the Municipal Guard and a representative of the sub-prefecture, but otherwise the balance is tipped much more heavily towards civil society than is the case in the CONSEG.20  The Secretariat claims that 2,870 people participated in 56 meetings between October 2002 and December 2003, an average of 50 per meeting, of which two-thirds were civil society representatives. It seems that the participatory, democratic ethos employed by the PT in its consultation exercises in other areas of municipal governance has influenced its conduct of the civil society-state partnership in this newer field of citizen security (Baiocchi, 2003).

Community policing

Analysis of policing in Brazil since the return to democracy have tended to emphasis its authoritarian characteristics, its ineffectiveness, and the degree to which the police actually contribute to criminal activity through corruption and organized crime, and routinely commit gross human rights violations such as torture and summary executions of criminal suspects. In particular, attention has focused on the militarized police, a uniformed state-level force responsible for carrying out preventive policing, with its military structure, hierarchy, code of conduct, training and corporate ethos.

Various studies of extrajudicial executions demonstrate the belligerent attitude that the military police adopt in relation to the community (Cano, 1997) and suggest that this is a residue of the National Security Doctrine of the military period by which the civilian population was viewed with suspicion, as the potential “enemy”, to be controlled and contained. This antagonistic stance of the police vis-à-vis the citizens whose security they are supposed to be safeguarding came to be viewed by critics and reformers as counter-productive, and in violation of Brazil’s commitments to human rights and civil liberties. It was within this environment that tentative experiments with community policing have been undertaken.

Community policing theory presupposes a quite different relationship between police and public. It is based on principles of trust and collaboration, ongoing interactions with civil society, attentiveness to the expressed needs and priorities of the population, information sharing leading to intelligence-led policing, conflict mediation and resolution, and crime prevention rather post hoc repression. It was pioneered in Rio de Janeiro in the early 1990s by the then Commander of the Military Police, Carlos Magno Nazareth Cerqueira, backed by a local human rights NGO, Viva Rio, under the leftwing administration of Leonel Brizola (1991-1994). The first projects were implemented in a piecemeal manner in a number of neighborhoods in Rio de Janeiro city. The first major scheme was set up in Copacabana but lasted a mere ten months, dismantled by the incoming administration of Marcello Alencar, who adopted a “tough on crime” stance, giving the new Secretary of Public Security carte blanche with a shoot-to-kill policy (Musumeci et al., 1996).

In 2001 Rio de Janeiro tried another community policing project, this time in the small central favella of Cantagalo, run by a military police major connected with a group of justice system reformers who fell out with the Garotinho state government. This initiative attempted to counter the mainstream policing practices in Rio’s favella which in the past have consisted of large-scale armed “blitzes”, firefights with the drug traffickers, followed by withdrawal. The police began by taking over a vast community centre and abandoned hotel at the top of the hill and running cultural, educational and training activities for the local youths, taking the place of local NGOs who are too scared yet to operate in the favella.

Similarly, in violent low-income areas in São Paulo, police have ended up calling in their own institution’s social services: military police doctors, dentists, and physical education teachers. As police are often the only agency of the state physically present in the more marginalized neighborhoods, it is evident that community policing projects require the collaboration not just of the local population, but also of other parts of the state apparatus in a multi-service approach in improving, simultaneously, quality of life, social capital and citizenship trust in and access to justice and rule of law services.

The core question about community policing in Brazil concerns its still marginal status. The Cantagalo project was isolated from the mainstream of policing activity in Rio de Janeiro, and boycotted by the municipal government, for reasons of territorial possessiveness and electoral competition (a long-standing rivalry between successive governors and mayors), and thus denied many vital social services that would have bolstered its legitimacy and effectiveness.21  Although initially the project started well, with 50 police officers purged on charges of corruption and violence, old habits died hard and police abuses gradually escalated again (Global Justice, 2004, p. 38).

Some community policing schemes are in name alone.22  In some 100 neighborhoods in the state of São Paulo mobile police cabins have been set up. However, as police only leave their post – reluctantly – when approached by a member of the public requesting assistance, they can scarcely establish the indispensable, durable and organic links with the local population.23  A comparison of public attitudes to conventional and community police in Brazil shows that public trust in the latter can only be generated by increased visibility and outreach (Kahn, 2004). In short, without political backing and an upheaval in institutional cultures, the “community” will remain the enemy of the police force, not a partner.


Prisons and the penal system

Although successful crime prevention and resolution requires cooperation from the local population, this is less evident in the case of sentencing and imprisonment, for the punishment of offenders has generally been abrogated by the state as its exclusive prerogative. Nonetheless, this is now being challenged in Brazil, initially in response to a crisis of state capacity – the rising concern about prison conditions and the collateral effects of rioting, break-outs and hostage taking episodes that characterized the late 1990s, and then inspired by global penal reform movements and radical new ideas such as restorative justice, that put the victim, offender and community, not the state, centre-stage.


It was not until the latter half of the 1990s that public attention shifted to the fate of those held in police custody or in the prison system. The pioneers in raising public consciousness were undoubtedly the members of the Catholic Church’s “Prison Ministry” (Pastoral Carcerária)24  which had some 3,000 lay and religious volunteers around the country regularly visiting prisons, offering practical and spiritual support and bearing witness to the daily abuses suffered by prisoners. In 1997 the plight of detainees was taken up by the National Council of Bishops (CNBB) as the theme of their Lenten campaign. They also began increasingly to seek allies in the Bar Association, in young and enthusiastic prosecutors, and in the judiciary.25  Lobbying of international organizations resulted in visits and reports by the Inter-American Human Rights Commission, Human Rights Watch, Amnesty International, the United National High Commissioner for Human Rights and the UN Special Rapporteur on Torture, as well as by the Human Rights Committees of a number of state legislatures.

In fact, civil society oversight of the prison system had already been established in the 1984 Law on Sentence Serving (LEP), which required the judge of every circuit with a prison to appoint a Community Council (Conselho da Comunidade), composed of representatives of the local community, to visit the prison at regular intervals, inspect conditions, and assist prisoners.

Despite the support of both the Cardoso and Lula governments for this scheme, relatively few have been set up. The Ministry of Justice holds no information26  as to how many exist and those that do operate generally do so in a vacuum without institutional support and connections.27  For example, the one in Rio, which is relatively active, lost its offices in the local Justice Secretariat. It also stopped reporting to the local judge due to his hostility to their work.

This case illustrates well a common problem of state-civil society interaction. One part of the state apparatus (the executive, and the national law) supports these groups, in principle. However, they can only be brought into existence by the local judiciary, which either is ignorant or resistant to having others intrude on “their patch”. Without training, guidelines for their activity, an established feedback mechanism into the local state and sufficient autonomy and backup to resist pressure from those officials they may criticize, these Conselhos have been doomed to be a dead-letter.28  It is perhaps understandable why some have decided to substitute the state altogether inside some prisons rather than engage with it as an outsider.

Given the weakness of the Conselho structure and the relative strength of the Pastoral, local prison administrators were tempted to simply “hand over” inspection responsibility to the Catholic Church, thus absolving the state from a degree of responsibility either to set up effective internal monitoring procedures, or to strengthen the institutional apparatus by which civil society could properly exercise its prerogative of oversight. Despite all the activism in this area in the late 1990s, very little progress has been made on either fronts.29

Community-run prisons

Periodically the capacity of the Brazilian state to run the prison system – notorious for its overcrowding, endemic violence, appalling conditions of detention, poor management and inability to alter offending behavior (Amnesty International 1999; Human Rights Watch 1998) – has been called into question. The private security industry is, after all, booming in Brazil30  in response to the deficiencies of the police, yet privatization of the prison system has been periodically debated since the 1980s but always rejected.31

One of the most surprising features of Brazil’s prison system is the existence of an innovative partnership between state and civil society in the area of prison administration. Whilst prisons are generally managed either by the public sector, or via contracts with the private sector, a third paradigm is possible.

The first experiment in community participation in running prisons occurred in the 1970s, in São José dos Campos, São Paulo state, where a Catholic group took over completely the running of a decrepit and overcrowded local jail. Dr. Nagashi Furukawa, then a local district justice in Bragança Paulista, visited the institution in the 90’s, and thereafter established contact with a group in his town that defended similar concepts. His initiative resulted in the first cooperation agreement between the state and a NGO.

The Bragança Paulista jail was completely refurbished, applying a model that came to draw the both domestic and international attention. Shortly afterwards, upon being appointed Secretary of Prison Affairs of the state of São Paulo, Dr. Furukawa, he began to replicate this successful experience. So far, 20 Resocialization Centers (Centros de Ressocialização – CR) have been set up, each holding approximately 210 inmates,32  administered in an innovative partnership between the state prison authorities and a dedicated local NGO, on the basis of formal co-operation agreements.*

The non-profit organization deals with the day-to-day running of the prison and rehabilitation of prisoners whilst discipline and security remain under state control. Most of the CR have been purpose-built to a new architectural design, but a number, such as the two original public jails in São José dos Campos and Bragança Paulista, have been upgraded and converted.

On the basis of field research carried out in four CR in October 2004, 33  they appear to be outstanding in relation to protection of the human rights of prisoners and staff, elimination of violence and drug abuse in prison, decent conditions of detention, potential for significantly reduced levels of re-offending, excellent social, education, occupational and psychological support given to offenders and their families, value-for-money in terms of quality and costs,34  increased transparency and checks-and-balances for both treatment of detainees and use of public resources and improved community relations with the justice system.35  Recidivism is claimed to be a third of national levels.36  This is in fact a fairly modest performance, as compared to other successful measures, within the framework of a model conceived so as to avoid family breakdown, unemployability, institutionalization, drug consumption and poor self-esteem that conventional prisons invariably produce.

Another form of cooperation between the civil society and the state with a view to face the problems of the prison system are the APACs (Associação de Proteção e Assistência Carcerária – Associations for Carcerary Protection and Assistance).

Although the original APAC model is essentially faith-based, reliant on “saturating the prison environment with religious programming and instruction” (Johnson, 2000), the CR rely principally on two factors for rehabilitation: work (in some CR 100 percent of prisoners work) and rebuilding family relations (the CR only take offenders whose families live in the vicinity).37  Visiting hours are generous and most families spend several hours each Sunday on the visit, and the money earned in prison is often of assistance to the low income families, which the NGO’s social work staff go to great lengths to support whilst their relative is in prison. The families form a bridge to the local community, helping it overcome hostility and demonization of the prison and its inmates.

Equally, both the CR and APAC units deliberately subvert the construction of a “prison culture”. Local business also benefits materially, as the NGOs have much greater flexibility than the state to purchase goods and services locally. The APAC founders from São José dos Campos have now sought out new partnerships in Minas Gerais, mainly via the judiciary, not prison authorities, and currently run a number of units with prison guards within the perimeters (the CR allow prisoners to unlock internal doors, but prison guards monitor all activity).

The NGO and the state authorities also provide an excellent system of checks and balances for one another as the contract makes very careful stipulations in relation to accounting and transparency. This is perhaps the truest partnership, or “co-production” (Joshi & Moore, 2004; Masud, 2002) of all the cases analyzed, and the one in which the balance is tipped most heavily toward civil society. As in the case of the community police councils set up in São Paulo, two factors were crucial: the presence of a committed change agents, and the political-institutional space and backing in which to trial a new model. It also highlights how spaces within a rather fragmented justice system may be partially appropriated by civil society. This may be due to fragility and neglect of the state but may also constitute a case of the state “inviting participation” on grounds that are much more equal and collaborative.

Nonetheless, they remain invisible within the prison system as a whole. There is no mention of them in the planning and policy documents and statements that issue from the Ministry of Justice, nor any empirical evaluation studies. It seems perverse that the state would not actively claim ownership of successful facilities under its aegis.



The unevenness of the fabric of local civil society affects the capacity for all these partnerships in public policy to succeed. Even the police are conscious of the degree to which it is actually creating social capital. The São Paulo state founding policy document on the CONSEG cites Putnam’s seminal work38  on the relationship between social capital and social development and notes “the police will tend to be more effective when they help communities to help themselves”.39

The regulation of the CONSEG in Santa Catarina lists as one of the core objectives “to develop civil and community spirit in the area”.40  Thus any state entity that wants to build in community consultation as part of its practice inevitably ends up attempting to foment and build the civic groups it wants as its partners. How to do this without co-option has been the perennial challenge facing all the areas of public policy in Brazil in which the “council” model is employed for civil society input.

Alongside the capacity of civil society to articulate its needs and interests, the other major problem lies with the receptivity of the state agencies. For example, at present the CONSEG and prison community councils will only be as successful as the local police commanders and judges allow them to be. The resistance of state agencies to change may be cast in many ways – as path dependency, as bureaucratic cultures, as territorial defensiveness – and I have argued that the very nature of the policy area of crime and violence exacerbates these tendencies. Nonetheless, change agents, both individual and collective, have proven adept at finding spaces and places within the state apparatus where the state authorities are indifferent and exhausted (the APAC prisons), have escaped the dominant institutional vices (the municipal guards), or are looking for radical new policy approaches (CR). Where there is local political support, these spaces offer valuable arenas for forging new forms of state-civil society partnership.

In the cases I have examined, partnership takes different forms depending on the role of civil society (critical watchdog, advice and support, or co-production), and this in turn affects the power asymmetries at play.

The example of the CONSEG is the clearest case of state actors capturing civil society – but this turns out to be a mutually beneficial arrangement due to the exclusionary design of the councils. The failure of the prison councils to make any progress must be attributed to inertia on the part of the judiciary. The local judges most likely see no benefit to themselves in setting up the councils (whereas the police get quite tangible goods from setting up the CONSEG) and thus have been boycotted this provision in the law, despite all exhortations from their superiors.

While it is to be expected that the state, with its superior resources, might always have the upper hand, this is not the case in the APAC and CR prisons, where the local community has managed to mobilize human capital resources that have made significant improvements both to the human rights and treatment of the detainees, and their prospects for rejoining their families and not reoffending. Where these partnerships work, they can make considerable contributions to improvement of citizen security in Brazil.

• • •


* See Last access on 14 March 2005. [E.N.]

1. For a critical discussion of the idea of “participation” in development policy circles see A. Cornwall, 2002.

2. For example, women’s organizations estimated that around 80 percent of the amendments they put forward had been retained in the final version.

3. The first two national advisory councils, in health and education, were set up in 1937 with the overhaul of the state under Vargas. There are now 25, most set up in the 1990s. The 1988 Constitution also provided for municipal and state-level councils.

4. There is by now a huge literature on the participatory budget process, but for a good critical overview see Baiocchi, 2003.

5. For an analysis of the PT’s modification of the women’s council model see Macaulay, 2003a.

6. For details of other civil society initiatives for transparency in the justice sector see documents on the website of the Due Process of Law Foundation Last access on 15 March 2005.

7. The state-level Military Police, which constitute some 80 percent of the force in Brazil, used their lobbying influence among senators to block proposals put forward under the Cardoso government to “deconstitutionalize” them, that is, to remove all mention of the military police in the constitutional text, thus allowing state governments to retain, abolish, or merge them with the Civil Police.

8. On the debates about accountability and the judiciary see Macaulay, 2003b.

9. In 2002, the PT-linked think-tank Instituto da Cidadania produced a 120 page set of recommendations for reform of the justice system, drawn up by leading policy experts. This became the basis for the Lula government’s Integrated Public Security Plan.

10. Rio de Janeiro, in March 1999 under Anthony Garotinho (PDT); Minas Gerais in 1997 (Eduardo Azeredo, PSDB); Pará in 1997 (Almir Gabriel, PSDB); Rio Grande do Sul in August 1999 (Olívio Dutra, PT); and others in Pernambuco, Espírito Santo, Rio Grande do Norte, Mato Grosso, Bahia and Ceará.

11. Exceptions are those in Pará, subordinated to the state council on law and order (CONSEP), and Minas Gerais, linked to the governor’s office.

12. They receive all manner of communications from the public in relation to the police. Priority is given, however, to serious allegations regarding the right to life, and police corruption.

13. See Lemgruber et al. (2003) for an in-depth study of the ouvidorias; and Macaulay (2002) for a comparison with other forms of police oversight.

14. However, individuals with a criminal record are excluded from the program, thus depriving a good number of police torture victims of this protection.

15. It is noticeable that the successors to the first ouvidor in Rio de Janeiro and São Paulo, Julita Lemgruber and Benedito Mariano, use the media a lot less. In 2001 the then Rio ouvidorrejected what he dismissed as work “just to get newspaper coverage”. On the other hand, media coverage gives a degree of visibility and protection to the ombudsman – he was forced to resign shortly afterwards due to lack of political backing.

16. In 2004 the state of Paraná had 280 CONSEG with 46 in Curitiba and 74 in the metropolitan region.

17. A Notícia, 16 May 2002. As of May 2002 Santa Catarina had 31, but planned to install one in every municipality. The law authorizing them was only passed in March 2001. Similar levels of crime reduction are cited for Embu in São Paulo state.

18. Regulamentação dos CONSEG, Resolução SSP n. 47, 18 March 1999; and Decree n. 25366, 11 June 1986.

19. Similar problems are reported for Chile by Frühling (2003, p. 38).

20. “Projeto do Programa das Comissões Civis Comunitárias”, unpublished internal document, Prefeitura do Município de São Paulo.

21. Interview with Major Antonio Carballo, Cantagalo, July 2001.

22. For details of other projects see Mesquita & Loche, 2003, pp. 193-199.

23. Information from Guaracy Mingardi, July 2001.

24. Much of the Pastoral’s effectiveness must be attributed to the inspired leadership and political savvy of its longtime leader, Father Francisco “Chico” Reardon, an Irish-American priest and naturalized Brazilian who, sadly, died in 1999.

25. Other human rights groups did visit places of detention intermittently, generally following some violent episode. None, however, had the consistent presence of the Pastoral.

26. I have data only for São Paulo state where there are 54 court districts with a functional Council, 23 with a dormant Council, and 62 with no Council. No qualitative data are available.

27. Interview with Tania Kolker, vice-chair of the Rio de Janeiro Conselho da Comunidade, July 2001.

28. There is also minimal interaction with the state authorities responsible, on paper, for prison inspection: the local prison judge, the internal affairs department of the state prison administration, local Penitentiary Council (essentially a parole board) and public prosecution service.

29. That said, there is currently a UK government funded project aimed at setting up a prisons inspectorate in Brazil, at state level initially, and discussions are ongoing concerning capacity building for the Conselhos.

30. According to data from the Private Security Companies Union in 1985 the ratio of police to private guards was 3:1. By 2000 this had reversed. Some 1,200 private companies were employing 400,000 registered guards plus 600,000 unregistered guards, making the industry worth U$ 4.5 billion in 2000.

31. Brazil has six semi-privatized prisons in Paraná state.

32. The United Nations recommended that prisons have no more than 500 inmates, as authorities tend to lose control of large units.

33. Funded by a grant from the UK Socio-Legal Studies Association.

34. The NGO receives a per capita allowance for each inmate for food, building maintenance and so forth. As the purchase of goods and services is being done by a private, not public body, the NGO is not tied into government supply contracts, and can sack staff for poor performance. The cost per prisoner in a CR is half that in a state-run prison, and one third of that in the few semi-privatized units.

35. The warden of Bragança, a civil police officer, admitted in a conversation in 1999 that he had staffing problems “two are drunkards, two are nuts and the other two are ok and have to keep an eye on the others”.  Clearly local civil society input here far outweighs that of the conventional prison staff, who must adapt the CR’s unusual ethos.

36. There is no systematic measurement of recidivism, either by facility, by state or nationally. The national databases required to track repeat offenders have not yet been put in place.

37. The APAC prison in Caruaru, Pernambuco, organized father-child art workshops and escorted day trips to the zoo.

38. David Putnam, Comunidade e democracia: uma experiência da Itália. Rio de Janeiro: FGV, 1996.

39. See “Informativo Institucional” at Last access on 15 March 2005.

40. “Regulamento dos Conselhos Comunitários de Segurança”. Secretaria de Estado de Segurança Pública de Santa Catarina, Conselho Superior de Segurança Pública, May 2001.

• • •


Amnesty International. No One Sleeps here Safely: Human Rights Violations against Detainees.London: Amnesty International, 1999.

Baiocchi, Gianpaolo (ed.). Radicals in Power: The Workers’ Party (PT) and Experiments with Urban Democracy in Brazil. London: Zed Press, 2003.

Cano, Ignacio. Letalidade da ação policial no Rio de Janeiro. Rio de Janeiro: ISER, 1997.

Cerqueira, C.M.N. “Questões preliminares para a discussão de uma proposta de diretrizes constitucionais sobre a segurança pública”. Revista Brasileira de Ciências Criminais, 6(2), 1998.

Chevigny, Paul. Edge of the Knife: Police Violence and Accountability in Six Cities of the Americas.New York: New Press, 1995.

Christie, Nils. “Conflicts as Property”. British Journal of Criminology, 17(1), 1977.

Cornwall, Andrea. “Making Spaces, Changing Places: Situating Participation in Development”. IDS Working Paper 170, 2002.

Draibe, Sônia Miriam. “A nova institucionalidade do sistema brasileiro de políticas sociais: os conselhos nacionais de políticas setoriais”. Caderno de Pesquisa 35, UNICAMP, 1998.

Frühling, Hugo. “Police Reform and Democratization”. In: H. Frühling et al. (eds.), Crime and Violence in Latin America: Citizen Security, Democracy and the State. Washington, DC/Baltimore: Woodrow Wilson Center Press – Johns Hopkins University Press, 2003.

Global Justice. Rio Report: Police Violence and Public Insecurity. Rio de Janeiro: Global Justice, 2004.

Human Rights Watch. Police Abuse in Brazil: Summary Executions in Rio de Janeiro. New York: Human Rights Watch, 1997.

. Behind Bars in Brazil. New York: Human Rights Watch, 1998.

Johnson, B.R. “Assessing the Impact of Religious Programs and Prison Industry on Recidivism: An Exploratory Study”. (Unpublished paper.) Washington, DC: Prison Fellowship International,2000.

Joshi, Anuradha & Moore, Mick. “Institutionalized Co-production: Unorthodox Public Service Delivery in Challenging Environments”. Journal of Development Studies 40(4), pp. 31-49, 2004.

Kahn, Tulio. “Segurança pública e trabalho policial no Brasil”. Working Paper CBS 51-2004. Oxford: Centre for Brazilian Studies, 2004. Available at Last access on 2 March 2005.

Lemgruber, Julita; Musumeci, Leonarda & Cano, Ignacio. Quem vigia os vigias: Um estudo sobre controle externo da polícia no Brasil. Rio de Janeiro:Record, 2003.

Macaulay, Fiona. “Problems of Police Oversight in Brazil”. Working Paper CBS 33-2002. Oxford: Centre for Brazilian Studies, 2002. Available at Last access on 2 March 2005.

. “Sexual Politics, Party Politics: The PT Government’s Policies on Gender Equity and Equality in Brazil”. Working Paper CBS 46-2003. Oxford: Centre for Brazilian Studies, 2003a. Available at Last access on 2 March 2005.

. “Democratization and the Judiciary: Competing Reform Agendas”. In: M. D. Kinzo & J. Dunkerley (eds.), Brazil since 1985: Economy, Polity and Society. London: Institute of Latin American Studies, 2003b.

Masud, Mohammad O. “Co-producing Citizen Security: The Citizen-Police Liaison Committee in Karachi”. IDS Working Paper 172. Brighton (RU): Institute of Development Studies, University of Sussex, 2002. Available at Last access on 2 March 2005.

Mesquita Neto, Paulo de & Loche, Adriana. “Police-Community Partnerships in Brazil”. In: H. Frühling et al. (eds.). Crime and Violence in Latin America: Citizen Security, Democracy and the State. Washington, DC/Baltimore: Woodrow Wilson Center Press – Johns Hopkins University Press, 2003.

Musumeci, L. et al. “Segurança pública e cidadania: A experiência de policiamento comunitário em Copacabana (1995-1996). Relatório final do monitoramento qualitativo”. Rio de Janeiro: ISER, 1996.

Neild, R. “From National Security to Citizen Security: Civil Society and the Evolution of Public Order Debates”, 1999. Available at Last access on 2 March 2005.

. Themes and Debates in Public Security Reform: External Controls. Washington, DC: Washington Office on Latin America, 2000.

Pereira, Anthony. “An Ugly Democracy? State Violence and the Rule of Law in Post-Authoritarian Brazil”. In: P. Kingstone & T. Powers (eds.). Democratic Brazil. Pittsburgh: University of Pittsburgh Press, 2000.

Sadek, Maria Tereza. O Judiciário em debate. São Paulo: IDESP, 1995.

. O Ministério Público e a Justiça no Brasil. São Paulo: IDESP, 1997.

. Delegados de polícia. São Paulo: IDESP, 2003.

Tatagiba, Luciana. “Os conselhos gestores e a democratização das políticas públicas no Brasil”. In: E. Dagnino (ed.). Sociedade civil e espaços públicos no Brasil. São Paulo: Paz e Terra, 2002.

Zehr, Howard. Changing Lenses: A New Focus for Crime and Justice. Scottsdale, Pa: Herald Press, 1990.

Fiona Macaulay

Lecturer in Development Studies, Department of Peace Studies, University of Bradford, United Kingdom.

Original article in English.